public inbox for git-commits@fedoraproject.org
help / color / mirror / Atom feed
* [rpms/chromium] epel9-next: - Update to 138.0.7204.49
@ 2026-08-07 16:07 Than Ngo
  0 siblings, 0 replies; only message in thread
From: Than Ngo @ 2026-08-07 16:07 UTC (permalink / raw)
  To: git-commits

          A new commit has been pushed.

          Repo   : rpms/chromium
          Branch : epel9-next
          Commit : 99ff6b4d8515b7bb058e46146cda885a2b4d2466
          Author : Than Ngo <than@redhat.com>
          Date   : 2025-06-24T23:43:49+02:00
          Stats  : +17348/-14682 in 8 file(s)
          URL    : https://src.fedoraproject.org/rpms/chromium/c/99ff6b4d8515b7bb058e46146cda885a2b4d2466?branch=epel9-next

          Log:
          - Update to 138.0.7204.49
* CVE-2025-6555: Use after free in Animation
* CVE-2025-6556: Insufficient policy enforcement in Loader
* CVE-2025-6557: Insufficient data validation in DevTools

---
diff --git a/0001-Add-PPC64-support-for-boringssl.patch b/0001-Add-PPC64-support-for-boringssl.patch
index 85c38d9..248ad06 100644
--- a/0001-Add-PPC64-support-for-boringssl.patch
+++ b/0001-Add-PPC64-support-for-boringssl.patch
@@ -1,7 +1,7 @@
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/abi_self_test.cc
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/abi_self_test.cc
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/crypto/abi_self_test.cc
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/abi_self_test.cc
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/crypto/abi_self_test.cc
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/abi_self_test.cc
 @@ -521,3 +521,289 @@ TEST(ABITest, AArch64) {
    CHECK_ABI_NO_UNWIND(abi_test_clobber_v15_upper);
  }
@@ -292,10 +292,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/abi_self_test.cc
 +  CHECK_ABI_NO_UNWIND(abi_test_clobber_lr);
 +}
 +#endif   // OPENSSL_PPC64LE && SUPPORTS_ABI_TEST
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/cpu_ppc64le.cc
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/cpu_ppc64le.cc
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/cpu_ppc64le.cc
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/cpu_ppc64le.cc
 @@ -0,0 +1,38 @@
 +/* Copyright (c) 2016, Google Inc.
 + *
@@ -335,10 +335,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/cpu_ppc64le.cc
 +}
 +
 +#endif  // OPENSSL_PPC64LE
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/crypto.cc
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/crypto.cc
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/crypto/crypto.cc
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/crypto.cc
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/crypto/crypto.cc
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/crypto.cc
 @@ -67,6 +67,10 @@ uint32_t OPENSSL_get_ia32cap(int idx) {
    return OPENSSL_ia32cap_P[idx];
  }
@@ -347,13 +347,13 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/crypto.cc
 +
 +HIDDEN unsigned long OPENSSL_ppc64le_hwcap2 = 0;
 +
- #elif defined(OPENSSL_ARM) || defined(OPENSSL_AARCH64)
- 
- #if defined(OPENSSL_STATIC_ARMCAP)
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/aes/asm/aesp8-ppc.pl
+ #elif (defined(OPENSSL_ARM) || defined(OPENSSL_AARCH64)) && \
+     !defined(OPENSSL_STATIC_ARMCAP)
+ HIDDEN uint32_t OPENSSL_armcap_P = 0;
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/aes/asm/aesp8-ppc.pl
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/aes/asm/aesp8-ppc.pl
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/aes/asm/aesp8-ppc.pl
 @@ -0,0 +1,3809 @@
 +#! /usr/bin/env perl
 +# Copyright 2014-2018 The OpenSSL Project Authors. All Rights Reserved.
@@ -4164,10 +4164,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/aes/as
 +}
 +
 +close STDOUT or die "error closing STDOUT: $!";
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/aes/internal.h
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/aes/internal.h
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/crypto/fipsmodule/aes/internal.h
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/aes/internal.h
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/crypto/fipsmodule/aes/internal.h
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/aes/internal.h
 @@ -81,6 +81,12 @@ inline int vpaes_capable(void) { return
  inline int vpaes_capable(void) { return CRYPTO_is_NEON_capable(); }
  #endif
@@ -4195,10 +4195,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/aes/in
  #endif
  #endif  // OPENSSL_NO_ASM
  
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/bcm.cc
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/bcm.cc
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/crypto/fipsmodule/bcm.cc
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/bcm.cc
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/crypto/fipsmodule/bcm.cc
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/bcm.cc
 @@ -104,6 +104,7 @@
  #include "self_check/fips.cc.inc"
  #include "self_check/self_check.cc.inc"
@@ -4207,10 +4207,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/bcm.cc
  #include "sha/sha1.cc.inc"
  #include "sha/sha256.cc.inc"
  #include "sha/sha512.cc.inc"
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/bn/bn.cc.inc
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/bn/bn.cc.inc
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/crypto/fipsmodule/bn/bn.cc.inc
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/bn/bn.cc.inc
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/crypto/fipsmodule/bn/bn.cc.inc
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/bn/bn.cc.inc
 @@ -330,6 +330,23 @@ int bn_expand(BIGNUM *bn, size_t bits) {
  }
  
@@ -4235,11 +4235,11 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/bn/bn.
    if ((size_t)bn->width <= words) {
      if (!bn_wexpand(bn, words)) {
        return 0;
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/cipher/e_aes.cc.inc
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/cipher/e_aes.cc.inc
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/crypto/fipsmodule/cipher/e_aes.cc.inc
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/cipher/e_aes.cc.inc
-@@ -1231,6 +1231,8 @@ int EVP_has_aes_hardware(void) {
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/crypto/fipsmodule/cipher/e_aes.cc.inc
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/cipher/e_aes.cc.inc
+@@ -1229,6 +1229,8 @@ int EVP_has_aes_hardware(void) {
    return hwaes_capable() && crypto_gcm_clmul_enabled();
  #elif defined(OPENSSL_ARM) || defined(OPENSSL_AARCH64)
    return hwaes_capable() && CRYPTO_is_ARMv8_PMULL_capable();
@@ -4248,10 +4248,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/cipher
  #else
    return 0;
  #endif
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/aes/asm/ghashp8-ppc.pl
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/aes/asm/ghashp8-ppc.pl
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/aes/asm/ghashp8-ppc.pl
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/aes/asm/ghashp8-ppc.pl
 @@ -0,0 +1,671 @@
 +#! /usr/bin/env perl
 +# Copyright 2014-2016 The OpenSSL Project Authors. All Rights Reserved.
@@ -4924,10 +4924,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/aes/as
 +}
 +
 +close STDOUT or die "error closing STDOUT: $!"; # enforce flush
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/aes/gcm.cc.inc
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/aes/gcm.cc.inc
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/crypto/fipsmodule/aes/gcm.cc.inc
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/aes/gcm.cc.inc
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/crypto/fipsmodule/aes/gcm.cc.inc
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/aes/gcm.cc.inc
 @@ -233,6 +233,13 @@ void CRYPTO_ghash_init(gmult_func *out_m
      *out_hash = gcm_ghash_neon;
      return;
@@ -4942,10 +4942,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/aes/gc
  #endif
  
    gcm_init_nohw(out_table, H);
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/aes/gcm_test.cc
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/aes/gcm_test.cc
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/crypto/fipsmodule/aes/gcm_test.cc
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/aes/gcm_test.cc
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/crypto/fipsmodule/aes/gcm_test.cc
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/aes/gcm_test.cc
 @@ -169,5 +169,15 @@ TEST(GCMTest, ABI) {
      }
    }
@@ -4962,10 +4962,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/aes/gc
 +#endif  // GHASH_ASM_PPC64LE
  }
  #endif  // SUPPORTS_ABI_TEST && !OPENSSL_NO_ASM
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/rand/getrandom_fillin.h
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/rand/getrandom_fillin.h
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/crypto/rand/getrandom_fillin.h
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/rand/getrandom_fillin.h
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/crypto/rand/getrandom_fillin.h
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/rand/getrandom_fillin.h
 @@ -30,6 +30,8 @@
  #define EXPECTED_NR_getrandom 278
  #elif defined(OPENSSL_ARM)
@@ -4975,10 +4975,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/rand/getrandom_fi
  #elif defined(OPENSSL_RISCV64)
  #define EXPECTED_NR_getrandom 278
  #endif
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/rand/rand.cc.inc
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/rand/rand.cc.inc
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/crypto/fipsmodule/rand/rand.cc.inc
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/rand/rand.cc.inc
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/crypto/fipsmodule/rand/rand.cc.inc
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/rand/rand.cc.inc
 @@ -431,6 +431,11 @@ bcm_infallible BCM_rand_bytes_with_addit
      // Take a read lock around accesses to |state->drbg|. This is needed to
      // avoid returning bad entropy if we race with
@@ -4991,10 +4991,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/rand/r
      CRYPTO_MUTEX_lock_read(&state->clear_drbg_lock);
  #endif
      if (!CTR_DRBG_reseed(&state->drbg, seed, reseed_additional_data,
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/sha/internal.h
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/sha/internal.h
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/crypto/fipsmodule/sha/internal.h
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/sha/internal.h
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/crypto/fipsmodule/sha/internal.h
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/sha/internal.h
 @@ -23,6 +23,16 @@
  extern "C" {
  #endif
@@ -5012,10 +5012,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/sha/in
  // Define SHA{n}[_{variant}]_ASM if sha{n}_block_data_order[_{variant}] is
  // defined in assembly.
  
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/sha/sha1-altivec.cc.inc
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/sha/sha1-altivec.cc.inc
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/sha/sha1-altivec.cc.inc
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/sha/sha1-altivec.cc.inc
 @@ -0,0 +1,369 @@
 +/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.ccom)
 + * All rights reserved.
@@ -5386,10 +5386,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/sha/sh
 +#undef BODY_20_39
 +#undef BODY_40_59
 +#undef BODY_60_79
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/internal.h
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/internal.h
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/crypto/internal.h
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/internal.h
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/crypto/internal.h
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/internal.h
 @@ -59,8 +59,9 @@ extern "C" {
  #if !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_STATIC_ARMCAP) && \
      (defined(OPENSSL_X86) || defined(OPENSSL_X86_64) ||            \
@@ -5402,7 +5402,7 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/internal.h
  #define NEED_CPUID
  
  // OPENSSL_cpuid_setup initializes the platform-specific feature cache. This
-@@ -1401,6 +1402,16 @@ inline int CRYPTO_is_ARMv8_SHA512_capabl
+@@ -1394,6 +1395,16 @@ inline int CRYPTO_is_ARMv8_SHA512_capabl
  #endif  // OPENSSL_ARM || OPENSSL_AARCH64
  
  
@@ -5419,10 +5419,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/internal.h
  #if defined(BORINGSSL_DISPATCH_TEST)
  // Runtime CPU dispatch testing support
  
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/perlasm/ppc-xlate.pl
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/perlasm/ppc-xlate.pl
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/perlasm/ppc-xlate.pl
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/perlasm/ppc-xlate.pl
 @@ -0,0 +1,320 @@
 +#! /usr/bin/env perl
 +# Copyright 2006-2016 The OpenSSL Project Authors. All Rights Reserved.
@@ -5744,10 +5744,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/perlasm/ppc-xlate
 +___
 +
 +close STDOUT or die "error closing STDOUT: $!";
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/test/abi_test.h
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/test/abi_test.h
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/crypto/test/abi_test.h
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/test/abi_test.h
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/crypto/test/abi_test.h
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/test/abi_test.h
 @@ -179,7 +179,78 @@ struct alignas(16) Reg128 {
    CALLER_STATE_REGISTER(uint64_t, x28)                               \
    CALLER_STATE_REGISTER(uint64_t, x29)
@@ -5854,10 +5854,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/test/abi_test.h
    static_assert(sizeof...(args) <= 8,
                  "too many arguments for abi_test_trampoline");
  
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/test/asm/trampoline-ppc.pl
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/test/asm/trampoline-ppc.pl
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/test/asm/trampoline-ppc.pl
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/test/asm/trampoline-ppc.pl
 @@ -0,0 +1,262 @@
 +#!/usr/bin/env perl
 +# Copyright (c) 2019, Google Inc.
@@ -6121,10 +6121,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/test/asm/trampoli
 +
 +print $code;
 +close STDOUT or die "error closing STDOUT: $!";
-Index: chromium-136.0.7103.48/third_party/boringssl/src/include/openssl/target.h
+Index: chromium-138.0.7204.35/third_party/boringssl/src/include/openssl/target.h
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/include/openssl/target.h
-+++ chromium-136.0.7103.48/third_party/boringssl/src/include/openssl/target.h
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/include/openssl/target.h
++++ chromium-138.0.7204.35/third_party/boringssl/src/include/openssl/target.h
 @@ -34,6 +34,9 @@
  #elif defined(__ARMEL__) || defined(_M_ARM)
  #define OPENSSL_32_BIT
@@ -6135,10 +6135,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/include/openssl/target.h
  #elif defined(__MIPSEL__) && !defined(__LP64__)
  #define OPENSSL_32_BIT
  #define OPENSSL_MIPS
-Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/acvp/modulewrapper/main.cc
+Index: chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/acvp/modulewrapper/main.cc
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/util/fipstools/acvp/modulewrapper/main.cc
-+++ chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/acvp/modulewrapper/main.cc
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/util/fipstools/acvp/modulewrapper/main.cc
++++ chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/acvp/modulewrapper/main.cc
 @@ -35,6 +35,8 @@ int main(int argc, char **argv) {
      puts("ARM (32-bit)");
  #elif defined(OPENSSL_AARCH64)
@@ -6148,10 +6148,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/acvp/modu
  #else
  #error "FIPS build not supported on this architecture"
  #endif
-Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/delocate.go
+Index: chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/delocate.go
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/util/fipstools/delocate/delocate.go
-+++ chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/delocate.go
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/util/fipstools/delocate/delocate.go
++++ chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/delocate.go
 @@ -56,7 +56,8 @@ type stringWriter interface {
  type processorType int
  
@@ -6210,7 +6210,7 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
  	return statement, nil
  }
  
-@@ -701,6 +713,191 @@ func (d *delocation) processAarch64Instr
+@@ -703,6 +715,191 @@ func (d *delocation) processAarch64Instr
  	return statement, nil
  }
  
@@ -6402,7 +6402,7 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
  func (d *delocation) gatherOffsets(symRef *node32, offsets string) (*node32, string) {
  	for symRef != nil && symRef.pegRule == ruleOffset {
  		offset := d.contents(symRef)
-@@ -755,6 +952,215 @@ func (d *delocation) parseMemRef(memRef
+@@ -757,6 +954,215 @@ func (d *delocation) parseMemRef(memRef
  	return
  }
  
@@ -6618,7 +6618,7 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
  /* Intel */
  
  type instructionType int
-@@ -1368,6 +1774,8 @@ func writeAarch64Function(w stringWriter
+@@ -1370,6 +1776,8 @@ func writeAarch64Function(w stringWriter
  func transform(w stringWriter, inputs []inputFile) error {
  	// symbols contains all defined symbols.
  	symbols := make(map[string]struct{})
@@ -6627,7 +6627,7 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
  	// fileNumbers is the set of IDs seen in .file directives.
  	fileNumbers := make(map[int]struct{})
  	// maxObservedFileNumber contains the largest seen file number in a
-@@ -1400,6 +1808,25 @@ func transform(w stringWriter, inputs []
+@@ -1402,6 +1810,25 @@ func transform(w stringWriter, inputs []
  		}, ruleStatement, ruleSymbolDefiningDirective, ruleSymbolDefiningDirectiveName)
  
  		forEachPath(input.ast.up, func(node *node32) {
@@ -6653,7 +6653,7 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
  			assertNodeType(node, ruleLocationDirective)
  			directive := input.contents[node.begin:node.end]
  			if !strings.HasPrefix(directive, ".file") {
-@@ -1447,11 +1874,13 @@ func transform(w stringWriter, inputs []
+@@ -1449,11 +1876,13 @@ func transform(w stringWriter, inputs []
  
  	d := &delocation{
  		symbols:             symbols,
@@ -6667,7 +6667,7 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
  		gotExternalsNeeded:  make(map[string]struct{}),
  		gotOffsetsNeeded:    make(map[string]struct{}),
  		gotOffOffsetsNeeded: make(map[string]struct{}),
-@@ -1486,6 +1915,22 @@ func transform(w stringWriter, inputs []
+@@ -1488,6 +1917,22 @@ func transform(w stringWriter, inputs []
  	for _, name := range redirectorNames {
  		redirector := d.redirectors[name]
  		switch d.processor {
@@ -6690,7 +6690,7 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
  		case aarch64:
  			writeAarch64Function(w, redirector, func(w stringWriter) {
  				w.WriteString("\tb " + name + "\n")
-@@ -1510,6 +1955,13 @@ func transform(w stringWriter, inputs []
+@@ -1512,6 +1957,13 @@ func transform(w stringWriter, inputs []
  		target := d.bssAccessorsNeeded[name]
  
  		switch d.processor {
@@ -6704,7 +6704,7 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
  		case x86_64:
  			w.WriteString(".type " + funcName + ", @function\n")
  			w.WriteString(funcName + ":\n")
-@@ -1525,6 +1977,26 @@ func transform(w stringWriter, inputs []
+@@ -1527,6 +1979,26 @@ func transform(w stringWriter, inputs []
  	}
  
  	switch d.processor {
@@ -6731,7 +6731,7 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
  	case aarch64:
  		externalNames := sortedSet(d.gotExternalsNeeded)
  		for _, symbol := range externalNames {
-@@ -1816,6 +2288,10 @@ func localTargetName(name string) string
+@@ -1818,6 +2290,10 @@ func localTargetName(name string) string
  	return ".L" + name + "_local_target"
  }
  
@@ -6742,7 +6742,7 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
  func isSynthesized(symbol string) bool {
  	return strings.HasSuffix(symbol, "_bss_get") ||
  		strings.HasPrefix(symbol, "BORINGSSL_bcm_text_")
-@@ -1870,6 +2346,8 @@ func detectProcessor(input inputFile) pr
+@@ -1872,6 +2348,8 @@ func detectProcessor(input inputFile) pr
  		switch instructionName {
  		case "movq", "call", "leaq":
  			return x86_64
@@ -6751,10 +6751,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
  		case "str", "bl", "ldr", "st1":
  			return aarch64
  		}
-Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/delocate.peg
+Index: chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/delocate.peg
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/util/fipstools/delocate/delocate.peg
-+++ chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/delocate.peg
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/util/fipstools/delocate/delocate.peg
++++ chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/delocate.peg
 @@ -12,9 +12,9 @@
  # See the License for the specific language governing permissions and
  # limitations under the License.
@@ -6768,10 +6768,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
  
  # To regenerate delocate.peg.go:
  #
-Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/delocate_test.go
+Index: chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/delocate_test.go
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/util/fipstools/delocate/delocate_test.go
-+++ chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/delocate_test.go
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/util/fipstools/delocate/delocate_test.go
++++ chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/delocate_test.go
 @@ -39,6 +39,11 @@ func (test *delocateTest) Path(file stri
  
  var delocateTests = []delocateTest{
@@ -6784,10 +6784,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
  	{"x86_64-Basic", []string{"in.s"}, "out.s"},
  	{"x86_64-BSS", []string{"in.s"}, "out.s"},
  	{"x86_64-GOTRewrite", []string{"in.s"}, "out.s"},
-Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-GlobalEntry/in.s
+Index: chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-GlobalEntry/in.s
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-GlobalEntry/in.s
++++ chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-GlobalEntry/in.s
 @@ -0,0 +1,9 @@
 +	.text
 +foo:
@@ -6798,10 +6798,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
 +	.localentry foo,.-foo
 +.LVL0:
 +	bl
-Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-GlobalEntry/out.s
+Index: chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-GlobalEntry/out.s
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-GlobalEntry/out.s
++++ chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-GlobalEntry/out.s
 @@ -0,0 +1,62 @@
 +.text
 +.file 1 "inserted_by_delocate.cc"
@@ -6865,19 +6865,19 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
 +.byte 0xff
 +.byte 0x31
 +.byte 0x80
-Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-LoadToR0/in.s
+Index: chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-LoadToR0/in.s
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-LoadToR0/in.s
++++ chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-LoadToR0/in.s
 @@ -0,0 +1,4 @@
 +	.text
 +foo:
 +	addis 22,2,bar@toc@ha
 +	ld 0,bar@toc@l(22)
-Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-LoadToR0/out.s
+Index: chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-LoadToR0/out.s
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-LoadToR0/out.s
++++ chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-LoadToR0/out.s
 @@ -0,0 +1,72 @@
 +.text
 +.file 1 "inserted_by_delocate.cc"
@@ -6951,10 +6951,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
 +.byte 0xff
 +.byte 0x31
 +.byte 0x80
-Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-Sample/in.s
+Index: chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-Sample/in.s
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-Sample/in.s
++++ chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-Sample/in.s
 @@ -0,0 +1,161 @@
 +	.file	"foo.cc"
 +	.abiversion 2
@@ -7117,10 +7117,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
 +	.size	exported_function,.-exported_function
 +	.ident	"GCC: (Ubuntu 4.9.2-10ubuntu13) 4.9.2"
 +	.section	.note.GNU-stack,"",@progbits
-Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-Sample/out.s
+Index: chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-Sample/out.s
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-Sample/out.s
++++ chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-Sample/out.s
 @@ -0,0 +1,552 @@
 +.text
 +.file 1 "inserted_by_delocate.cc"
@@ -7674,10 +7674,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
 +.byte 0xff
 +.byte 0x31
 +.byte 0x80
-Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-Sample2/in.s
+Index: chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-Sample2/in.s
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-Sample2/in.s
++++ chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-Sample2/in.s
 @@ -0,0 +1,226 @@
 +	.file	"foo.cc"
 +	.abiversion 2
@@ -7905,10 +7905,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
 +	.zero	20
 +	.ident	"GCC: (Ubuntu 4.9.2-10ubuntu13) 4.9.2"
 +	.section	.note.GNU-stack,"",@progbits
-Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-Sample2/out.s
+Index: chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-Sample2/out.s
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-Sample2/out.s
++++ chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-Sample2/out.s
 @@ -0,0 +1,677 @@
 +.text
 +.file 1 "inserted_by_delocate.cc"
@@ -8587,10 +8587,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
 +.byte 0xff
 +.byte 0x31
 +.byte 0x80
-Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-TOCWithOffset/in.s
+Index: chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-TOCWithOffset/in.s
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-TOCWithOffset/in.s
++++ chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-TOCWithOffset/in.s
 @@ -0,0 +1,23 @@
 +	.text
 +foo:
@@ -8615,10 +8615,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
 +
 +	addis 4, 2, 1+foo-2@toc@ha+3
 +	addi 4, 4, 1+foo-2@toc@l+3
-Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-TOCWithOffset/out.s
+Index: chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-TOCWithOffset/out.s
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-TOCWithOffset/out.s
++++ chromium-138.0.7204.35/third_party/boringssl/src/util/fipstools/delocate/testdata/ppc64le-TOCWithOffset/out.s
 @@ -0,0 +1,178 @@
 +.text
 +.file 1 "inserted_by_delocate.cc"
@@ -8798,11 +8798,11 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/fipstools/delocate/
 +.byte 0xff
 +.byte 0x31
 +.byte 0x80
-Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/sha/sha1.cc.inc
+Index: chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/sha/sha1.cc.inc
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/crypto/fipsmodule/sha/sha1.cc.inc
-+++ chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/sha/sha1.cc.inc
-@@ -367,6 +367,10 @@ static void sha1_block_data_order(uint32
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/crypto/fipsmodule/sha/sha1.cc.inc
++++ chromium-138.0.7204.35/third_party/boringssl/src/crypto/fipsmodule/sha/sha1.cc.inc
+@@ -377,6 +377,10 @@ static void sha1_block_data_order(uint32
      return;
    }
  #endif
@@ -8813,10 +8813,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/crypto/fipsmodule/sha/sh
    sha1_block_data_order_nohw(state, data, num);
  }
  
-Index: chromium-136.0.7103.48/third_party/boringssl/src/build.json
+Index: chromium-138.0.7204.35/third_party/boringssl/src/build.json
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/build.json
-+++ chromium-136.0.7103.48/third_party/boringssl/src/build.json
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/build.json
++++ chromium-138.0.7204.35/third_party/boringssl/src/build.json
 @@ -128,6 +128,10 @@
              {"src": "crypto/fipsmodule/sha/asm/sha512-armv4.pl"},
              {"src": "crypto/fipsmodule/aes/asm/vpaes-armv7.pl"}
@@ -8828,7 +8828,7 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/build.json
          "perlasm_x86": [
              {"src": "crypto/fipsmodule/aes/asm/aesni-x86.pl"},
              {"src": "crypto/fipsmodule/bn/asm/bn-586.pl"},
-@@ -234,6 +238,7 @@
+@@ -237,6 +241,7 @@
              "crypto/cpu_arm_freebsd.cc",
              "crypto/cpu_arm_linux.cc",
              "crypto/cpu_intel.cc",
@@ -8836,7 +8836,7 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/build.json
              "crypto/crypto.cc",
              "crypto/curve25519/curve25519.cc",
              "crypto/curve25519/curve25519_64_adx.cc",
-@@ -791,6 +796,9 @@
+@@ -798,6 +803,9 @@
          "perlasm_arm": [
              {"src": "crypto/test/asm/trampoline-armv4.pl"}
          ],
@@ -8846,10 +8846,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/build.json
          "perlasm_x86": [
              {"src": "crypto/test/asm/trampoline-x86.pl"}
          ],
-Index: chromium-136.0.7103.48/third_party/boringssl/src/util/pregenerate/build.go
+Index: chromium-138.0.7204.35/third_party/boringssl/src/util/pregenerate/build.go
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/util/pregenerate/build.go
-+++ chromium-136.0.7103.48/third_party/boringssl/src/util/pregenerate/build.go
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/util/pregenerate/build.go
++++ chromium-138.0.7204.35/third_party/boringssl/src/util/pregenerate/build.go
 @@ -38,6 +38,7 @@ type InputTarget struct {
  	// architecture.
  	PerlasmAarch64 []PerlasmSource `json:"perlasm_aarch64,omitempty"`
@@ -8868,10 +8868,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/util/pregenerate/build.g
  	for _, p := range in.PerlasmX86 {
  		addPerlasmTask(&out.Asm, &p, "-apple.S", []string{"macosx", "-fPIC"})
  		addPerlasmTask(&out.Asm, &p, "-linux.S", []string{"elf", "-fPIC"})
-Index: chromium-136.0.7103.48/third_party/boringssl/README.ppc64le
+Index: chromium-138.0.7204.35/third_party/boringssl/README.ppc64le
 ===================================================================
 --- /dev/null
-+++ chromium-136.0.7103.48/third_party/boringssl/README.ppc64le
++++ chromium-138.0.7204.35/third_party/boringssl/README.ppc64le
 @@ -0,0 +1,8 @@
 +==============================================================
 +To recreate boringssl pregenerated files patch for ppc64le:
@@ -8881,10 +8881,10 @@ Index: chromium-136.0.7103.48/third_party/boringssl/README.ppc64le
 +go run ./util/pregenerate
 +cd ../../../../
 +diff -urN chromium-*/third_party/boringssl/src/gen.orig chromium-*/third_party/boringssl/src/gen
-Index: chromium-136.0.7103.48/third_party/boringssl/src/gen/sources.gni
+Index: chromium-138.0.7204.35/third_party/boringssl/src/gen/sources.gni
 ===================================================================
---- chromium-136.0.7103.48.orig/third_party/boringssl/src/gen/sources.gni
-+++ chromium-136.0.7103.48/third_party/boringssl/src/gen/sources.gni
+--- chromium-138.0.7204.35.orig/third_party/boringssl/src/gen/sources.gni
++++ chromium-138.0.7204.35/third_party/boringssl/src/gen/sources.gni
 @@ -121,6 +121,7 @@ bcm_sources_asm = [
    "gen/bcm/aesv8-gcm-armv8-apple.S",
    "gen/bcm/aesv8-gcm-armv8-linux.S",
@@ -8901,7 +8901,7 @@ Index: chromium-136.0.7103.48/third_party/boringssl/src/gen/sources.gni
    "gen/bcm/ghash-ssse3-x86-apple.S",
    "gen/bcm/ghash-ssse3-x86-linux.S",
    "gen/bcm/ghash-ssse3-x86_64-apple.S",
-@@ -333,6 +335,7 @@ crypto_sources = [
+@@ -336,6 +338,7 @@ crypto_sources = [
    "crypto/cpu_arm_freebsd.cc",
    "crypto/cpu_arm_linux.cc",
    "crypto/cpu_intel.cc",

diff --git a/0002-regenerate-xnn-buildgn.patch b/0002-regenerate-xnn-buildgn.patch
index 78977f4..dd44958 100644
--- a/0002-regenerate-xnn-buildgn.patch
+++ b/0002-regenerate-xnn-buildgn.patch
@@ -1,8 +1,10 @@
 File regenerated by running 'python3 generate_build_gn.py'
 
---- a/third_party/xnnpack/BUILD.gn
-+++ b/third_party/xnnpack/BUILD.gn
-@@ -25,7 +25,7 @@
+Index: chromium-138.0.7204.35/third_party/xnnpack/BUILD.gn
+===================================================================
+--- chromium-138.0.7204.35.orig/third_party/xnnpack/BUILD.gn
++++ chromium-138.0.7204.35/third_party/xnnpack/BUILD.gn
+@@ -25,7 +25,7 @@ config("xnnpack_config") {
      "src",
    ]
  
@@ -11,7 +13,7 @@ File regenerated by running 'python3 generate_build_gn.py'
      "-Wno-unused-function",
      "-Wno-deprecated-comma-subscript",
    ]
-@@ -54,11 +54,15 @@
+@@ -54,11 +54,15 @@ config("xnnpack_config") {
    }
  
    if (current_cpu == "x86" || current_cpu == "x64") {
@@ -28,7 +30,7 @@ File regenerated by running 'python3 generate_build_gn.py'
    xnnpack_deps = [
      ":configs_x64",
      ":enums_x64",
-@@ -545,7 +549,7 @@
+@@ -560,7 +564,7 @@ if (current_cpu == "x64" || current_cpu
      ":xx-fill_x64",
      ":xx-pad_sse2-no-sse3",
      ":xx-pad_x64",
@@ -37,7 +39,7 @@ File regenerated by running 'python3 generate_build_gn.py'
    ]
  
    xnnpack_standalone_deps = [
-@@ -572,8 +576,8 @@
+@@ -587,8 +591,8 @@ if (current_cpu == "x64" || current_cpu
      ":f16-qu8-vcvt_x64_standalone",
      ":f16-raddstoreexpminusmax_f16c-fma-avx2_standalone",
      ":f16-rdminmax_x64_standalone",
@@ -47,7 +49,7 @@ File regenerated by running 'python3 generate_build_gn.py'
      ":f16-rminmax_f16c-no-avx2-no-fma_standalone",
      ":f16-rminmax_x64_standalone",
      ":f16-vapproxgelu_x64_standalone",
-@@ -776,47 +780,47 @@
+@@ -794,47 +798,47 @@ if (current_cpu == "x64" || current_cpu
      ":qd8-f16-qb4w-gemm_f16c-fma-avx2_standalone",
      ":qd8-f16-qc4w-gemm_avx2-avxvnni-f16c-fma_standalone",
      ":qd8-f16-qc4w-gemm_f16c-fma-avx2_standalone",
@@ -108,7 +110,7 @@ File regenerated by running 'python3 generate_build_gn.py'
      ":qd8-f32-qc8w-igemm_sse2-no-sse3_standalone",
      ":qd8-f32-qc8w-igemm_sse4.1-no-sse4.2_standalone",
      ":qd8-f32-qc8w-igemm_x64_standalone",
-@@ -847,9 +851,9 @@
+@@ -865,9 +869,9 @@ if (current_cpu == "x64" || current_cpu
      ":qs8-qc8w-gemm_avx2-avxvnni-f16c-fma_standalone",
      ":qs8-qc8w-gemm_avx2-avxvnniint8-f16c-fma_standalone",
      ":qs8-qc8w-gemm_f16c-fma-avx2_standalone",
@@ -120,7 +122,7 @@ File regenerated by running 'python3 generate_build_gn.py'
      ":qs8-qc8w-gemm_sse2-no-sse3_standalone",
      ":qs8-qc8w-gemm_sse4.1-no-sse4.2_standalone",
      ":qs8-qc8w-gemm_x64_standalone",
-@@ -857,9 +861,9 @@
+@@ -875,9 +879,9 @@ if (current_cpu == "x64" || current_cpu
      ":qs8-qc8w-igemm_avx2-avxvnni-f16c-fma_standalone",
      ":qs8-qc8w-igemm_avx2-avxvnniint8-f16c-fma_standalone",
      ":qs8-qc8w-igemm_f16c-fma-avx2_standalone",
@@ -132,7 +134,7 @@ File regenerated by running 'python3 generate_build_gn.py'
      ":qs8-qc8w-igemm_sse2-no-sse3_standalone",
      ":qs8-qc8w-igemm_sse4.1-no-sse4.2_standalone",
      ":qs8-qc8w-igemm_x64_standalone",
-@@ -872,8 +876,8 @@
+@@ -890,8 +894,8 @@ if (current_cpu == "x64" || current_cpu
      ":qs8-rdsum_x64_standalone",
      ":qs8-rsum_avx2-avxvnni-f16c-fma_standalone",
      ":qs8-rsum_f16c-fma-avx2_standalone",
@@ -142,7 +144,7 @@ File regenerated by running 'python3 generate_build_gn.py'
      ":qs8-rsum_ssse3-no-sse4.1_standalone",
      ":qs8-rsum_x64_standalone",
      ":qs8-vadd_avx-no-avx2-no-f16c-no-fma_standalone",
-@@ -1021,8 +1025,8 @@
+@@ -1051,8 +1055,8 @@ if (current_cpu == "x64" || current_cpu
      ":x64-transposec_x64_standalone",
      ":x8-lut_avx-no-avx2-no-f16c-no-fma_standalone",
      ":x8-lut_f16c-fma-avx2_standalone",
@@ -152,7 +154,7 @@ File regenerated by running 'python3 generate_build_gn.py'
      ":x8-lut_x64_standalone",
      ":x8-packq_x64_standalone",
      ":x8-packw_x64_standalone",
-@@ -1034,9 +1038,11 @@
+@@ -1064,9 +1068,11 @@ if (current_cpu == "x64" || current_cpu
      ":xx-fill_x64_standalone",
      ":xx-pad_sse2-no-sse3_standalone",
      ":xx-pad_x64_standalone",
@@ -166,7 +168,7 @@ File regenerated by running 'python3 generate_build_gn.py'
    xnnpack_deps = [
      ":configs_arm64",
      ":enums_arm64",
-@@ -1235,7 +1241,7 @@
+@@ -1278,7 +1284,7 @@ if (current_cpu == "x64" || current_cpu
      ":xx-copy_arm64",
      ":xx-fill_arm64",
      ":xx-pad_arm64",
@@ -175,7 +177,7 @@ File regenerated by running 'python3 generate_build_gn.py'
    ]
  
    xnnpack_standalone_deps = [
-@@ -1248,12 +1254,12 @@
+@@ -1291,12 +1297,12 @@ if (current_cpu == "x64" || current_cpu
      ":f16-f32-vcvt_arm64_standalone",
      ":f16-f32acc-rdsum_arch=armv8.2-a+fp16_standalone",
      ":f16-f32acc-rsum_arch=armv8.2-a+fp16_standalone",
@@ -190,7 +192,7 @@ File regenerated by running 'python3 generate_build_gn.py'
      ":f16-maxpool_arch=armv8.2-a+fp16_standalone",
      ":f16-qs8-vcvt_arch=armv8.2-a+fp16_standalone",
      ":f16-qs8-vcvt_arm64_standalone",
-@@ -1341,12 +1347,12 @@
+@@ -1387,12 +1393,12 @@ if (current_cpu == "x64" || current_cpu
      ":qd8-f16-qc4w-gemm_arch=armv8.2-a+fp16_standalone",
      ":qd8-f16-qc4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone",
      ":qd8-f16-qc8w-gemm_arch=armv8.2-a+dotprod+fp16_standalone",
@@ -205,7 +207,7 @@ File regenerated by running 'python3 generate_build_gn.py'
      ":qd8-f16-qc8w-igemm_arch=armv8.2-a+i8mm+fp16_standalone",
      ":qd8-f32-qb4w-gemm_arch=armv8.2-a+dotprod_standalone",
      ":qd8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone",
-@@ -1436,7 +1442,247 @@
+@@ -1492,7 +1498,259 @@ if (current_cpu == "x64" || current_cpu
      ":xx-copy_arm64_standalone",
      ":xx-fill_arm64_standalone",
      ":xx-pad_arm64_standalone",
@@ -291,6 +293,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    ":qs8-vlrelu_ppc64",
 +    ":qs8-vmul_ppc64",
 +    ":qs8-vmulc_ppc64",
++    ":qs8-vprelu_ppc64",
++    ":qs8-vpreluc_ppc64",
++    ":qs8-vrpreluc_ppc64",
 +    ":qu8-dwconv_ppc64",
 +    ":qu8-f32-vcvt_ppc64",
 +    ":qu8-gemm_ppc64",
@@ -303,6 +308,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    ":qu8-vlrelu_ppc64",
 +    ":qu8-vmul_ppc64",
 +    ":qu8-vmulc_ppc64",
++    ":qu8-vprelu_ppc64",
++    ":qu8-vpreluc_ppc64",
++    ":qu8-vrpreluc_ppc64",
 +    ":reference_ppc64",
 +    ":s8-ibilinear_ppc64",
 +    ":s8-maxpool_ppc64",
@@ -410,6 +418,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    ":qs8-vlrelu_ppc64_standalone",
 +    ":qs8-vmul_ppc64_standalone",
 +    ":qs8-vmulc_ppc64_standalone",
++    ":qs8-vprelu_ppc64_standalone",
++    ":qs8-vpreluc_ppc64_standalone",
++    ":qs8-vrpreluc_ppc64_standalone",
 +    ":qu8-dwconv_ppc64_standalone",
 +    ":qu8-f32-vcvt_ppc64_standalone",
 +    ":qu8-gemm_ppc64_standalone",
@@ -422,6 +433,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    ":qu8-vlrelu_ppc64_standalone",
 +    ":qu8-vmul_ppc64_standalone",
 +    ":qu8-vmulc_ppc64_standalone",
++    ":qu8-vprelu_ppc64_standalone",
++    ":qu8-vpreluc_ppc64_standalone",
++    ":qu8-vrpreluc_ppc64_standalone",
 +    ":reference_ppc64_standalone",
 +    ":s8-ibilinear_ppc64_standalone",
 +    ":s8-maxpool_ppc64_standalone",
@@ -454,7 +468,7 @@ File regenerated by running 'python3 generate_build_gn.py'
    ]
  } else {
    xnnpack_deps = []
-@@ -1451,8 +1697,8 @@
+@@ -1507,8 +1765,8 @@ source_set("xnnpack") {
    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
    sources = [
@@ -465,7 +479,7 @@ File regenerated by running 'python3 generate_build_gn.py'
      "src/src/allocator.c",
      "src/src/cache.c",
      "src/src/datatype.c",
-@@ -1472,15 +1718,15 @@
+@@ -1528,15 +1786,15 @@ source_set("xnnpack") {
      "src/src/runtime.c",
      "src/src/sanitizers.c",
      "src/src/subgraph.c",
@@ -487,7 +501,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
    public_configs = [ ":xnnpack_config" ]
  }
-@@ -1494,8 +1740,8 @@
+@@ -1550,8 +1808,8 @@ source_set("xnnpack_standalone") {
    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
    sources = [
@@ -498,7 +512,7 @@ File regenerated by running 'python3 generate_build_gn.py'
      "src/src/allocator.c",
      "src/src/cache.c",
      "src/src/datatype.c",
-@@ -1515,15 +1761,15 @@
+@@ -1571,15 +1829,15 @@ source_set("xnnpack_standalone") {
      "src/src/runtime.c",
      "src/src/sanitizers.c",
      "src/src/subgraph.c",
@@ -520,7 +534,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
    public_configs = [ ":xnnpack_config" ]
  
-@@ -1533,39160 +1779,47034 @@
+@@ -1589,40466 +1847,48672 @@ source_set("xnnpack_standalone") {
  }
  
  if (current_cpu == "x64" || current_cpu == "x86") {
@@ -1881,7 +1895,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-f32acc-rsum/gen/f16-f32acc-rsum-avx512skx-u64-acc4.c",
+-      "src/src/f16-f32acc-rsum/gen/f16-f32acc-rsum-avx512skx-u32-acc2.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -1927,7 +1941,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-f32acc-rsum/gen/f16-f32acc-rsum-avx512skx-u64-acc4.c",
+-      "src/src/f16-f32acc-rsum/gen/f16-f32acc-rsum-avx512skx-u32-acc2.c",
 -    ]
 +source_set("f16-f32acc-gemm_f16c-fma-avx2") {
 +  cflags = [
@@ -2421,7 +2435,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-f32acc-rsum/gen/f16-f32acc-rsum-avx512skx-u64-acc4.c"
++    "src/src/f16-f32acc-rsum/gen/f16-f32acc-rsum-avx512skx-u32-acc2.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -2473,7 +2487,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-f32acc-rsum/gen/f16-f32acc-rsum-avx512skx-u64-acc4.c"
++    "src/src/f16-f32acc-rsum/gen/f16-f32acc-rsum-avx512skx-u32-acc2.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -3593,12 +3607,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mf16c",
 -      "-mfma",
 -    ]
--
++source_set("f16-rminmax_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
++  cflags = [
++    "-mavx512bw",
++    "-mavx512cd",
++    "-mavx512dq",
++    "-mavx512f",
++    "-mavx512vl",
++    "-mf16c",
++    "-mfma"
++  ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f16-velu/gen/f16-velu-avx2-rr1-p3-u16.c",
 -    ]
--
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-rminmax/gen/f16-rmax-avx512skx-u64-acc4.c",
++    "src/src/f16-rminmax/gen/f16-rmin-avx512skx-u64-acc4.c",
++    "src/src/f16-rminmax/gen/f16-rminmax-avx512skx-u64-acc4.c"
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -3609,7 +3639,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("f16-rminmax_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f16-velu_f16c-fma-avx2_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f16-velu/gen/f16-velu-avx2-rr1-p3-u16.c",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("f16-rminmax_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 +  cflags = [
 +    "-mavx512bw",
 +    "-mavx512cd",
@@ -3620,8 +3678,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "-mfma"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/f16-rminmax/gen/f16-rmax-avx512skx-u64-acc4.c",
@@ -3629,20 +3695,78 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/f16-rminmax/gen/f16-rminmax-avx512skx-u64-acc4.c"
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("f16-velu_f16c-fma-avx2_standalone") {
--    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-    public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-  source_set("f16-vexp_x64") {
+-    cflags = []
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f16-vexp/gen/f16-vexp-scalar-poly-3.c",
+-    ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
 -    ]
++source_set("f16-rminmax_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-avx512fp16") {
++  cflags = [
++    "-mavx512bw",
++    "-mavx512cd",
++    "-mavx512dq",
++    "-mavx512f",
++    "-mavx512fp16",
++    "-mavx512vl",
++    "-mavx512vnni",
++    "-mf16c",
++    "-mfma",
++    "-mgfni"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-rminmax/gen/f16-rmax-avx512fp16-u128-acc4.c",
++    "src/src/f16-rminmax/gen/f16-rmin-avx512fp16-u128-acc4.c",
++    "src/src/f16-rminmax/gen/f16-rminmax-avx512fp16-u128-acc4.c"
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f16-vexp_x64_standalone") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-velu/gen/f16-velu-avx2-rr1-p3-u16.c",
+-      "src/src/f16-vexp/gen/f16-vexp-scalar-poly-3.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -3666,15 +3790,18 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-rminmax_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
++source_set("f16-rminmax_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-avx512fp16_standalone") {
 +  cflags = [
 +    "-mavx512bw",
 +    "-mavx512cd",
 +    "-mavx512dq",
 +    "-mavx512f",
++    "-mavx512fp16",
 +    "-mavx512vl",
++    "-mavx512vnni",
 +    "-mf16c",
-+    "-mfma"
++    "-mfma",
++    "-mgfni"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -3683,12 +3810,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-rminmax/gen/f16-rmax-avx512skx-u64-acc4.c",
-+    "src/src/f16-rminmax/gen/f16-rmin-avx512skx-u64-acc4.c",
-+    "src/src/f16-rminmax/gen/f16-rminmax-avx512skx-u64-acc4.c"
++    "src/src/f16-rminmax/gen/f16-rmax-avx512fp16-u128-acc4.c",
++    "src/src/f16-rminmax/gen/f16-rmin-avx512fp16-u128-acc4.c",
++    "src/src/f16-rminmax/gen/f16-rminmax-avx512fp16-u128-acc4.c"
 +  ]
  
--  source_set("f16-vexp_x64") {
+-  source_set("f16-vgelu_x64") {
 -    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -3696,7 +3823,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vexp/gen/f16-vexp-scalar-poly-3.c",
+-      "src/src/f16-vgelu/gen/f16-vgelu-scalar-rational-6-4-div.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -3724,25 +3851,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f16-vexp_x64_standalone") {
+-  source_set("f16-vgelu_x64_standalone") {
 -    cflags = []
--
++source_set("f16-rminmax_f16c-no-avx2-no-fma") {
++  cflags = [
++    "-mf16c",
++    "-mno-avx2",
++    "-mno-fma"
++  ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vexp/gen/f16-vexp-scalar-poly-3.c",
+-      "src/src/f16-vgelu/gen/f16-vgelu-scalar-rational-6-4-div.c",
 -    ]
-+source_set("f16-rminmax_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-avx512fp16") {
-+  cflags = [
-+    "-mavx512bw",
-+    "-mavx512cd",
-+    "-mavx512dq",
-+    "-mavx512f",
-+    "-mavx512fp16",
-+    "-mavx512vl",
-+    "-mavx512vnni",
-+    "-mf16c",
-+    "-mfma",
-+    "-mgfni"
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-rminmax/f16-rmax-f16c-u32.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -3755,22 +3879,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-rminmax/gen/f16-rmax-avx512fp16-u128-acc4.c",
-+    "src/src/f16-rminmax/gen/f16-rmin-avx512fp16-u128-acc4.c",
-+    "src/src/f16-rminmax/gen/f16-rminmax-avx512fp16-u128-acc4.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -3778,28 +3891,34 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  source_set("f16-vgelu_x64") {
--    cflags = []
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f16-vgelu/gen/f16-vgelu-scalar-rational-6-4-div.c",
+-  source_set("f16-vhswish_f16c-no-avx2-no-fma") {
+-    cflags = [
+-      "-mf16c",
+-      "-mno-avx2",
+-      "-mno-fma",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-rminmax_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-avx512fp16_standalone") {
++source_set("f16-rminmax_f16c-no-avx2-no-fma_standalone") {
 +  cflags = [
-+    "-mavx512bw",
-+    "-mavx512cd",
-+    "-mavx512dq",
-+    "-mavx512f",
-+    "-mavx512fp16",
-+    "-mavx512vl",
-+    "-mavx512vnni",
 +    "-mf16c",
-+    "-mfma",
-+    "-mgfni"
++    "-mno-avx2",
++    "-mno-fma"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f16-vhswish/gen/f16-vhswish-f16c-u16.c",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-rminmax/f16-rmax-f16c-u32.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -3812,22 +3931,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-rminmax/gen/f16-rmax-avx512fp16-u128-acc4.c",
-+    "src/src/f16-rminmax/gen/f16-rmin-avx512fp16-u128-acc4.c",
-+    "src/src/f16-rminmax/gen/f16-rminmax-avx512fp16-u128-acc4.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("f16-vgelu_x64_standalone") {
--    cflags = []
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -3835,11 +3944,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-  # This is a target that cannot depend on //base.
+-  source_set("f16-vhswish_f16c-no-avx2-no-fma_standalone") {
+-    cflags = [
+-      "-mf16c",
+-      "-mno-avx2",
+-      "-mno-fma",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vgelu/gen/f16-vgelu-scalar-rational-6-4-div.c",
+-      "src/src/f16-vhswish/gen/f16-vhswish-f16c-u16.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -3851,17 +3972,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++source_set("f16-rminmax_x64") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("f16-rminmax_f16c-no-avx2-no-fma") {
-+  cflags = [
-+    "-mf16c",
-+    "-mno-avx2",
-+    "-mno-fma"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -3870,10 +3984,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-rminmax/f16-rmax-f16c-u32.c"
++    "src/src/f16-rminmax/gen/f16-rmax-scalar-u2-acc2.c",
++    "src/src/f16-rminmax/gen/f16-rmin-scalar-u2-acc2.c",
++    "src/src/f16-rminmax/gen/f16-rminmax-scalar-u2-acc2.c"
 +  ]
  
--  source_set("f16-vhswish_f16c-no-avx2-no-fma") {
+-  source_set("f16-vlrelu_f16c-no-avx2-no-fma") {
 -    cflags = [
 -      "-mf16c",
 -      "-mno-avx2",
@@ -3885,7 +4001,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vhswish/gen/f16-vhswish-f16c-u16.c",
+-      "src/src/f16-vlrelu/gen/f16-vlrelu-f16c-u16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -3910,32 +4026,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("f16-rminmax_f16c-no-avx2-no-fma_standalone") {
++source_set("f16-rminmax_x64_standalone") {
 +  cflags = [
-+    "-mf16c",
-+    "-mno-avx2",
-+    "-mno-fma"
-+  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f16-vhswish_f16c-no-avx2-no-fma_standalone") {
+-  source_set("f16-vlrelu_f16c-no-avx2-no-fma_standalone") {
 -    cflags = [
 -      "-mf16c",
 -      "-mno-avx2",
 -      "-mno-fma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-rminmax/f16-rmax-f16c-u32.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vhswish/gen/f16-vhswish-f16c-u16.c",
+-      "src/src/f16-vlrelu/gen/f16-vlrelu-f16c-u16.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-rminmax/gen/f16-rmax-scalar-u2-acc2.c",
++    "src/src/f16-rminmax/gen/f16-rmin-scalar-u2-acc2.c",
++    "src/src/f16-rminmax/gen/f16-rminmax-scalar-u2-acc2.c"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -3947,6 +4059,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -3965,18 +4081,18 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("f16-vlrelu_f16c-no-avx2-no-fma") {
+-  source_set("f16-vmulcaddc_f16c-fma-no-avx2") {
 -    cflags = [
 -      "-mf16c",
+-      "-mfma",
 -      "-mno-avx2",
--      "-mno-fma",
 -    ]
-+source_set("f16-rminmax_x64") {
++source_set("f16-vapproxgelu_x64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vlrelu/gen/f16-vlrelu-f16c-u16.c",
+-      "src/src/f16-vmulcaddc/gen/f16-vmulcaddc-c8-minmax-fma3-2x.c",
 -    ]
 +  ]
  
@@ -3992,9 +4108,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-rminmax/gen/f16-rmax-scalar-u2-acc2.c",
-+    "src/src/f16-rminmax/gen/f16-rmin-scalar-u2-acc2.c",
-+    "src/src/f16-rminmax/gen/f16-rminmax-scalar-u2-acc2.c"
++    "src/src/f16-vapproxgelu/gen/f16-vapproxgelu-scalar-rational-6-4-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -4004,11 +4118,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f16-vlrelu_f16c-no-avx2-no-fma_standalone") {
+-  source_set("f16-vmulcaddc_f16c-fma-no-avx2_standalone") {
 -    cflags = [
 -      "-mf16c",
+-      "-mfma",
 -      "-mno-avx2",
--      "-mno-fma",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -4019,7 +4133,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vlrelu/gen/f16-vlrelu-f16c-u16.c",
+-      "src/src/f16-vmulcaddc/gen/f16-vmulcaddc-c8-minmax-fma3-2x.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -4035,7 +4149,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-rminmax_x64_standalone") {
++source_set("f16-vapproxgelu_x64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -4047,16 +4161,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-rminmax/gen/f16-rmax-scalar-u2-acc2.c",
-+    "src/src/f16-rminmax/gen/f16-rmin-scalar-u2-acc2.c",
-+    "src/src/f16-rminmax/gen/f16-rminmax-scalar-u2-acc2.c"
++    "src/src/f16-vapproxgelu/gen/f16-vapproxgelu-scalar-rational-6-4-div.c"
 +  ]
  
--  source_set("f16-vmulcaddc_f16c-fma-no-avx2") {
+-  source_set("f16-vrnd_f16c-no-avx2-no-fma") {
 -    cflags = [
 -      "-mf16c",
--      "-mfma",
 -      "-mno-avx2",
+-      "-mno-fma",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -4064,7 +4176,10 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vmulcaddc/gen/f16-vmulcaddc-c8-minmax-fma3-2x.c",
+-      "src/src/f16-vrnd/gen/f16-vrndd-f16c-u16.c",
+-      "src/src/f16-vrnd/gen/f16-vrndne-f16c-u16.c",
+-      "src/src/f16-vrnd/gen/f16-vrndu-f16c-u16.c",
+-      "src/src/f16-vrnd/gen/f16-vrndz-f16c-u16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -4092,19 +4207,33 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f16-vmulcaddc_f16c-fma-no-avx2_standalone") {
+-  source_set("f16-vrnd_f16c-no-avx2-no-fma_standalone") {
 -    cflags = [
 -      "-mf16c",
--      "-mfma",
 -      "-mno-avx2",
+-      "-mno-fma",
 -    ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vmulcaddc/gen/f16-vmulcaddc-c8-minmax-fma3-2x.c",
+-      "src/src/f16-vrnd/gen/f16-vrndd-f16c-u16.c",
+-      "src/src/f16-vrnd/gen/f16-vrndne-f16c-u16.c",
+-      "src/src/f16-vrnd/gen/f16-vrndu-f16c-u16.c",
+-      "src/src/f16-vrnd/gen/f16-vrndz-f16c-u16.c",
 -    ]
-+source_set("f16-vapproxgelu_x64") {
++source_set("f16-vbinary_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-avx512fp16") {
 +  cflags = [
++    "-mavx512bw",
++    "-mavx512cd",
++    "-mavx512dq",
++    "-mavx512f",
++    "-mavx512fp16",
++    "-mavx512vl",
++    "-mavx512vnni",
++    "-mf16c",
++    "-mfma",
++    "-mgfni"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -4116,28 +4245,38 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vapproxgelu/gen/f16-vapproxgelu-scalar-rational-6-4-div.c"
++    "src/src/f16-vbinary/gen/f16-vadd-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vaddc-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vdiv-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vdivc-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vmax-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vmaxc-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vmin-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vminc-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vmul-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vmulc-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vprelu-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vpreluc-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vrdivc-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vrpreluc-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vrsubc-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vsqrdiff-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vsqrdiffc-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vsub-avx512fp16-u64.c",
++    "src/src/f16-vbinary/gen/f16-vsubc-avx512fp16-u64.c"
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("f16-vrnd_f16c-no-avx2-no-fma") {
--    cflags = [
--      "-mf16c",
--      "-mno-avx2",
--      "-mno-fma",
--    ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -4145,108 +4284,21 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f16-vrnd/gen/f16-vrndd-f16c-u16.c",
--      "src/src/f16-vrnd/gen/f16-vrndne-f16c-u16.c",
--      "src/src/f16-vrnd/gen/f16-vrndu-f16c-u16.c",
--      "src/src/f16-vrnd/gen/f16-vrndz-f16c-u16.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-vapproxgelu_x64_standalone") {
-+  cflags = [
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
-+  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("f16-vrnd_f16c-no-avx2-no-fma_standalone") {
+-  source_set("f16-vrsqrt_f16c-no-avx2-no-fma") {
 -    cflags = [
 -      "-mf16c",
 -      "-mno-avx2",
 -      "-mno-fma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vapproxgelu/gen/f16-vapproxgelu-scalar-rational-6-4-div.c"
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f16-vrnd/gen/f16-vrndd-f16c-u16.c",
--      "src/src/f16-vrnd/gen/f16-vrndne-f16c-u16.c",
--      "src/src/f16-vrnd/gen/f16-vrndu-f16c-u16.c",
--      "src/src/f16-vrnd/gen/f16-vrndz-f16c-u16.c",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
 +}
  
--  source_set("f16-vrsqrt_f16c-no-avx2-no-fma") {
--    cflags = [
--      "-mf16c",
--      "-mno-avx2",
--      "-mno-fma",
--    ]
--
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f16-vrsqrt/gen/f16-vrsqrt-f16c-rsqrt-u32.c",
 -    ]
--
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+source_set("f16-vbinary_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-avx512fp16") {
++# This is a target that cannot depend on //base.
++source_set("f16-vbinary_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-avx512fp16_standalone") {
 +  cflags = [
 +    "-mavx512bw",
 +    "-mavx512cd",
@@ -4260,8 +4312,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "-mgfni"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/f16-vbinary/gen/f16-vadd-avx512fp16-u64.c",
@@ -4285,6 +4345,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/f16-vbinary/gen/f16-vsubc-avx512fp16-u64.c"
 +  ]
  
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
 -  # This is a target that cannot depend on //base.
 -  source_set("f16-vrsqrt_f16c-no-avx2-no-fma_standalone") {
 -    cflags = [
@@ -4292,21 +4358,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-avx2",
 -      "-mno-fma",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f16-vrsqrt/gen/f16-vrsqrt-f16c-rsqrt-u32.c",
--    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f16-vrsqrt/gen/f16-vrsqrt-f16c-rsqrt-u32.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -4317,23 +4381,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
 +}
  
 -    public_configs = [ ":xnnpack_config" ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-vbinary_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-avx512fp16_standalone") {
++source_set("f16-vbinary_f16c-no-avx2-no-fma") {
 +  cflags = [
-+    "-mavx512bw",
-+    "-mavx512cd",
-+    "-mavx512dq",
-+    "-mavx512f",
-+    "-mavx512fp16",
-+    "-mavx512vl",
-+    "-mavx512vnni",
 +    "-mf16c",
-+    "-mfma",
-+    "-mgfni"
++    "-mno-avx2",
++    "-mno-fma"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -4342,25 +4400,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vbinary/gen/f16-vadd-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vaddc-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vdiv-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vdivc-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vmax-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vmaxc-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vmin-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vminc-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vmul-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vmulc-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vprelu-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vpreluc-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vrdivc-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vrpreluc-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vrsubc-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vsqrdiff-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vsqrdiffc-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vsub-avx512fp16-u64.c",
-+    "src/src/f16-vbinary/gen/f16-vsubc-avx512fp16-u64.c"
++    "src/src/f16-vbinary/gen/f16-vadd-f16c-u16.c",
++    "src/src/f16-vbinary/gen/f16-vaddc-f16c-u16.c",
++    "src/src/f16-vbinary/gen/f16-vdiv-f16c-u8.c",
++    "src/src/f16-vbinary/gen/f16-vdivc-f16c-u8.c",
++    "src/src/f16-vbinary/gen/f16-vmax-f16c-u16.c",
++    "src/src/f16-vbinary/gen/f16-vmaxc-f16c-u16.c",
++    "src/src/f16-vbinary/gen/f16-vmin-f16c-u16.c",
++    "src/src/f16-vbinary/gen/f16-vminc-f16c-u16.c",
++    "src/src/f16-vbinary/gen/f16-vmul-f16c-u16.c",
++    "src/src/f16-vbinary/gen/f16-vmulc-f16c-u16.c",
++    "src/src/f16-vbinary/gen/f16-vprelu-f16c-u16.c",
++    "src/src/f16-vbinary/gen/f16-vpreluc-f16c-u16.c",
++    "src/src/f16-vbinary/gen/f16-vrdivc-f16c-u8.c",
++    "src/src/f16-vbinary/gen/f16-vrpreluc-f16c-u16.c",
++    "src/src/f16-vbinary/gen/f16-vrsubc-f16c-u16.c",
++    "src/src/f16-vbinary/gen/f16-vsqrdiff-f16c-u16.c",
++    "src/src/f16-vbinary/gen/f16-vsqrdiffc-f16c-u16.c",
++    "src/src/f16-vbinary/gen/f16-vsub-f16c-u16.c",
++    "src/src/f16-vbinary/gen/f16-vsubc-f16c-u16.c"
 +  ]
  
 -  source_set("f16-vsigmoid_f16c-fma-avx2") {
@@ -4381,7 +4439,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -4395,12 +4453,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++# This is a target that cannot depend on //base.
++source_set("f16-vbinary_f16c-no-avx2-no-fma_standalone") {
++  cflags = [
++    "-mf16c",
++    "-mno-avx2",
++    "-mno-fma"
++  ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set("f16-vsigmoid_f16c-fma-avx2_standalone") {
@@ -4409,28 +4472,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mf16c",
 -      "-mfma",
 -    ]
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f16-vsigmoid/gen/f16-vsigmoid-avx2-rr1-p2-rcp-u32.c",
--    ]
-+source_set("f16-vbinary_f16c-no-avx2-no-fma") {
-+  cflags = [
-+    "-mf16c",
-+    "-mno-avx2",
-+    "-mno-fma"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/f16-vbinary/gen/f16-vadd-f16c-u16.c",
@@ -4454,38 +4495,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/f16-vbinary/gen/f16-vsubc-f16c-u16.c"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--  source_set("f16-vsin_x64") {
--    cflags = []
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vsin/gen/f16-vsin-scalar-rational-3-2-div.c",
+-      "src/src/f16-vsigmoid/gen/f16-vsigmoid-avx2-rr1-p2-rcp-u32.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-vbinary_f16c-no-avx2-no-fma_standalone") {
-+  cflags = [
-+    "-mf16c",
-+    "-mno-avx2",
-+    "-mno-fma"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -4495,40 +4511,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vbinary/gen/f16-vadd-f16c-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vaddc-f16c-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vdiv-f16c-u8.c",
-+    "src/src/f16-vbinary/gen/f16-vdivc-f16c-u8.c",
-+    "src/src/f16-vbinary/gen/f16-vmax-f16c-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vmaxc-f16c-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vmin-f16c-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vminc-f16c-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vmul-f16c-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vmulc-f16c-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vprelu-f16c-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vpreluc-f16c-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vrdivc-f16c-u8.c",
-+    "src/src/f16-vbinary/gen/f16-vrpreluc-f16c-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vrsubc-f16c-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vsqrdiff-f16c-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vsqrdiffc-f16c-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vsub-f16c-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vsubc-f16c-u16.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("f16-vsin_x64_standalone") {
--    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -4536,28 +4520,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f16-vsin/gen/f16-vsin-scalar-rational-3-2-div.c",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
  
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
  
--    public_configs = [ ":xnnpack_config" ]
+-  source_set("f16-vsin_x64") {
+-    cflags = []
 +source_set("f16-vclamp_f16c-no-avx2-no-fma") {
 +  cflags = [
 +    "-mf16c",
@@ -4565,29 +4540,31 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "-mno-fma"
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f16-vsin/gen/f16-vsin-scalar-rational-3-2-div.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/f16-vclamp/gen/f16-vclamp-f16c-u16.c"
 +  ]
  
--  source_set("f16-vsqrt_f16c-no-avx2-no-fma") {
--    cflags = [
--      "-mf16c",
--      "-mno-avx2",
--      "-mno-fma",
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f16-vsqrt/gen/f16-vsqrt-f16c-rsqrt-u32.c",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -4595,21 +4572,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
+-  # This is a target that cannot depend on //base.
+-  source_set("f16-vsin_x64_standalone") {
+-    cflags = []
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f16-vsin/gen/f16-vsin-scalar-rational-3-2-div.c",
+-    ]
 +# This is a target that cannot depend on //base.
 +source_set("f16-vclamp_f16c-no-avx2-no-fma_standalone") {
 +  cflags = [
@@ -4618,26 +4590,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "-mno-fma"
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("f16-vsqrt_f16c-no-avx2-no-fma_standalone") {
--    cflags = [
--      "-mf16c",
--      "-mno-avx2",
--      "-mno-fma",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vclamp/gen/f16-vclamp-f16c-u16.c"
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f16-vsqrt/gen/f16-vsqrt-f16c-rsqrt-u32.c",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -4648,6 +4600,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vclamp/gen/f16-vclamp-f16c-u16.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -4666,18 +4627,18 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("f16-vtanh_f16c-fma-no-avx2") {
+-  source_set("f16-vsqrt_f16c-no-avx2-no-fma") {
 -    cflags = [
 -      "-mf16c",
--      "-mfma",
 -      "-mno-avx2",
+-      "-mno-fma",
 -    ]
 +source_set("f16-vcos_x64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vtanh/gen/f16-vtanh-fma3-polynomial-p19h9t2-u32.c",
+-      "src/src/f16-vsqrt/gen/f16-vsqrt-f16c-rsqrt-u32.c",
 -    ]
 +  ]
  
@@ -4703,11 +4664,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f16-vtanh_f16c-fma-no-avx2_standalone") {
+-  source_set("f16-vsqrt_f16c-no-avx2-no-fma_standalone") {
 -    cflags = [
 -      "-mf16c",
--      "-mfma",
 -      "-mno-avx2",
+-      "-mno-fma",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -4718,7 +4679,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vtanh/gen/f16-vtanh-fma3-polynomial-p19h9t2-u32.c",
+-      "src/src/f16-vsqrt/gen/f16-vsqrt-f16c-rsqrt-u32.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -4749,11 +4710,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/f16-vcos/gen/f16-vcos-scalar-rational-3-2-div.c"
 +  ]
  
--  source_set("f16-vtanh_f16c-no-avx2-no-fma") {
+-  source_set("f16-vtanh_f16c-fma-no-avx2") {
 -    cflags = [
 -      "-mf16c",
+-      "-mfma",
 -      "-mno-avx2",
--      "-mno-fma",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -4761,7 +4722,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vtanh/gen/f16-vtanh-f16c-expm1minus-rr1-p3h2ts-rcp-u24.c",
+-      "src/src/f16-vtanh/gen/f16-vtanh-fma3-polynomial-p19h9t2-u32.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -4789,11 +4750,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f16-vtanh_f16c-no-avx2-no-fma_standalone") {
+-  source_set("f16-vtanh_f16c-fma-no-avx2_standalone") {
 -    cflags = [
 -      "-mf16c",
+-      "-mfma",
 -      "-mno-avx2",
--      "-mno-fma",
 -    ]
 +source_set("f16-velu_f16c-fma-avx2") {
 +  cflags = [
@@ -4804,7 +4765,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vtanh/gen/f16-vtanh-f16c-expm1minus-rr1-p3h2ts-rcp-u24.c",
+-      "src/src/f16-vtanh/gen/f16-vtanh-fma3-polynomial-p19h9t2-u32.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
@@ -4840,7 +4801,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
--  source_set("f16-vunary_f16c-no-avx2-no-fma") {
+-  source_set("f16-vtanh_f16c-no-avx2-no-fma") {
 -    cflags = [
 -      "-mf16c",
 -      "-mno-avx2",
@@ -4856,7 +4817,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vunary/gen/f16-vsqr-f16c-u16.c",
+-      "src/src/f16-vtanh/gen/f16-vtanh-f16c-expm1minus-rr1-p3h2ts-rcp-u24.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
@@ -4887,7 +4848,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f16-vunary_f16c-no-avx2-no-fma_standalone") {
+-  source_set("f16-vtanh_f16c-no-avx2-no-fma_standalone") {
 -    cflags = [
 -      "-mf16c",
 -      "-mno-avx2",
@@ -4897,7 +4858,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vunary/gen/f16-vsqr-f16c-u16.c",
+-      "src/src/f16-vtanh/gen/f16-vtanh-f16c-expm1minus-rr1-p3h2ts-rcp-u24.c",
 -    ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
@@ -4929,10 +4890,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/f16-vexp/gen/f16-vexp-scalar-poly-3.c"
 +  ]
  
--  source_set("f16-vunary_sse2-no-sse3") {
+-  source_set("f16-vunary_f16c-no-avx2-no-fma") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mf16c",
+-      "-mno-avx2",
+-      "-mno-fma",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -4940,8 +4902,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vunary/gen/f16-vabs-sse2-u16.c",
--      "src/src/f16-vunary/gen/f16-vneg-sse2-u16.c",
+-      "src/src/f16-vunary/gen/f16-vsqr-f16c-u16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -4970,17 +4931,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  cflags = [
  
 -  # This is a target that cannot depend on //base.
--  source_set("f16-vunary_sse2-no-sse3_standalone") {
+-  source_set("f16-vunary_f16c-no-avx2-no-fma_standalone") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mf16c",
+-      "-mno-avx2",
+-      "-mno-fma",
 -    ]
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f16-vunary/gen/f16-vabs-sse2-u16.c",
--      "src/src/f16-vunary/gen/f16-vneg-sse2-u16.c",
+-      "src/src/f16-vunary/gen/f16-vsqr-f16c-u16.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
@@ -5019,7 +4980,7 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("f32-argmaxpool_sse2-no-sse3") {
+-  source_set("f16-vunary_sse2-no-sse3") {
 -    cflags = [
 -      "-mno-sse3",
 -      "-msse2",
@@ -5029,7 +4990,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-sse2-c4.c",
+-      "src/src/f16-vunary/gen/f16-vabs-sse2-u16.c",
+-      "src/src/f16-vunary/gen/f16-vneg-sse2-u16.c",
 -    ]
 +  ]
  
@@ -5055,7 +5017,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-argmaxpool_sse2-no-sse3_standalone") {
+-  source_set("f16-vunary_sse2-no-sse3_standalone") {
 -    cflags = [
 -      "-mno-sse3",
 -      "-msse2",
@@ -5069,7 +5031,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-sse2-c4.c",
+-      "src/src/f16-vunary/gen/f16-vabs-sse2-u16.c",
+-      "src/src/f16-vunary/gen/f16-vneg-sse2-u16.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -5100,15 +5063,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/f16-vgelu/gen/f16-vgelu-scalar-rational-6-4-div.c"
 +  ]
  
--  source_set("f32-argmaxpool_x64") {
--    cflags = []
+-  source_set("f32-argmaxpool_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c",
+-      "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-sse2-c4.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -5136,12 +5102,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-argmaxpool_x64_standalone") {
--    cflags = []
+-  source_set("f32-argmaxpool_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c",
+-      "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-sse2-c4.c",
 -    ]
 +source_set("f16-vhswish_f16c-no-avx2-no-fma") {
 +  cflags = [
@@ -5181,17 +5150,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  source_set("f32-avgpool_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("f32-argmaxpool_x64") {
+-    cflags = []
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-sse2-u4.c",
+-      "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c",
 -    ]
 +# This is a target that cannot depend on //base.
 +source_set("f16-vhswish_f16c-no-avx2-no-fma_standalone") {
@@ -5223,11 +5189,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-avgpool_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("f32-argmaxpool_x64_standalone") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -5237,7 +5200,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-sse2-u4.c",
+-      "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
  
@@ -5273,15 +5236,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/f16-vlrelu/gen/f16-vlrelu-f16c-u16.c"
 +  ]
  
--  source_set("f32-avgpool_x64") {
--    cflags = []
+-  source_set("f32-avgpool_avx-no-avx2-no-f16c-no-fma") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-scalar-u1.c",
+-      "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-avx-u8.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -5314,8 +5282,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-avgpool_x64_standalone") {
--    cflags = []
+-  source_set("f32-avgpool_avx-no-avx2-no-f16c-no-fma_standalone") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/f16-vlrelu/gen/f16-vlrelu-f16c-u16.c"
@@ -5323,7 +5296,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-scalar-u1.c",
+-      "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-avx-u8.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -5357,11 +5330,8 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("f32-conv-hwc2chw_sse-no-sse2") {
--    cflags = [
--      "-mno-sse2",
--      "-msse",
--    ]
+-  source_set("f32-avgpool_avx512f") {
+-    cflags = [ "-mavx512f" ]
 +source_set("f16-vmulcaddc_f16c-fma-no-avx2") {
 +  cflags = [
 +    "-mf16c",
@@ -5371,7 +5341,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-sse-2x2.c",
+-      "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-avx512f-u16.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
@@ -5402,17 +5372,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-conv-hwc2chw_sse-no-sse2_standalone") {
--    cflags = [
--      "-mno-sse2",
--      "-msse",
--    ]
+-  source_set("f32-avgpool_avx512f_standalone") {
+-    cflags = [ "-mavx512f" ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-sse-2x2.c",
+-      "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-avx512f-u16.c",
 -    ]
 +# This is a target that cannot depend on //base.
 +source_set("f16-vmulcaddc_f16c-fma-no-avx2_standalone") {
@@ -5453,13 +5420,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--  source_set("f32-conv-hwc2chw_x64") {
--    cflags = []
+-  source_set("f32-avgpool_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  public_configs = [ ":xnnpack_config" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-scalar-1x1.c",
+-      "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-sse2-u4.c",
 -    ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
@@ -5494,15 +5464,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-conv-hwc2chw_x64_standalone") {
--    cflags = []
+-  source_set("f32-avgpool_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-scalar-1x1.c",
+-      "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-sse2-u4.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -5545,21 +5518,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/f16-vrnd/gen/f16-vrndz-f16c-u16.c"
 +  ]
  
--  source_set("f32-dwconv2d-chw_sse-no-sse2") {
--    cflags = [
--      "-mno-sse2",
--      "-msse",
--    ]
+-  source_set("f32-avgpool_x64") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-sse-2x4-acc2.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-sse-1x4-acc3.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-sse-4x4.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-sse-2x4.c",
+-      "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-scalar-u1.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -5587,10 +5554,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-dwconv2d-chw_sse-no-sse2_standalone") {
--    cflags = [
--      "-mno-sse2",
--      "-msse",
+-  source_set("f32-avgpool_x64_standalone") {
+-    cflags = []
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-scalar-u1.c",
 -    ]
 +source_set("f16-vrsqrt_f16c-no-avx2-no-fma") {
 +  cflags = [
@@ -5599,18 +5568,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "-mno-fma"
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-sse-2x4-acc2.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-sse-1x4-acc3.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-sse-4x4.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-sse-2x4.c",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vrsqrt/gen/f16-vrsqrt-f16c-rsqrt-u32.c"
-+  ]
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -5621,11 +5578,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vrsqrt/gen/f16-vrsqrt-f16c-rsqrt-u32.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -5633,17 +5599,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-  source_set("f32-conv-hwc2chw_sse-no-sse2") {
+-    cflags = [
+-      "-mno-sse2",
+-      "-msse",
+-    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
--  source_set("f32-dwconv2d-chw_ssse3-no-sse4.1") {
--    cflags = [
--      "-mno-sse4.1",
--      "-mssse3",
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-sse-2x2.c",
 -    ]
 +# This is a target that cannot depend on //base.
 +source_set("f16-vrsqrt_f16c-no-avx2-no-fma_standalone") {
@@ -5653,15 +5619,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "-mno-fma"
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-ssse3-2x4-acc2.c",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vrsqrt/gen/f16-vrsqrt-f16c-rsqrt-u32.c"
-+  ]
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -5672,10 +5629,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vrsqrt/gen/f16-vrsqrt-f16c-rsqrt-u32.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-conv-hwc2chw_sse-no-sse2_standalone") {
+-    cflags = [
+-      "-mno-sse2",
+-      "-msse",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -5683,20 +5653,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-sse-2x2.c",
+-    ]
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-   }
++  }
 +}
  
--  # This is a target that cannot depend on //base.
--  source_set("f32-dwconv2d-chw_ssse3-no-sse4.1_standalone") {
--    cflags = [
--      "-mno-sse4.1",
--      "-mssse3",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
 +source_set("f16-vsigmoid_f16c-fma-avx2") {
 +  cflags = [
 +    "-mavx2",
@@ -5704,30 +5682,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "-mfma"
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-ssse3-2x4-acc2.c",
--    ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/f16-vsigmoid/gen/f16-vsigmoid-avx2-rr1-p2-rcp-u32.c"
 +  ]
  
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
+-  source_set("f32-conv-hwc2chw_x64") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-scalar-1x1.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -5735,28 +5708,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
--
--  source_set("f32-dwconv2d-chw_x64") {
--    cflags = []
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-2x1-acc2.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-4x1.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-1x1-acc2.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-2x1-acc2.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-1x1-acc5.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-2x1-acc2.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-1x1-acc5.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-2x1-acc2.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -5767,6 +5718,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +# This is a target that cannot depend on //base.
 +source_set("f16-vsigmoid_f16c-fma-avx2_standalone") {
 +  cflags = [
@@ -5775,27 +5731,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "-mfma"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-conv-hwc2chw_x64_standalone") {
+-    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/f16-vsigmoid/gen/f16-vsigmoid-avx2-rr1-p2-rcp-u32.c"
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("f32-dwconv2d-chw_x64_standalone") {
--    cflags = []
--
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-2x1-acc2.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-4x1.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-1x1-acc2.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-2x1-acc2.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-1x1-acc5.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-2x1-acc2.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-1x1-acc5.c",
--      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-2x1-acc2.c",
+-      "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-scalar-1x1.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -5829,22 +5775,20 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("f32-dwconv_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("f32-dwconv2d-chw_sse-no-sse2") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mno-sse2",
+-      "-msse",
 -    ]
 +source_set("f16-vsin_x64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-dwconv/gen/f32-dwconv-25p16c-minmax-avx.c",
--      "src/src/f32-dwconv/gen/f32-dwconv-3p16c-minmax-avx.c",
--      "src/src/f32-dwconv/gen/f32-dwconv-4p16c-minmax-avx.c",
--      "src/src/f32-dwconv/gen/f32-dwconv-9p16c-minmax-avx.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-sse-2x4-acc2.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-sse-1x4-acc3.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-sse-4x4.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-sse-2x4.c",
 -    ]
 +  ]
  
@@ -5870,12 +5814,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-dwconv_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("f32-dwconv2d-chw_sse-no-sse2_standalone") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mno-sse2",
+-      "-msse",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -5886,10 +5828,10 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-dwconv/gen/f32-dwconv-25p16c-minmax-avx.c",
--      "src/src/f32-dwconv/gen/f32-dwconv-3p16c-minmax-avx.c",
--      "src/src/f32-dwconv/gen/f32-dwconv-4p16c-minmax-avx.c",
--      "src/src/f32-dwconv/gen/f32-dwconv-9p16c-minmax-avx.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-sse-2x4-acc2.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-sse-1x4-acc3.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-sse-4x4.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-sse-2x4.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -5915,20 +5857,211 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vsin/gen/f16-vsin-scalar-rational-3-2-div.c"
++  ]
+ 
+-  source_set("f32-dwconv2d-chw_ssse3-no-sse4.1") {
+-    cflags = [
+-      "-mno-sse4.1",
+-      "-mssse3",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-ssse3-2x4-acc2.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
 -
--  source_set("f32-dwconv_avx512f") {
--    cflags = [ "-mavx512f" ]
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-dwconv2d-chw_ssse3-no-sse4.1_standalone") {
+-    cflags = [
+-      "-mno-sse4.1",
+-      "-mssse3",
+-    ]
++source_set("f16-vsqrt_f16c-no-avx2-no-fma") {
++  cflags = [
++    "-mf16c",
++    "-mno-avx2",
++    "-mno-fma"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-ssse3-2x4-acc2.c",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vsqrt/gen/f16-vsqrt-f16c-rsqrt-u32.c"
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
+-
+-  source_set("f32-dwconv2d-chw_x64") {
+-    cflags = []
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-dwconv/gen/f32-dwconv-25p16c-minmax-avx512f.c",
--      "src/src/f32-dwconv/gen/f32-dwconv-3p16c-minmax-avx512f.c",
--      "src/src/f32-dwconv/gen/f32-dwconv-4p16c-minmax-avx512f.c",
--      "src/src/f32-dwconv/gen/f32-dwconv-9p16c-minmax-avx512f.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-2x1-acc2.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-4x1.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-1x1-acc2.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-2x1-acc2.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-1x1-acc5.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-2x1-acc2.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-1x1-acc5.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-2x1-acc2.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
 -    ]
++# This is a target that cannot depend on //base.
++source_set("f16-vsqrt_f16c-no-avx2-no-fma_standalone") {
++  cflags = [
++    "-mf16c",
++    "-mno-avx2",
++    "-mno-fma"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vsin/gen/f16-vsin-scalar-rational-3-2-div.c"
++    "src/src/f16-vsqrt/gen/f16-vsqrt-f16c-rsqrt-u32.c"
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-dwconv2d-chw_x64_standalone") {
+-    cflags = []
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-2x1-acc2.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-4x1.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-1x1-acc2.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-2x1-acc2.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-1x1-acc5.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-2x1-acc2.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-1x1-acc5.c",
+-      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-2x1-acc2.c",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  source_set("f32-dwconv_avx-no-avx2-no-f16c-no-fma") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
++source_set("f16-vtanh_f16c-fma-no-avx2") {
++  cflags = [
++    "-mf16c",
++    "-mfma",
++    "-mno-avx2"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-dwconv/gen/f32-dwconv-25p16c-minmax-avx.c",
+-      "src/src/f32-dwconv/gen/f32-dwconv-3p16c-minmax-avx.c",
+-      "src/src/f32-dwconv/gen/f32-dwconv-4p16c-minmax-avx.c",
+-      "src/src/f32-dwconv/gen/f32-dwconv-9p16c-minmax-avx.c",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vtanh/gen/f16-vtanh-fma3-polynomial-p19h9t2-u32.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -5951,11 +6084,61 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-dwconv_avx512f_standalone") {
+-  source_set("f32-dwconv_avx-no-avx2-no-f16c-no-fma_standalone") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-dwconv/gen/f32-dwconv-25p16c-minmax-avx.c",
+-      "src/src/f32-dwconv/gen/f32-dwconv-3p16c-minmax-avx.c",
+-      "src/src/f32-dwconv/gen/f32-dwconv-4p16c-minmax-avx.c",
+-      "src/src/f32-dwconv/gen/f32-dwconv-9p16c-minmax-avx.c",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("f16-vtanh_f16c-fma-no-avx2_standalone") {
++  cflags = [
++    "-mf16c",
++    "-mfma",
++    "-mno-avx2"
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vtanh/gen/f16-vtanh-fma3-polynomial-p19h9t2-u32.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
+-
+-  source_set("f32-dwconv_avx512f") {
 -    cflags = [ "-mavx512f" ]
 -
 -    sources = [
@@ -5965,7 +6148,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-dwconv/gen/f32-dwconv-4p16c-minmax-avx512f.c",
 -      "src/src/f32-dwconv/gen/f32-dwconv-9p16c-minmax-avx512f.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -5975,15 +6163,41 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
  
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-dwconv_avx512f_standalone") {
+-    cflags = [ "-mavx512f" ]
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-dwconv/gen/f32-dwconv-25p16c-minmax-avx512f.c",
+-      "src/src/f32-dwconv/gen/f32-dwconv-3p16c-minmax-avx512f.c",
+-      "src/src/f32-dwconv/gen/f32-dwconv-4p16c-minmax-avx512f.c",
+-      "src/src/f32-dwconv/gen/f32-dwconv-9p16c-minmax-avx512f.c",
+-    ]
+-
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
+-
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("f16-vsqrt_f16c-no-avx2-no-fma") {
++source_set("f16-vtanh_f16c-no-avx2-no-fma") {
 +  cflags = [
 +    "-mf16c",
 +    "-mno-avx2",
@@ -5996,7 +6210,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vsqrt/gen/f16-vsqrt-f16c-rsqrt-u32.c"
++    "src/src/f16-vtanh/gen/f16-vtanh-f16c-expm1minus-rr1-p3h2ts-rcp-u24.c"
 +  ]
  
 -  source_set("f32-dwconv_f16c-fma-no-avx2") {
@@ -6039,7 +6253,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("f16-vsqrt_f16c-no-avx2-no-fma_standalone") {
++source_set("f16-vtanh_f16c-no-avx2-no-fma_standalone") {
 +  cflags = [
 +    "-mf16c",
 +    "-mno-avx2",
@@ -6055,7 +6269,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vsqrt/gen/f16-vsqrt-f16c-rsqrt-u32.c"
++    "src/src/f16-vtanh/gen/f16-vtanh-f16c-expm1minus-rr1-p3h2ts-rcp-u24.c"
 +  ]
  
 -    sources = [
@@ -6102,11 +6316,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse2",
 -      "-msse",
 -    ]
-+source_set("f16-vtanh_f16c-fma-no-avx2") {
++source_set("f16-vunary_f16c-no-avx2-no-fma") {
 +  cflags = [
 +    "-mf16c",
-+    "-mfma",
-+    "-mno-avx2"
++    "-mno-avx2",
++    "-mno-fma"
 +  ]
  
 -    sources = [
@@ -6118,7 +6332,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vtanh/gen/f16-vtanh-fma3-polynomial-p19h9t2-u32.c"
++    "src/src/f16-vunary/gen/f16-vsqr-f16c-u16.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -6161,11 +6375,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-dwconv/gen/f32-dwconv-9p8c-minmax-sse.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-vtanh_f16c-fma-no-avx2_standalone") {
++source_set("f16-vunary_f16c-no-avx2-no-fma_standalone") {
 +  cflags = [
 +    "-mf16c",
-+    "-mfma",
-+    "-mno-avx2"
++    "-mno-avx2",
++    "-mno-fma"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -6180,7 +6394,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vtanh/gen/f16-vtanh-fma3-polynomial-p19h9t2-u32.c"
++    "src/src/f16-vunary/gen/f16-vsqr-f16c-u16.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -6251,11 +6465,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-dwconv/gen/f32-dwconv-9p1c-minmax-scalar-acc2.c",
 -      "src/src/f32-dwconv/gen/f32-dwconv-9p1c-scalar-acc2.c",
 -    ]
-+source_set("f16-vtanh_f16c-no-avx2-no-fma") {
++source_set("f16-vunary_sse2-no-sse3") {
 +  cflags = [
-+    "-mf16c",
-+    "-mno-avx2",
-+    "-mno-fma"
++    "-mno-sse3",
++    "-msse2"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -6270,7 +6483,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vtanh/gen/f16-vtanh-f16c-expm1minus-rr1-p3h2ts-rcp-u24.c"
++    "src/src/f16-vunary/gen/f16-vabs-sse2-u16.c",
++    "src/src/f16-vunary/gen/f16-vneg-sse2-u16.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -6304,11 +6518,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-avx-u24.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-vtanh_f16c-no-avx2-no-fma_standalone") {
++source_set("f16-vunary_sse2-no-sse3_standalone") {
 +  cflags = [
-+    "-mf16c",
-+    "-mno-avx2",
-+    "-mno-fma"
++    "-mno-sse3",
++    "-msse2"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -6323,7 +6536,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vtanh/gen/f16-vtanh-f16c-expm1minus-rr1-p3h2ts-rcp-u24.c"
++    "src/src/f16-vunary/gen/f16-vabs-sse2-u16.c",
++    "src/src/f16-vunary/gen/f16-vneg-sse2-u16.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -6369,11 +6583,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("f16-vunary_f16c-no-avx2-no-fma") {
++source_set("f32-argmaxpool_sse2-no-sse3") {
 +  cflags = [
-+    "-mf16c",
-+    "-mno-avx2",
-+    "-mno-fma"
++    "-mno-sse3",
++    "-msse2"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -6394,7 +6607,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vunary/gen/f16-vsqr-f16c-u16.c"
++    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-sse2-c4.c"
 +  ]
  
 -    sources = [
@@ -6440,11 +6653,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mfma",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-vunary_f16c-no-avx2-no-fma_standalone") {
++source_set("f32-argmaxpool_sse2-no-sse3_standalone") {
 +  cflags = [
-+    "-mf16c",
-+    "-mno-avx2",
-+    "-mno-fma"
++    "-mno-sse3",
++    "-msse2"
 +  ]
  
 -    sources = [
@@ -6453,7 +6665,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vunary/gen/f16-vsqr-f16c-u16.c"
++    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-sse2-c4.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -6469,15 +6681,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
- 
--    public_configs = [ ":xnnpack_config" ]
++
 +  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -6494,20 +6706,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-avx2",
 -      "-mno-fma",
 -    ]
-+source_set("f16-vunary_sse2-no-sse3") {
++source_set("f32-argmaxpool_x64") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
-+  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-f16c-u16.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vunary/gen/f16-vabs-sse2-u16.c",
-+    "src/src/f16-vunary/gen/f16-vneg-sse2-u16.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -6520,18 +6725,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set("f32-f16-vcvt_f16c-no-avx2-no-fma_standalone") {
@@ -6540,19 +6743,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-avx2",
 -      "-mno-fma",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-f16c-u16.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-vunary_sse2-no-sse3_standalone") {
-+  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -6564,26 +6767,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vunary/gen/f16-vabs-sse2-u16.c",
-+    "src/src/f16-vunary/gen/f16-vneg-sse2-u16.c"
-+  ]
++# This is a target that cannot depend on //base.
++source_set("f32-argmaxpool_x64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c"
 +  ]
  
 -  source_set("f32-f16-vcvt_sse2-no-sse3") {
@@ -6591,16 +6788,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-sse2-u16.c",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -6612,18 +6813,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("f32-argmaxpool_sse2-no-sse3") {
-+  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-sse2-c4.c"
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
 -  source_set("f32-f16-vcvt_sse2-no-sse3_standalone") {
@@ -6631,19 +6827,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-sse2-u16.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++source_set("f32-avgpool_avx-no-avx2-no-f16c-no-fma") {
++  cflags = [
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -6656,24 +6850,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-argmaxpool_sse2-no-sse3_standalone") {
-+  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-avx-u8.c"
 +  ]
  
+-    public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-sse2-c4.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -  source_set("f32-f16-vcvt_sse4.1-no-sse4.2") {
@@ -6681,19 +6876,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse4.2",
 -      "-msse4.1",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-sse41-u8.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++# This is a target that cannot depend on //base.
++source_set("f32-avgpool_avx-no-avx2-no-f16c-no-fma_standalone") {
++  cflags = [
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -6706,13 +6902,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-avx-u8.c"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set("f32-f16-vcvt_sse4.1-no-sse4.2_standalone") {
@@ -6720,13 +6919,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse4.2",
 -      "-msse4.1",
 -    ]
--
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-sse41-u8.c",
 -    ]
-+source_set("f32-argmaxpool_x64") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -6738,24 +6942,36 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c"
++source_set("f32-avgpool_avx512f") {
++  cflags = [
++    "-mavx512f"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-avx512f-u16.c"
++  ]
+ 
+-  source_set("f32-f16-vcvt_x64") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("f32-f16-vcvt_x64") {
--    cflags = []
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-scalar-fabsf-u2.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -6763,13 +6979,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-scalar-fabsf-u2.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -6780,12 +6989,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-argmaxpool_x64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
++# This is a target that cannot depend on //base.
++source_set("f32-avgpool_avx512f_standalone") {
++  cflags = [
++    "-mavx512f"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -6793,7 +7005,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c"
++    "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-avx512f-u16.c"
 +  ]
  
 -    sources = [
@@ -7205,9 +7417,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-gemm/gen/f32-gemm-1x4-minmax-scalar.c",
 -      "src/src/f32-gemm/gen/f32-gemm-1x4-relu-scalar.c",
 -      "src/src/f32-gemm/gen/f32-gemm-1x4-scalar.c",
--      "src/src/f32-gemm/gen/f32-gemm-2x4-minmax-scalar.c",
--      "src/src/f32-gemm/gen/f32-gemm-2x4-relu-scalar.c",
--      "src/src/f32-gemm/gen/f32-gemm-2x4-scalar.c",
 -      "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-scalar.c",
 -      "src/src/f32-gemm/gen/f32-gemm-4x2-scalar.c",
 -      "src/src/f32-gemm/gen/f32-gemm-4x4-minmax-scalar.c",
@@ -7248,9 +7457,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-gemm/gen/f32-gemm-1x4-minmax-scalar.c",
 -      "src/src/f32-gemm/gen/f32-gemm-1x4-relu-scalar.c",
 -      "src/src/f32-gemm/gen/f32-gemm-1x4-scalar.c",
--      "src/src/f32-gemm/gen/f32-gemm-2x4-minmax-scalar.c",
--      "src/src/f32-gemm/gen/f32-gemm-2x4-relu-scalar.c",
--      "src/src/f32-gemm/gen/f32-gemm-2x4-scalar.c",
 -      "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-scalar.c",
 -      "src/src/f32-gemm/gen/f32-gemm-4x2-scalar.c",
 -      "src/src/f32-gemm/gen/f32-gemm-4x4-minmax-scalar.c",
@@ -8042,9 +8248,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-igemm/gen/f32-igemm-1x4-minmax-scalar.c",
 -      "src/src/f32-igemm/gen/f32-igemm-1x4-relu-scalar.c",
 -      "src/src/f32-igemm/gen/f32-igemm-1x4-scalar.c",
--      "src/src/f32-igemm/gen/f32-igemm-2x4-minmax-scalar.c",
--      "src/src/f32-igemm/gen/f32-igemm-2x4-relu-scalar.c",
--      "src/src/f32-igemm/gen/f32-igemm-2x4-scalar.c",
 -      "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-scalar.c",
 -      "src/src/f32-igemm/gen/f32-igemm-4x2-scalar.c",
 -      "src/src/f32-igemm/gen/f32-igemm-4x4-minmax-scalar.c",
@@ -8086,9 +8289,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-igemm/gen/f32-igemm-1x4-minmax-scalar.c",
 -      "src/src/f32-igemm/gen/f32-igemm-1x4-relu-scalar.c",
 -      "src/src/f32-igemm/gen/f32-igemm-1x4-scalar.c",
--      "src/src/f32-igemm/gen/f32-igemm-2x4-minmax-scalar.c",
--      "src/src/f32-igemm/gen/f32-igemm-2x4-relu-scalar.c",
--      "src/src/f32-igemm/gen/f32-igemm-2x4-scalar.c",
 -      "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-scalar.c",
 -      "src/src/f32-igemm/gen/f32-igemm-4x2-scalar.c",
 -      "src/src/f32-igemm/gen/f32-igemm-4x4-minmax-scalar.c",
@@ -9691,9 +9891,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/f32-gemm/gen/f32-gemm-1x4-minmax-scalar.c",
 +    "src/src/f32-gemm/gen/f32-gemm-1x4-relu-scalar.c",
 +    "src/src/f32-gemm/gen/f32-gemm-1x4-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-relu-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-scalar.c",
 +    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-scalar.c",
 +    "src/src/f32-gemm/gen/f32-gemm-4x2-scalar.c",
 +    "src/src/f32-gemm/gen/f32-gemm-4x4-minmax-scalar.c",
@@ -9753,9 +9950,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/f32-gemm/gen/f32-gemm-1x4-minmax-scalar.c",
 +    "src/src/f32-gemm/gen/f32-gemm-1x4-relu-scalar.c",
 +    "src/src/f32-gemm/gen/f32-gemm-1x4-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-relu-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-scalar.c",
 +    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-scalar.c",
 +    "src/src/f32-gemm/gen/f32-gemm-4x2-scalar.c",
 +    "src/src/f32-gemm/gen/f32-gemm-4x4-minmax-scalar.c",
@@ -9862,7 +10056,6 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u1.c",
 -      "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u4.c",
 -      "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-lrintf-u4.c",
 -    ]
@@ -9884,34 +10077,30 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
  
 -  # This is a target that cannot depend on //base.
 -  source_set("f32-qs8-vcvt_x64_standalone") {
 -    cflags = []
--
++  public_configs = [ ":xnnpack_config" ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u1.c",
 -      "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u4.c",
 -      "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-lrintf-u4.c",
 -    ]
-+source_set("f32-ibilinear-chw_x64") {
-+  cflags = [
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -9923,21 +10112,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
++source_set("f32-ibilinear-chw_x64") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-scalar-p4.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-scalar-p4.c"
++  ]
  
 -  source_set("f32-qu8-vcvt_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
@@ -9946,6 +10134,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-avx-u32.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -9953,13 +10149,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-avx-u32.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -9970,13 +10159,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-ibilinear-chw_x64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  ]
++# This is a target that cannot depend on //base.
++source_set("f32-ibilinear-chw_x64_standalone") {
++  cflags = [
  
 -  # This is a target that cannot depend on //base.
 -  source_set("f32-qu8-vcvt_avx-no-avx2-no-f16c-no-fma_standalone") {
@@ -9986,18 +10176,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-scalar-p4.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-avx-u32.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-scalar-p4.c"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -10009,6 +10197,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -10016,16 +10209,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
- 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  public_configs = [ ":xnnpack_config" ]
  
 -  source_set("f32-qu8-vcvt_f16c-fma-avx2") {
 -    cflags = [
@@ -10033,7 +10221,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mf16c",
 -      "-mfma",
 -    ]
--
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-avx2-u64.c",
@@ -10290,14 +10482,63 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f32-qu8-vcvt_x64") {
 -    cflags = []
++source_set("f32-igemm_avx-no-avx2-no-f16c-no-fma") {
++  cflags = [
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u4.c",
+-      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-lrintf-u4.c",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-igemm/gen/f32-igemm-1x16-minmax-avx-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-avx-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-5x16-minmax-avx-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-5x8-minmax-avx-broadcast.c"
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
 -
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-qu8-vcvt_x64_standalone") {
+-    cflags = []
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u1.c",
 -      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u4.c",
 -      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-lrintf-u4.c",
 -    ]
-+source_set("f32-igemm_avx-no-avx2-no-f16c-no-fma") {
++# This is a target that cannot depend on //base.
++source_set("f32-igemm_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
 +    "-mavx",
 +    "-mno-avx2",
@@ -10313,7 +10554,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
@@ -10324,20 +10565,69 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("f32-qu8-vcvt_x64_standalone") {
--    cflags = []
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-  source_set("f32-raddstoreexpminusmax_avx512f") {
+-    cflags = [ "-mavx512f" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-avx512f-rr2-p5-u64-acc2.c",
+-    ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
 -
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++source_set("f32-igemm_avx512f") {
++  cflags = [
++    "-mavx512f"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-igemm/gen/f32-igemm-1x16-minmax-avx512f-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x32-minmax-avx512f-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-7x16-minmax-avx512f-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-7x32-minmax-avx512f-broadcast.c"
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-raddstoreexpminusmax_avx512f_standalone") {
+-    cflags = [ "-mavx512f" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u1.c",
--      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u4.c",
--      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-lrintf-u4.c",
+-      "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-avx512f-rr2-p5-u64-acc2.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -10361,12 +10651,9 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-igemm_avx-no-avx2-no-f16c-no-fma_standalone") {
++source_set("f32-igemm_avx512f_standalone") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mavx512f"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -10375,21 +10662,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-igemm/gen/f32-igemm-1x16-minmax-avx-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-avx-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-5x16-minmax-avx-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-5x8-minmax-avx-broadcast.c"
++    "src/src/f32-igemm/gen/f32-igemm-1x16-minmax-avx512f-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x32-minmax-avx512f-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-7x16-minmax-avx512f-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-7x32-minmax-avx512f-broadcast.c"
 +  ]
  
--  source_set("f32-raddstoreexpminusmax_avx512f") {
--    cflags = [ "-mavx512f" ]
+-  source_set("f32-raddstoreexpminusmax_f16c-fma-avx2") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-avx512f-rr2-p5-u64-acc2.c",
+-      "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-avx2-rr2-p5-u32-acc2.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -10417,64 +10708,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-raddstoreexpminusmax_avx512f_standalone") {
--    cflags = [ "-mavx512f" ]
-+source_set("f32-igemm_avx512f") {
-+  cflags = [
-+    "-mavx512f"
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-avx512f-rr2-p5-u64-acc2.c",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-igemm/gen/f32-igemm-1x16-minmax-avx512f-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x32-minmax-avx512f-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-7x16-minmax-avx512f-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-7x32-minmax-avx512f-broadcast.c"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--  source_set("f32-raddstoreexpminusmax_f16c-fma-avx2") {
+-  source_set("f32-raddstoreexpminusmax_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-igemm_avx512f_standalone") {
++source_set("f32-igemm_f16c-fma-no-avx2") {
 +  cflags = [
-+    "-mavx512f"
++    "-mf16c",
++    "-mfma",
++    "-mno-avx2"
 +  ]
  
 -    sources = [
@@ -10483,10 +10727,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-igemm/gen/f32-igemm-1x16-minmax-avx512f-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x32-minmax-avx512f-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-7x16-minmax-avx512f-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-7x32-minmax-avx512f-broadcast.c"
++    "src/src/f32-igemm/gen/f32-igemm-10x8-minmax-fma3-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x16-minmax-fma3-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x16s4-minmax-fma3-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-fma3-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x16s4-minmax-fma3-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-fma3-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-5x16-minmax-fma3-broadcast-prfm.c",
++    "src/src/f32-igemm/gen/f32-igemm-5x8-minmax-fma3-broadcast.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -10497,67 +10745,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("f32-raddstoreexpminusmax_f16c-fma-avx2_standalone") {
--    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-avx2-rr2-p5-u32-acc2.c",
--    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+source_set("f32-igemm_f16c-fma-no-avx2") {
-+  cflags = [
-+    "-mf16c",
-+    "-mfma",
-+    "-mno-avx2"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-igemm/gen/f32-igemm-10x8-minmax-fma3-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x16-minmax-fma3-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x16s4-minmax-fma3-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-fma3-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x16s4-minmax-fma3-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-fma3-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-5x16-minmax-fma3-broadcast-prfm.c",
-+    "src/src/f32-igemm/gen/f32-igemm-5x8-minmax-fma3-broadcast.c"
++    "//third_party/pthreadpool",
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -10576,19 +10775,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-avx256skx-rr2-p5-u32-acc2.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++# This is a target that cannot depend on //base.
++source_set("f32-igemm_f16c-fma-no-avx2_standalone") {
++  cflags = [
++    "-mf16c",
++    "-mfma",
++    "-mno-avx2"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -10601,18 +10800,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-igemm/gen/f32-igemm-10x8-minmax-fma3-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x16-minmax-fma3-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x16s4-minmax-fma3-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-fma3-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x16s4-minmax-fma3-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-fma3-broadcast.c",
++    "src/src/f32-igemm/gen/f32-igemm-5x16-minmax-fma3-broadcast-prfm.c",
++    "src/src/f32-igemm/gen/f32-igemm-5x8-minmax-fma3-broadcast.c"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+# This is a target that cannot depend on //base.
-+source_set("f32-igemm_f16c-fma-no-avx2_standalone") {
-+  cflags = [
-+    "-mf16c",
-+    "-mfma",
-+    "-mno-avx2"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
@@ -10626,25 +10830,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-igemm/gen/f32-igemm-10x8-minmax-fma3-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x16-minmax-fma3-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x16s4-minmax-fma3-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-fma3-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x16s4-minmax-fma3-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-fma3-broadcast.c",
-+    "src/src/f32-igemm/gen/f32-igemm-5x16-minmax-fma3-broadcast-prfm.c",
-+    "src/src/f32-igemm/gen/f32-igemm-5x8-minmax-fma3-broadcast.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-avx256skx-rr2-p5-u32-acc2.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -10656,52 +10853,21 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-   }
++  }
 +}
  
--  source_set("f32-raddstoreexpminusmax_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-sse2-rr2-p5-u16-acc2.c",
--    ]
--
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
 +source_set("f32-igemm_sse-no-sse2") {
 +  cflags = [
 +    "-mno-sse2",
 +    "-msse"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
@@ -10710,8 +10876,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-sse-load1.c"
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("f32-raddstoreexpminusmax_sse2-no-sse3_standalone") {
+-  source_set("f32-raddstoreexpminusmax_sse2-no-sse3") {
 -    cflags = [
 -      "-mno-sse3",
 -      "-msse2",
@@ -10739,12 +10904,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    public_configs = [ ":xnnpack_config" ]
+-  }
 +# This is a target that cannot depend on //base.
 +source_set("f32-igemm_sse-no-sse2_standalone") {
 +  cflags = [
@@ -10752,10 +10918,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "-msse"
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-raddstoreexpminusmax_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-sse-load1.c",
@@ -10763,22 +10931,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-sse-load1.c"
 +  ]
  
--  source_set("f32-raddstoreexpminusmax_x64") {
--    cflags = []
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-scalar-rr2-p5-u4-acc2.c",
+-      "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-sse2-rr2-p5-u16-acc2.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -10788,26 +10947,36 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
    }
 +}
  
--  # This is a target that cannot depend on //base.
--  source_set("f32-raddstoreexpminusmax_x64_standalone") {
+-  source_set("f32-raddstoreexpminusmax_x64") {
 -    cflags = []
--
++source_set("f32-igemm_x64") {
++  cflags = [
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-scalar-rr2-p5-u4-acc2.c",
 -    ]
-+source_set("f32-igemm_x64") {
-+  cflags = [
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -10817,19 +10986,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/f32-igemm/gen/f32-igemm-1x4-minmax-scalar.c",
 +    "src/src/f32-igemm/gen/f32-igemm-1x4-relu-scalar.c",
 +    "src/src/f32-igemm/gen/f32-igemm-1x4-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-relu-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-scalar.c",
 +    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-scalar.c",
 +    "src/src/f32-igemm/gen/f32-igemm-4x2-scalar.c",
 +    "src/src/f32-igemm/gen/f32-igemm-4x4-minmax-scalar.c",
@@ -10837,21 +11000,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/f32-igemm/gen/f32-igemm-4x4-scalar.c"
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
+-    public_configs = [ ":xnnpack_config" ]
 -  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("f32-rdminmax_avx-no-avx2-no-f16c-no-fma") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-raddstoreexpminusmax_x64_standalone") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -10861,8 +11018,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-avx-c32.c",
--      "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-avx-c32.c",
+-      "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-scalar-rr2-p5-u4-acc2.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -10875,32 +11031,24 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
 +source_set("f32-igemm_x64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("f32-rdminmax_avx-no-avx2-no-f16c-no-fma_standalone") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/f32-igemm/gen/f32-igemm-1x4-minmax-scalar.c",
 +    "src/src/f32-igemm/gen/f32-igemm-1x4-relu-scalar.c",
 +    "src/src/f32-igemm/gen/f32-igemm-1x4-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-relu-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-scalar.c",
 +    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-scalar.c",
 +    "src/src/f32-igemm/gen/f32-igemm-4x2-scalar.c",
 +    "src/src/f32-igemm/gen/f32-igemm-4x4-minmax-scalar.c",
@@ -10908,14 +11056,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/f32-igemm/gen/f32-igemm-4x4-scalar.c"
 +  ]
  
+-  source_set("f32-rdminmax_avx-no-avx2-no-f16c-no-fma") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-avx-c32.c",
 -      "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-avx-c32.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -10925,42 +11087,74 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
    }
 +}
  
--  source_set("f32-rdminmax_avx512f") {
--    cflags = [ "-mavx512f" ]
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-rdminmax_avx-no-avx2-no-f16c-no-fma_standalone") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-avx-c32.c",
+-      "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-avx-c32.c",
+-    ]
+-
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
+-
+-    public_configs = [ ":xnnpack_config" ]
 +source_set("f32-maxpool_sse2-no-sse3") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
 +  ]
  
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-maxpool/gen/f32-maxpool-9p-minmax-sse2-u4.c"
++  ]
+ 
+-  source_set("f32-rdminmax_avx512f") {
+-    cflags = [ "-mavx512f" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-avx512f-c32.c",
 -      "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-avx512f-c32.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-maxpool/gen/f32-maxpool-9p-minmax-sse2-u4.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -10973,36 +11167,34 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++# This is a target that cannot depend on //base.
++source_set("f32-maxpool_sse2-no-sse3_standalone") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
 +  ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set("f32-rdminmax_avx512f_standalone") {
 -    cflags = [ "-mavx512f" ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-maxpool/gen/f32-maxpool-9p-minmax-sse2-u4.c"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-avx512f-c32.c",
 -      "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-avx512f-c32.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-maxpool_sse2-no-sse3_standalone") {
-+  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -11014,15 +11206,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-maxpool/gen/f32-maxpool-9p-minmax-sse2-u4.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -12040,7 +12223,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-rsum/gen/f32-rsum-avx512f-u64-acc4.c",
+-      "src/src/f32-rsum/gen/f32-rsum-avx512f-u32-acc2.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -12089,7 +12272,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-rsum/gen/f32-rsum-avx512f-u64-acc4.c",
+-      "src/src/f32-rsum/gen/f32-rsum-avx512f-u32-acc2.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -12123,10 +12306,10 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("f32-rsum_sse-no-sse2") {
+-  source_set("f32-rsum_sse2-no-sse3") {
 -    cflags = [
--      "-mno-sse2",
--      "-msse",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +source_set("f32-qc8w-gemm_f16c-fma-no-avx2") {
 +  cflags = [
@@ -12137,7 +12320,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-rsum/gen/f32-rsum-sse-u16-acc4.c",
+-      "src/src/f32-rsum/gen/f32-rsum-sse2-u16-acc4.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
@@ -12169,17 +12352,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-rsum_sse-no-sse2_standalone") {
+-  source_set("f32-rsum_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mno-sse2",
--      "-msse",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-rsum/gen/f32-rsum-sse-u16-acc4.c",
+-      "src/src/f32-rsum/gen/f32-rsum-sse2-u16-acc4.c",
 -    ]
 +# This is a target that cannot depend on //base.
 +source_set("f32-qc8w-gemm_f16c-fma-no-avx2_standalone") {
@@ -12577,7 +12760,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-avx512f-rational-12-10-nr.c",
+-      "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-avx512f-rational-12-10-div.c",
 -    ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
@@ -12617,7 +12800,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-avx512f-rational-12-10-nr.c",
+-      "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-avx512f-rational-12-10-div.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -13150,7 +13333,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u1.c",
 +    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u4.c",
 +    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-lrintf-u4.c"
 +  ]
@@ -13223,7 +13405,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u1.c",
 +    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u4.c",
 +    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-lrintf-u4.c"
 +  ]
@@ -13493,7 +13674,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vclamp/gen/f32-vclamp-avx-u16.c",
+-      "src/src/f32-vclamp/gen/f32-vclamp-avx.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -13540,7 +13721,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vclamp/gen/f32-vclamp-avx-u16.c",
+-      "src/src/f32-vclamp/gen/f32-vclamp-avx.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -13589,7 +13770,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vclamp/gen/f32-vclamp-avx512f-u16.c",
+-      "src/src/f32-vclamp/gen/f32-vclamp-avx512f.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
@@ -13627,7 +13808,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vclamp/gen/f32-vclamp-avx512f-u16.c",
+-      "src/src/f32-vclamp/gen/f32-vclamp-avx512f.c",
 -    ]
 +# This is a target that cannot depend on //base.
 +source_set("f32-qu8-vcvt_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
@@ -13672,16 +13853,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--  source_set("f32-vclamp_sse-no-sse2") {
+-  source_set("f32-vclamp_sse2-no-sse3") {
 -    cflags = [
--      "-mno-sse2",
--      "-msse",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vclamp/gen/f32-vclamp-sse-u8.c",
+-      "src/src/f32-vclamp/gen/f32-vclamp-sse2.c",
 -    ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
@@ -13712,10 +13893,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vclamp_sse-no-sse2_standalone") {
+-  source_set("f32-vclamp_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mno-sse2",
--      "-msse",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -13723,7 +13904,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vclamp/gen/f32-vclamp-sse-u8.c",
+-      "src/src/f32-vclamp/gen/f32-vclamp-sse2.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -13770,7 +13951,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vclamp/gen/f32-vclamp-scalar-u4.c",
+-      "src/src/f32-vclamp/gen/f32-vclamp-scalar.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -13803,7 +13984,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vclamp/gen/f32-vclamp-scalar-u4.c",
+-      "src/src/f32-vclamp/gen/f32-vclamp-scalar.c",
 -    ]
 +source_set("f32-qu8-vcvt_x64") {
 +  cflags = [
@@ -13823,7 +14004,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u1.c",
 +    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u4.c",
 +    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-lrintf-u4.c"
 +  ]
@@ -13875,7 +14055,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = [ "-mavx512f" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u1.c",
 +    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u4.c",
 +    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-lrintf-u4.c"
 +  ]
@@ -14645,7 +14824,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vcos/gen/f32-vcos-avx512f-rational-5-4-nr.c",
+-      "src/src/f32-vcos/gen/f32-vcos-avx512f-rational-5-4-div.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
@@ -14684,7 +14863,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vcos/gen/f32-vcos-avx512f-rational-5-4-nr.c",
+-      "src/src/f32-vcos/gen/f32-vcos-avx512f-rational-5-4-div.c",
 -    ]
 +# This is a target that cannot depend on //base.
 +source_set("f32-rdminmax_avx512f_standalone") {
@@ -14983,22 +15162,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-velu/gen/f32-velu-avx-rr2-lut4-p4-perm-u32.c",
--    ]
--
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
 +source_set("f32-rdsum_avx-no-avx2-no-f16c-no-fma") {
 +  cflags = [
 +    "-mavx",
@@ -15006,51 +15169,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "-mno-f16c",
 +    "-mno-fma"
 +  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
++
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-avx-c32.c"
 +  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("f32-velu_avx-no-avx2-no-f16c-no-fma_standalone") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
++
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-velu/gen/f32-velu-avx-rr2-lut4-p4-perm-u32.c",
--    ]
++
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool",
 +  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
++
 +  public_configs = [ ":xnnpack_config" ]
 +}
- 
--    public_configs = [ ":xnnpack_config" ]
++
 +# This is a target that cannot depend on //base.
 +source_set("f32-rdsum_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
@@ -15059,133 +15197,69 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "-mno-f16c",
 +    "-mno-fma"
 +  ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
++
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-avx-c32.c"
 +  ]
- 
--  source_set("f32-velu_avx512f") {
--    cflags = [ "-mavx512f" ]
++
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-velu/gen/f32-velu-avx512f-rr1-p6-u64.c",
--    ]
++
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
++
 +  public_configs = [ ":xnnpack_config" ]
- 
--    public_configs = [ ":xnnpack_config" ]
++
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-   }
++  }
 +}
- 
--  # This is a target that cannot depend on //base.
--  source_set("f32-velu_avx512f_standalone") {
--    cflags = [ "-mavx512f" ]
++
 +source_set("f32-rdsum_avx512f") {
 +  cflags = [
 +    "-mavx512f"
 +  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-velu/gen/f32-velu-avx512f-rr1-p6-u64.c",
--    ]
++
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-avx512f-c64.c"
 +  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
++
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
++
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool",
 +  ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
++
 +  public_configs = [ ":xnnpack_config" ]
 +}
- 
--  source_set("f32-velu_f16c-fma-avx2") {
--    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
--    ]
++
 +# This is a target that cannot depend on //base.
 +source_set("f32-rdsum_avx512f_standalone") {
 +  cflags = [
 +    "-mavx512f"
 +  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-velu/gen/f32-velu-avx2-rr1-lut4-p4-perm-u32.c",
--    ]
++
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-avx512f-c64.c"
 +  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
++
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
++
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -15610,7 +15684,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rsum/gen/f32-rsum-avx512f-u64-acc4.c"
++    "src/src/f32-rsum/gen/f32-rsum-avx512f-u32-acc2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -15635,7 +15709,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rsum/gen/f32-rsum-avx512f-u64-acc4.c"
++    "src/src/f32-rsum/gen/f32-rsum-avx512f-u32-acc2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -15656,15 +15730,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-rsum_sse-no-sse2") {
++source_set("f32-rsum_sse2-no-sse3") {
 +  cflags = [
-+    "-mno-sse2",
-+    "-msse"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rsum/gen/f32-rsum-sse-u16-acc4.c"
++    "src/src/f32-rsum/gen/f32-rsum-sse2-u16-acc4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -15682,15 +15756,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-rsum_sse-no-sse2_standalone") {
++source_set("f32-rsum_sse2-no-sse3_standalone") {
 +  cflags = [
-+    "-mno-sse2",
-+    "-msse"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rsum/gen/f32-rsum-sse-u16-acc4.c"
++    "src/src/f32-rsum/gen/f32-rsum-sse2-u16-acc4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -15942,7 +16016,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-avx512f-rational-12-10-nr.c"
++    "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-avx512f-rational-12-10-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -15967,7 +16041,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-avx512f-rational-12-10-nr.c"
++    "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-avx512f-rational-12-10-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -16580,7 +16654,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vclamp/gen/f32-vclamp-avx-u16.c"
++    "src/src/f32-vclamp/gen/f32-vclamp-avx.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -16608,7 +16682,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vclamp/gen/f32-vclamp-avx-u16.c"
++    "src/src/f32-vclamp/gen/f32-vclamp-avx.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -16636,7 +16710,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vclamp/gen/f32-vclamp-avx512f-u16.c"
++    "src/src/f32-vclamp/gen/f32-vclamp-avx512f.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -16661,7 +16735,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vclamp/gen/f32-vclamp-avx512f-u16.c"
++    "src/src/f32-vclamp/gen/f32-vclamp-avx512f.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -16682,15 +16756,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vclamp_sse-no-sse2") {
++source_set("f32-vclamp_sse2-no-sse3") {
 +  cflags = [
-+    "-mno-sse2",
-+    "-msse"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vclamp/gen/f32-vclamp-sse-u8.c"
++    "src/src/f32-vclamp/gen/f32-vclamp-sse2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -16708,15 +16782,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vclamp_sse-no-sse2_standalone") {
++source_set("f32-vclamp_sse2-no-sse3_standalone") {
 +  cflags = [
-+    "-mno-sse2",
-+    "-msse"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vclamp/gen/f32-vclamp-sse-u8.c"
++    "src/src/f32-vclamp/gen/f32-vclamp-sse2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -16744,7 +16818,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vclamp/gen/f32-vclamp-scalar-u4.c"
++    "src/src/f32-vclamp/gen/f32-vclamp-scalar.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -16769,7 +16843,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vclamp/gen/f32-vclamp-scalar-u4.c"
++    "src/src/f32-vclamp/gen/f32-vclamp-scalar.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -17310,7 +17384,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vcos/gen/f32-vcos-avx512f-rational-5-4-nr.c"
++    "src/src/f32-vcos/gen/f32-vcos-avx512f-rational-5-4-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -17335,7 +17409,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vcos/gen/f32-vcos-avx512f-rational-5-4-nr.c"
++    "src/src/f32-vcos/gen/f32-vcos-avx512f-rational-5-4-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -17752,7 +17826,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u2.c",
 +    "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u4.c"
 +  ]
 +
@@ -17778,7 +17851,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u2.c",
 +    "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u4.c"
 +  ]
 +
@@ -18143,7 +18215,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vgelu/gen/f32-vgelu-avx512f-rational-12-10-nr.c"
++    "src/src/f32-vgelu/gen/f32-vgelu-avx512f-rational-12-10-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -18168,7 +18240,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vgelu/gen/f32-vgelu-avx512f-rational-12-10-nr.c"
++    "src/src/f32-vgelu/gen/f32-vgelu-avx512f-rational-12-10-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -18364,7 +18436,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vhswish/gen/f32-vhswish-avx-u16.c"
++    "src/src/f32-vhswish/gen/f32-vhswish-avx.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -18392,7 +18464,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vhswish/gen/f32-vhswish-avx-u16.c"
++    "src/src/f32-vhswish/gen/f32-vhswish-avx.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -18420,7 +18492,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vhswish/gen/f32-vhswish-avx512f-u16.c"
++    "src/src/f32-vhswish/gen/f32-vhswish-avx512f.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -18445,7 +18517,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vhswish/gen/f32-vhswish-avx512f-u16.c"
++    "src/src/f32-vhswish/gen/f32-vhswish-avx512f.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -18475,7 +18547,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vhswish/gen/f32-vhswish-fma3-u16.c"
++    "src/src/f32-vhswish/gen/f32-vhswish-fma3.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -18502,7 +18574,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vhswish/gen/f32-vhswish-fma3-u16.c"
++    "src/src/f32-vhswish/gen/f32-vhswish-fma3.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -18523,15 +18595,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vhswish_sse-no-sse2") {
++source_set("f32-vhswish_sse2-no-sse3") {
 +  cflags = [
-+    "-mno-sse2",
-+    "-msse"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vhswish/gen/f32-vhswish-sse-u8.c"
++    "src/src/f32-vhswish/gen/f32-vhswish-sse2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -18549,15 +18621,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vhswish_sse-no-sse2_standalone") {
++source_set("f32-vhswish_sse2-no-sse3_standalone") {
 +  cflags = [
-+    "-mno-sse2",
-+    "-msse"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vhswish/gen/f32-vhswish-sse-u8.c"
++    "src/src/f32-vhswish/gen/f32-vhswish-sse2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -18585,7 +18657,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vhswish/gen/f32-vhswish-scalar-u4.c"
++    "src/src/f32-vhswish/gen/f32-vhswish-scalar.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -18610,7 +18682,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vhswish/gen/f32-vhswish-scalar-u4.c"
++    "src/src/f32-vhswish/gen/f32-vhswish-scalar.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19289,14 +19361,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vrelu_x64") {
++source_set("f32-vrelu_avx-no-avx2-no-f16c-no-fma") {
 +  cflags = [
-+
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrelu/gen/f32-vrelu-scalar-u8.c"
++    "src/src/f32-vrelu/gen/f32-vrelu-avx.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19314,14 +19389,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vrelu_x64_standalone") {
++source_set("f32-vrelu_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
-+
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrelu/gen/f32-vrelu-scalar-u8.c"
++    "src/src/f32-vrelu/gen/f32-vrelu-avx.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19342,20 +19420,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vrnd_avx-no-avx2-no-f16c-no-fma") {
++source_set("f32-vrelu_avx512f") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mavx512f"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrnd/gen/f32-vrndd-avx-u16.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-avx-u16.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-avx-u16.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-avx-u16.c"
++    "src/src/f32-vrelu/gen/f32-vrelu-avx512f.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19373,20 +19445,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vrnd_avx-no-avx2-no-f16c-no-fma_standalone") {
++source_set("f32-vrelu_avx512f_standalone") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mavx512f"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrnd/gen/f32-vrndd-avx-u16.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-avx-u16.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-avx-u16.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-avx-u16.c"
++    "src/src/f32-vrelu/gen/f32-vrelu-avx512f.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19407,17 +19473,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vrnd_avx512f") {
++source_set("f32-vrelu_sse2-no-sse3") {
 +  cflags = [
-+    "-mavx512f"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrnd/gen/f32-vrndd-avx512f-u16.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-avx512f-u16.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-avx512f-u16.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-avx512f-u16.c"
++    "src/src/f32-vrelu/gen/f32-vrelu-sse2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19435,17 +19499,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vrnd_avx512f_standalone") {
++source_set("f32-vrelu_sse2-no-sse3_standalone") {
 +  cflags = [
-+    "-mavx512f"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrnd/gen/f32-vrndd-avx512f-u16.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-avx512f-u16.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-avx512f-u16.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-avx512f-u16.c"
++    "src/src/f32-vrelu/gen/f32-vrelu-sse2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19466,18 +19528,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vrnd_sse2-no-sse3") {
++source_set("f32-vrelu_x64") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrnd/gen/f32-vrndd-sse2-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-sse2-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-sse2-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-sse2-u8.c"
++    "src/src/f32-vrelu/gen/f32-vrelu-scalar.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19495,18 +19553,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vrnd_sse2-no-sse3_standalone") {
++source_set("f32-vrelu_x64_standalone") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrnd/gen/f32-vrndd-sse2-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-sse2-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-sse2-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-sse2-u8.c"
++    "src/src/f32-vrelu/gen/f32-vrelu-scalar.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19527,18 +19581,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vrnd_sse4.1-no-sse4.2") {
++source_set("f32-vrnd_avx-no-avx2-no-f16c-no-fma") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrnd/gen/f32-vrndd-sse41-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-sse41-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-sse41-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-sse41-u8.c"
++    "src/src/f32-vrnd/gen/f32-vrndd-avx-u16.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-avx-u16.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-avx-u16.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-avx-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19556,18 +19612,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vrnd_sse4.1-no-sse4.2_standalone") {
++source_set("f32-vrnd_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrnd/gen/f32-vrndd-sse41-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-sse41-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-sse41-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-sse41-u8.c"
++    "src/src/f32-vrnd/gen/f32-vrndd-avx-u16.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-avx-u16.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-avx-u16.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-avx-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19588,140 +19646,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vrnd_x64") {
-+  cflags = [
-+
-+  ]
-+
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u4.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+}
-+
-+# This is a target that cannot depend on //base.
-+source_set("f32-vrnd_x64_standalone") {
-+  cflags = [
-+
-+  ]
-+
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u4.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
-+
-+source_set("f32-vrsqrt_avx-no-avx2-no-f16c-no-fma") {
-+  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
-+  ]
-+
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx-rsqrt-u16.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+}
-+
-+# This is a target that cannot depend on //base.
-+source_set("f32-vrsqrt_avx-no-avx2-no-f16c-no-fma_standalone") {
-+  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
-+  ]
-+
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx-rsqrt-u16.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
-+
-+source_set("f32-vrsqrt_avx512f") {
++source_set("f32-vrnd_avx512f") {
 +  cflags = [
 +    "-mavx512f"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx512f-rsqrt-u32.c"
++    "src/src/f32-vrnd/gen/f32-vrndd-avx512f-u16.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-avx512f-u16.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-avx512f-u16.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-avx512f-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19739,14 +19674,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vrsqrt_avx512f_standalone") {
++source_set("f32-vrnd_avx512f_standalone") {
 +  cflags = [
 +    "-mavx512f"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx512f-rsqrt-u32.c"
++    "src/src/f32-vrnd/gen/f32-vrndd-avx512f-u16.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-avx512f-u16.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-avx512f-u16.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-avx512f-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19767,16 +19705,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vrsqrt_f16c-fma-no-avx2") {
++source_set("f32-vrnd_sse2-no-sse3") {
 +  cflags = [
-+    "-mf16c",
-+    "-mfma",
-+    "-mno-avx2"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-fma3-rsqrt-u16.c"
++    "src/src/f32-vrnd/gen/f32-vrndd-sse2-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-sse2-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-sse2-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-sse2-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19794,16 +19734,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vrsqrt_f16c-fma-no-avx2_standalone") {
++source_set("f32-vrnd_sse2-no-sse3_standalone") {
 +  cflags = [
-+    "-mf16c",
-+    "-mfma",
-+    "-mno-avx2"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-fma3-rsqrt-u16.c"
++    "src/src/f32-vrnd/gen/f32-vrndd-sse2-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-sse2-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-sse2-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-sse2-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19824,15 +19766,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vrsqrt_sse-no-sse2") {
++source_set("f32-vrnd_sse4.1-no-sse4.2") {
 +  cflags = [
-+    "-mno-sse2",
-+    "-msse"
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-sse-rsqrt-u8.c"
++    "src/src/f32-vrnd/gen/f32-vrndd-sse41-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-sse41-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-sse41-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-sse41-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19850,15 +19795,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vrsqrt_sse-no-sse2_standalone") {
++source_set("f32-vrnd_sse4.1-no-sse4.2_standalone") {
 +  cflags = [
-+    "-mno-sse2",
-+    "-msse"
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-sse-rsqrt-u8.c"
++    "src/src/f32-vrnd/gen/f32-vrndd-sse41-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-sse41-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-sse41-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-sse41-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19879,15 +19827,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vrsqrt_x64") {
++source_set("f32-vrnd_x64") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u4.c"
++    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19905,15 +19855,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vrsqrt_x64_standalone") {
++source_set("f32-vrnd_x64_standalone") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u4.c"
++    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19934,7 +19886,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vsigmoid_avx-no-avx2-no-f16c-no-fma") {
++source_set("f32-vrsqrt_avx-no-avx2-no-f16c-no-fma") {
 +  cflags = [
 +    "-mavx",
 +    "-mno-avx2",
@@ -19944,7 +19896,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx-rr2-p5-nr2-u16.c"
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx-rsqrt.c",
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx-sqrt.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19962,7 +19915,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsigmoid_avx-no-avx2-no-f16c-no-fma_standalone") {
++source_set("f32-vrsqrt_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
 +    "-mavx",
 +    "-mno-avx2",
@@ -19972,7 +19925,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx-rr2-p5-nr2-u16.c"
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx-rsqrt.c",
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx-sqrt.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -19993,14 +19947,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vsigmoid_avx512f") {
++source_set("f32-vrsqrt_avx512f") {
 +  cflags = [
 +    "-mavx512f"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx512f-rr2-lut32-p2-perm2-scalef-div-u64.c"
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx512f-rsqrt.c",
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx512f-sqrt.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20018,14 +19973,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsigmoid_avx512f_standalone") {
++source_set("f32-vrsqrt_avx512f_standalone") {
 +  cflags = [
 +    "-mavx512f"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx512f-rr2-lut32-p2-perm2-scalef-div-u64.c"
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx512f-rsqrt.c",
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx512f-sqrt.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20046,16 +20002,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vsigmoid_f16c-fma-avx2") {
++source_set("f32-vrsqrt_sse2-no-sse3") {
 +  cflags = [
-+    "-mavx2",
-+    "-mf16c",
-+    "-mfma"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx2-rr1-p5-div-u16.c"
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-sse2-sqrt.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20073,16 +20028,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsigmoid_f16c-fma-avx2_standalone") {
++source_set("f32-vrsqrt_sse2-no-sse3_standalone") {
 +  cflags = [
-+    "-mavx2",
-+    "-mf16c",
-+    "-mfma"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx2-rr1-p5-div-u16.c"
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-sse2-sqrt.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20103,15 +20057,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vsigmoid_sse2-no-sse3") {
++source_set("f32-vrsqrt_x64") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-sse2-rr2-lut64-p2-div-u8.c"
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-sqrt.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20129,15 +20083,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsigmoid_sse2-no-sse3_standalone") {
++source_set("f32-vrsqrt_x64_standalone") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-sse2-rr2-lut64-p2-div-u8.c"
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-sqrt.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20158,15 +20112,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vsigmoid_sse4.1-no-sse4.2") {
++source_set("f32-vsigmoid_avx-no-avx2-no-f16c-no-fma") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-sse41-rr2-lut64-p2-div-u8.c"
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx-rr2-p5-nr2-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20184,15 +20140,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsigmoid_sse4.1-no-sse4.2_standalone") {
++source_set("f32-vsigmoid_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-sse41-rr2-lut64-p2-div-u8.c"
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx-rr2-p5-nr2-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20213,14 +20171,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vsigmoid_x64") {
++source_set("f32-vsigmoid_avx512f") {
 +  cflags = [
-+
++    "-mavx512f"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c"
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx512f-rr2-lut32-p2-perm2-scalef-div-u64.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20238,14 +20196,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsigmoid_x64_standalone") {
++source_set("f32-vsigmoid_avx512f_standalone") {
 +  cflags = [
-+
++    "-mavx512f"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c"
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx512f-rr2-lut32-p2-perm2-scalef-div-u64.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20266,17 +20224,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vsin_avx-no-avx2-no-f16c-no-fma") {
++source_set("f32-vsigmoid_f16c-fma-avx2") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsin/gen/f32-vsin-avx-rational-5-4-div.c"
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx2-rr1-p5-div-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20294,17 +20251,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsin_avx-no-avx2-no-f16c-no-fma_standalone") {
++source_set("f32-vsigmoid_f16c-fma-avx2_standalone") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsin/gen/f32-vsin-avx-rational-5-4-div.c"
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx2-rr1-p5-div-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20325,14 +20281,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vsin_avx512f") {
++source_set("f32-vsigmoid_sse2-no-sse3") {
 +  cflags = [
-+    "-mavx512f"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsin/gen/f32-vsin-avx512f-rational-5-4-nr.c"
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-sse2-rr2-lut64-p2-div-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20350,14 +20307,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsin_avx512f_standalone") {
++source_set("f32-vsigmoid_sse2-no-sse3_standalone") {
 +  cflags = [
-+    "-mavx512f"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsin/gen/f32-vsin-avx512f-rational-5-4-nr.c"
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-sse2-rr2-lut64-p2-div-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20378,16 +20336,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vsin_f16c-fma-no-avx2") {
++source_set("f32-vsigmoid_sse4.1-no-sse4.2") {
 +  cflags = [
-+    "-mf16c",
-+    "-mfma",
-+    "-mno-avx2"
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsin/gen/f32-vsin-fma3-rational-5-4-div.c"
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-sse41-rr2-lut64-p2-div-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20405,16 +20362,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsin_f16c-fma-no-avx2_standalone") {
++source_set("f32-vsigmoid_sse4.1-no-sse4.2_standalone") {
 +  cflags = [
-+    "-mf16c",
-+    "-mfma",
-+    "-mno-avx2"
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsin/gen/f32-vsin-fma3-rational-5-4-div.c"
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-sse41-rr2-lut64-p2-div-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20435,15 +20391,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vsin_sse2-no-sse3") {
++source_set("f32-vsigmoid_x64") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsin/gen/f32-vsin-sse2-rational-5-4-div.c"
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20461,15 +20416,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsin_sse2-no-sse3_standalone") {
++source_set("f32-vsigmoid_x64_standalone") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsin/gen/f32-vsin-sse2-rational-5-4-div.c"
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20490,14 +20444,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vsin_x64") {
++source_set("f32-vsin_avx-no-avx2-no-f16c-no-fma") {
 +  cflags = [
-+
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsin/gen/f32-vsin-scalar-rational-5-4-div.c"
++    "src/src/f32-vsin/gen/f32-vsin-avx-rational-5-4-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20515,14 +20472,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsin_x64_standalone") {
++source_set("f32-vsin_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
-+
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsin/gen/f32-vsin-scalar-rational-5-4-div.c"
++    "src/src/f32-vsin/gen/f32-vsin-avx-rational-5-4-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20543,17 +20503,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vsqrt_avx-no-avx2-no-f16c-no-fma") {
++source_set("f32-vsin_avx512f") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mavx512f"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsqrt/gen/f32-vsqrt-avx-rsqrt-u16.c"
++    "src/src/f32-vsin/gen/f32-vsin-avx512f-rational-5-4-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20571,17 +20528,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsqrt_avx-no-avx2-no-f16c-no-fma_standalone") {
++source_set("f32-vsin_avx512f_standalone") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mavx512f"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsqrt/gen/f32-vsqrt-avx-rsqrt-u16.c"
++    "src/src/f32-vsin/gen/f32-vsin-avx512f-rational-5-4-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20602,14 +20556,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vsqrt_avx512f") {
++source_set("f32-vsin_f16c-fma-no-avx2") {
 +  cflags = [
-+    "-mavx512f"
++    "-mf16c",
++    "-mfma",
++    "-mno-avx2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsqrt/gen/f32-vsqrt-avx512f-rsqrt-u16.c"
++    "src/src/f32-vsin/gen/f32-vsin-fma3-rational-5-4-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20627,14 +20583,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsqrt_avx512f_standalone") {
++source_set("f32-vsin_f16c-fma-no-avx2_standalone") {
 +  cflags = [
-+    "-mavx512f"
++    "-mf16c",
++    "-mfma",
++    "-mno-avx2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsqrt/gen/f32-vsqrt-avx512f-rsqrt-u16.c"
++    "src/src/f32-vsin/gen/f32-vsin-fma3-rational-5-4-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20655,16 +20613,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vsqrt_f16c-fma-no-avx2") {
++source_set("f32-vsin_sse2-no-sse3") {
 +  cflags = [
-+    "-mf16c",
-+    "-mfma",
-+    "-mno-avx2"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsqrt/gen/f32-vsqrt-fma3-rsqrt-u16.c"
++    "src/src/f32-vsin/gen/f32-vsin-sse2-rational-5-4-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20682,16 +20639,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsqrt_f16c-fma-no-avx2_standalone") {
++source_set("f32-vsin_sse2-no-sse3_standalone") {
 +  cflags = [
-+    "-mf16c",
-+    "-mfma",
-+    "-mno-avx2"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsqrt/gen/f32-vsqrt-fma3-rsqrt-u16.c"
++    "src/src/f32-vsin/gen/f32-vsin-sse2-rational-5-4-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20712,15 +20668,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("f32-vsqrt_sse-no-sse2") {
++source_set("f32-vsin_x64") {
 +  cflags = [
-+    "-mno-sse2",
-+    "-msse"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsqrt/gen/f32-vsqrt-sse-rsqrt-u12.c"
++    "src/src/f32-vsin/gen/f32-vsin-scalar-rational-5-4-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20738,15 +20693,183 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsqrt_sse-no-sse2_standalone") {
++source_set("f32-vsin_x64_standalone") {
 +  cflags = [
-+    "-mno-sse2",
-+    "-msse"
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vsin/gen/f32-vsin-scalar-rational-5-4-div.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("f32-vsqrt_avx-no-avx2-no-f16c-no-fma") {
++  cflags = [
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vsqrt/gen/f32-vsqrt-avx-rsqrt.c",
++    "src/src/f32-vsqrt/gen/f32-vsqrt-avx-sqrt.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("f32-vsqrt_avx-no-avx2-no-f16c-no-fma_standalone") {
++  cflags = [
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsqrt/gen/f32-vsqrt-sse-rsqrt-u12.c"
++    "src/src/f32-vsqrt/gen/f32-vsqrt-avx-rsqrt.c",
++    "src/src/f32-vsqrt/gen/f32-vsqrt-avx-sqrt.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("f32-vsqrt_avx512f") {
++  cflags = [
++    "-mavx512f"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vsqrt/gen/f32-vsqrt-avx512f-rsqrt.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("f32-vsqrt_avx512f_standalone") {
++  cflags = [
++    "-mavx512f"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vsqrt/gen/f32-vsqrt-avx512f-rsqrt.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("f32-vsqrt_sse2-no-sse3") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vsqrt/gen/f32-vsqrt-sse2-sqrt.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("f32-vsqrt_sse2-no-sse3_standalone") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vsqrt/gen/f32-vsqrt-sse2-sqrt.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20774,7 +20897,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt-u1.c"
++    "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20799,7 +20922,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt-u1.c"
++    "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20886,7 +21009,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vtanh/gen/f32-vtanh-avx512f-rational-9-8-nr.c"
++    "src/src/f32-vtanh/gen/f32-vtanh-avx512f-rational-9-8-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -20911,7 +21034,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vtanh/gen/f32-vtanh-avx512f-rational-9-8-nr.c"
++    "src/src/f32-vtanh/gen/f32-vtanh-avx512f-rational-9-8-div.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -23865,7 +23988,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/include/xnnpack.h",
 +    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x2-minmax-scalar.c",
 +    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-2x2-minmax-scalar.c",
 +    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4-minmax-scalar.c"
 +  ]
 +
@@ -23893,7 +24015,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/include/xnnpack.h",
 +    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x2-minmax-scalar.c",
 +    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-2x2-minmax-scalar.c",
 +    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4-minmax-scalar.c"
 +  ]
 +
@@ -24373,7 +24494,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/include/xnnpack.h",
 +    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x2-minmax-scalar.c",
 +    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-2x2-minmax-scalar.c",
 +    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4-minmax-scalar.c"
 +  ]
 +
@@ -24401,7 +24521,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/include/xnnpack.h",
 +    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x2-minmax-scalar.c",
 +    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-2x2-minmax-scalar.c",
 +    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4-minmax-scalar.c"
 +  ]
 +
@@ -24732,10 +24851,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-dwconv/gen/qs8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
 +    "src/src/qs8-dwconv/gen/qs8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
 +
@@ -24762,10 +24879,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-dwconv/gen/qs8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
 +    "src/src/qs8-dwconv/gen/qs8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
 +
@@ -25142,7 +25257,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u1.c",
 +    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u4.c"
 +  ]
 +
@@ -25168,7 +25282,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u1.c",
 +    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u4.c"
 +  ]
 +
@@ -25747,13 +25860,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-lrintf.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
 +
@@ -25780,13 +25890,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-lrintf.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
 +
@@ -26386,9 +26493,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x2-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x2-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
 +
@@ -26414,9 +26519,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x2-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x2-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
 +
@@ -27016,9 +27119,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x2-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x2-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
 +
@@ -27044,9 +27145,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x2-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x2-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
 +
@@ -28815,7 +28914,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u1.c",
 +    "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c"
 +  ]
 +
@@ -28841,7 +28939,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u1.c",
 +    "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c"
 +  ]
 +
@@ -29151,8 +29248,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c",
-+    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-select-u4.c"
++    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -29177,8 +29273,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c",
-+    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-select-u4.c"
++    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -29643,68 +29738,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-dwconv_avx-no-avx2-no-f16c-no-fma") {
-+  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
-+  ]
-+
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p16c-minmax-fp32-avx-mul16.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-fp32-avx-mul16.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+}
-+
-+# This is a target that cannot depend on //base.
-+source_set("qu8-dwconv_avx-no-avx2-no-f16c-no-fma_standalone") {
-+  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
-+  ]
-+
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p16c-minmax-fp32-avx-mul16.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-fp32-avx-mul16.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
-+
-+source_set("qu8-dwconv_f16c-fma-avx2") {
++source_set("qs8-vprelu_f16c-fma-avx2") {
 +  cflags = [
 +    "-mavx2",
 +    "-mf16c",
@@ -29713,8 +29747,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p16c-minmax-fp32-avx2-mul32.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-fp32-avx2-mul32.c"
++    "src/src/qs8-vprelu/gen/qs8-vprelu-avx2-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -29732,7 +29765,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-dwconv_f16c-fma-avx2_standalone") {
++source_set("qs8-vprelu_f16c-fma-avx2_standalone") {
 +  cflags = [
 +    "-mavx2",
 +    "-mf16c",
@@ -29741,8 +29774,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p16c-minmax-fp32-avx2-mul32.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-fp32-avx2-mul32.c"
++    "src/src/qs8-vprelu/gen/qs8-vprelu-avx2-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -29763,83 +29795,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-dwconv_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
-+  cflags = [
-+    "-mavx512bw",
-+    "-mavx512cd",
-+    "-mavx512dq",
-+    "-mavx512f",
-+    "-mavx512vl",
-+    "-mf16c",
-+    "-mfma"
-+  ]
-+
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p32c-minmax-fp32-avx512skx-mul32.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p32c-minmax-fp32-avx512skx-mul32.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+}
-+
-+# This is a target that cannot depend on //base.
-+source_set("qu8-dwconv_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
++source_set("qs8-vprelu_x64") {
 +  cflags = [
-+    "-mavx512bw",
-+    "-mavx512cd",
-+    "-mavx512dq",
-+    "-mavx512f",
-+    "-mavx512vl",
-+    "-mf16c",
-+    "-mfma"
-+  ]
-+
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p32c-minmax-fp32-avx512skx-mul32.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p32c-minmax-fp32-avx512skx-mul32.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
 +
-+source_set("qu8-dwconv_sse2-no-sse3") {
-+  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-fp32-sse2-mul16.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p8c-minmax-fp32-sse2-mul16.c"
++    "src/src/qs8-vprelu/gen/qs8-vprelu-scalar-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -29857,16 +29820,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-dwconv_sse2-no-sse3_standalone") {
++source_set("qs8-vprelu_x64_standalone") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-fp32-sse2-mul16.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p8c-minmax-fp32-sse2-mul16.c"
++    "src/src/qs8-vprelu/gen/qs8-vprelu-scalar-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -29887,16 +29848,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-dwconv_sse4.1-no-sse4.2") {
++source_set("qs8-vpreluc_f16c-fma-avx2") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-fp32-sse41-mul16.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p8c-minmax-fp32-sse41-mul16.c"
++    "src/src/qs8-vpreluc/gen/qs8-vpreluc-avx2-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -29914,16 +29875,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-dwconv_sse4.1-no-sse4.2_standalone") {
++source_set("qs8-vpreluc_f16c-fma-avx2_standalone") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-fp32-sse41-mul16.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p8c-minmax-fp32-sse41-mul16.c"
++    "src/src/qs8-vpreluc/gen/qs8-vpreluc-avx2-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -29944,19 +29905,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-dwconv_x64") {
++source_set("qs8-vpreluc_x64") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
++    "src/src/qs8-vpreluc/gen/qs8-vpreluc-scalar-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -29974,19 +29930,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-dwconv_x64_standalone") {
++source_set("qs8-vpreluc_x64_standalone") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
++    "src/src/qs8-vpreluc/gen/qs8-vpreluc-scalar-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30007,66 +29958,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-f32-vcvt_avx-no-avx2-no-f16c-no-fma") {
-+  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
-+  ]
-+
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx-u32.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+}
-+
-+# This is a target that cannot depend on //base.
-+source_set("qu8-f32-vcvt_avx-no-avx2-no-f16c-no-fma_standalone") {
-+  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
-+  ]
-+
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx-u32.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
-+
-+source_set("qu8-f32-vcvt_f16c-fma-avx2") {
++source_set("qs8-vrpreluc_f16c-fma-avx2") {
 +  cflags = [
 +    "-mavx2",
 +    "-mf16c",
@@ -30075,7 +29967,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx2-u16.c"
++    "src/src/qs8-vrpreluc/gen/qs8-vrpreluc-avx2-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30093,7 +29985,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-f32-vcvt_f16c-fma-avx2_standalone") {
++source_set("qs8-vrpreluc_f16c-fma-avx2_standalone") {
 +  cflags = [
 +    "-mavx2",
 +    "-mf16c",
@@ -30102,7 +29994,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx2-u16.c"
++    "src/src/qs8-vrpreluc/gen/qs8-vrpreluc-avx2-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30123,80 +30015,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-f32-vcvt_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
++source_set("qs8-vrpreluc_x64") {
 +  cflags = [
-+    "-mavx512bw",
-+    "-mavx512cd",
-+    "-mavx512dq",
-+    "-mavx512f",
-+    "-mavx512vl",
-+    "-mf16c",
-+    "-mfma"
-+  ]
 +
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx512skx-u32.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+}
-+
-+# This is a target that cannot depend on //base.
-+source_set("qu8-f32-vcvt_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
-+  cflags = [
-+    "-mavx512bw",
-+    "-mavx512cd",
-+    "-mavx512dq",
-+    "-mavx512f",
-+    "-mavx512vl",
-+    "-mf16c",
-+    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx512skx-u32.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
-+
-+source_set("qu8-f32-vcvt_sse2-no-sse3") {
-+  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
-+  ]
-+
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-sse2-u32.c"
++    "src/src/qs8-vrpreluc/gen/qs8-vrpreluc-scalar-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30214,15 +30040,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-f32-vcvt_sse2-no-sse3_standalone") {
++source_set("qs8-vrpreluc_x64_standalone") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-sse2-u32.c"
++    "src/src/qs8-vrpreluc/gen/qs8-vrpreluc-scalar-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30243,117 +30068,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-f32-vcvt_sse4.1-no-sse4.2") {
-+  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
-+  ]
-+
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-sse41-u16.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+}
-+
-+# This is a target that cannot depend on //base.
-+source_set("qu8-f32-vcvt_sse4.1-no-sse4.2_standalone") {
-+  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
-+  ]
-+
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-sse41-u16.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
-+
-+source_set("qu8-f32-vcvt_x64") {
-+  cflags = [
-+
-+  ]
-+
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u1.c",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+}
-+
-+# This is a target that cannot depend on //base.
-+source_set("qu8-f32-vcvt_x64_standalone") {
-+  cflags = [
-+
-+  ]
-+
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u1.c",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
-+
-+source_set("qu8-gemm_avx-no-avx2-no-f16c-no-fma") {
++source_set("qu8-dwconv_avx-no-avx2-no-f16c-no-fma") {
 +  cflags = [
 +    "-mavx",
 +    "-mno-avx2",
@@ -30363,8 +30078,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-avx-ld128.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-2x4c8-minmax-fp32-avx-ld128.c"
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p16c-minmax-fp32-avx-mul16.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-fp32-avx-mul16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30382,7 +30097,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-gemm_avx-no-avx2-no-f16c-no-fma_standalone") {
++source_set("qu8-dwconv_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
 +    "-mavx",
 +    "-mno-avx2",
@@ -30392,8 +30107,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-avx-ld128.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-2x4c8-minmax-fp32-avx-ld128.c"
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p16c-minmax-fp32-avx-mul16.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-fp32-avx-mul16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30414,7 +30129,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-gemm_f16c-fma-avx2") {
++source_set("qu8-dwconv_f16c-fma-avx2") {
 +  cflags = [
 +    "-mavx2",
 +    "-mf16c",
@@ -30423,8 +30138,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x8c8-minmax-fp32-avx2.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-3x8c8-minmax-fp32-avx2.c"
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p16c-minmax-fp32-avx2-mul32.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-fp32-avx2-mul32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30442,7 +30157,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-gemm_f16c-fma-avx2_standalone") {
++source_set("qu8-dwconv_f16c-fma-avx2_standalone") {
 +  cflags = [
 +    "-mavx2",
 +    "-mf16c",
@@ -30451,8 +30166,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x8c8-minmax-fp32-avx2.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-3x8c8-minmax-fp32-avx2.c"
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p16c-minmax-fp32-avx2-mul32.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-fp32-avx2-mul32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30473,7 +30188,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
++source_set("qu8-dwconv_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 +  cflags = [
 +    "-mavx512bw",
 +    "-mavx512cd",
@@ -30486,8 +30201,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-7x16c8-minmax-fp32-avx512skx-prfm.c"
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p32c-minmax-fp32-avx512skx-mul32.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p32c-minmax-fp32-avx512skx-mul32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30505,7 +30220,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
++source_set("qu8-dwconv_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 +  cflags = [
 +    "-mavx512bw",
 +    "-mavx512cd",
@@ -30518,8 +30233,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-7x16c8-minmax-fp32-avx512skx-prfm.c"
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p32c-minmax-fp32-avx512skx-mul32.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p32c-minmax-fp32-avx512skx-mul32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30540,7 +30255,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-gemm_sse2-no-sse3") {
++source_set("qu8-dwconv_sse2-no-sse3") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
@@ -30548,8 +30263,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-sse2-ld64.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-3x4c8-minmax-fp32-sse2-ld64.c"
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-fp32-sse2-mul16.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p8c-minmax-fp32-sse2-mul16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30567,7 +30282,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-gemm_sse2-no-sse3_standalone") {
++source_set("qu8-dwconv_sse2-no-sse3_standalone") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
@@ -30575,40 +30290,77 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-sse2-ld64.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-3x4c8-minmax-fp32-sse2-ld64.c"
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-fp32-sse2-mul16.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p8c-minmax-fp32-sse2-mul16.c"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-velu/gen/f32-velu-avx-rr2-lut4-p4-perm-u32.c",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
 +
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-gemm_sse4.1-no-sse4.2") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-velu_avx-no-avx2-no-f16c-no-fma_standalone") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
++source_set("qu8-dwconv_sse4.1-no-sse4.2") {
 +  cflags = [
 +    "-mno-sse4.2",
 +    "-msse4.1"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-velu/gen/f32-velu-avx-rr2-lut4-p4-perm-u32.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-sse41-ld64.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-3x4c8-minmax-fp32-sse41-ld64.c"
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-fp32-sse41-mul16.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p8c-minmax-fp32-sse41-mul16.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -30624,7 +30376,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-gemm_sse4.1-no-sse4.2_standalone") {
++source_set("qu8-dwconv_sse4.1-no-sse4.2_standalone") {
 +  cflags = [
 +    "-mno-sse4.2",
 +    "-msse4.1"
@@ -30632,8 +30384,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-sse41-ld64.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-3x4c8-minmax-fp32-sse41-ld64.c"
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-fp32-sse41-mul16.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p8c-minmax-fp32-sse41-mul16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30646,27 +30398,47 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
 +
-+source_set("qu8-gemm_x64") {
++source_set("qu8-dwconv_x64") {
 +  cflags = [
-+
+ 
+-  source_set("f32-velu_avx512f") {
+-    cflags = [ "-mavx512f" ]
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-velu/gen/f32-velu-avx512f-rr1-p6-u64.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c"
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -30682,17 +30454,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-gemm_x64_standalone") {
++source_set("qu8-dwconv_x64_standalone") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c"
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30707,26 +30479,43 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
 +
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-igemm_avx-no-avx2-no-f16c-no-fma") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-velu_avx512f_standalone") {
+-    cflags = [ "-mavx512f" ]
++source_set("qu8-f32-vcvt_avx-no-avx2-no-f16c-no-fma") {
 +  cflags = [
 +    "-mavx",
 +    "-mno-avx2",
 +    "-mno-f16c",
 +    "-mno-fma"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-velu/gen/f32-velu-avx512f-rr1-p6-u64.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-avx-ld128.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-2x4c8-minmax-fp32-avx-ld128.c"
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx-u32.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -30742,7 +30531,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-igemm_avx-no-avx2-no-f16c-no-fma_standalone") {
++source_set("qu8-f32-vcvt_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
 +    "-mavx",
 +    "-mno-avx2",
@@ -30752,8 +30541,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-avx-ld128.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-2x4c8-minmax-fp32-avx-ld128.c"
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30766,27 +30554,50 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-igemm_f16c-fma-avx2") {
+ 
+-  source_set("f32-velu_f16c-fma-avx2") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++source_set("qu8-f32-vcvt_f16c-fma-avx2") {
 +  cflags = [
 +    "-mavx2",
 +    "-mf16c",
 +    "-mfma"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-velu/gen/f32-velu-avx2-rr1-lut4-p4-perm-u32.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x8c8-minmax-fp32-avx2.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-3x8c8-minmax-fp32-avx2.c"
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx2-u16.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -30802,7 +30613,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-igemm_f16c-fma-avx2_standalone") {
++source_set("qu8-f32-vcvt_f16c-fma-avx2_standalone") {
 +  cflags = [
 +    "-mavx2",
 +    "-mf16c",
@@ -30811,14 +30622,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x8c8-minmax-fp32-avx2.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-3x8c8-minmax-fp32-avx2.c"
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx2-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -30830,10 +30641,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-velu_f16c-fma-avx2_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++source_set("qu8-f32-vcvt_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 +  cflags = [
 +    "-mavx512bw",
 +    "-mavx512cd",
@@ -30843,13 +30661,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "-mf16c",
 +    "-mfma"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-velu/gen/f32-velu-avx2-rr1-lut4-p4-perm-u32.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-7x16c8-minmax-fp32-avx512skx-prfm.c"
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx512skx-u32.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -30865,7 +30696,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
++source_set("qu8-f32-vcvt_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 +  cflags = [
 +    "-mavx512bw",
 +    "-mavx512cd",
@@ -30878,8 +30709,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-7x16c8-minmax-fp32-avx512skx-prfm.c"
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx512skx-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30892,26 +30722,48 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-igemm_sse2-no-sse3") {
+ 
+-  source_set("f32-velu_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
++source_set("qu8-f32-vcvt_sse2-no-sse3") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-velu/gen/f32-velu-sse2-rr2-lut16-p3-u12.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-sse2-ld64.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-3x4c8-minmax-fp32-sse2-ld64.c"
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-sse2-u32.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -30927,7 +30779,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-igemm_sse2-no-sse3_standalone") {
++source_set("qu8-f32-vcvt_sse2-no-sse3_standalone") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
@@ -30935,8 +30787,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-sse2-ld64.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-3x4c8-minmax-fp32-sse2-ld64.c"
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-sse2-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -30951,24 +30802,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
 +
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-igemm_sse4.1-no-sse4.2") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-velu_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
++source_set("qu8-f32-vcvt_sse4.1-no-sse4.2") {
 +  cflags = [
 +    "-mno-sse4.2",
 +    "-msse4.1"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-velu/gen/f32-velu-sse2-rr2-lut16-p3-u12.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-sse41-ld64.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-3x4c8-minmax-fp32-sse41-ld64.c"
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-sse41-u16.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -30984,16 +30855,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-igemm_sse4.1-no-sse4.2_standalone") {
++source_set("qu8-f32-vcvt_sse4.1-no-sse4.2_standalone") {
 +  cflags = [
 +    "-mno-sse4.2",
 +    "-msse4.1"
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-sse41-ld64.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-3x4c8-minmax-fp32-sse41-ld64.c"
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-sse41-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31008,29 +30879,45 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
 +
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-igemm_x64") {
+ 
+-  source_set("f32-velu_x64") {
+-    cflags = []
++source_set("qu8-f32-vcvt_x64") {
 +  cflags = [
-+
-+  ]
-+
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c"
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u4.c",
+-    ]
 +  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -31042,17 +30929,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-igemm_x64_standalone") {
++source_set("qu8-f32-vcvt_x64_standalone") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c"
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31067,23 +30951,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
 +
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-rdsum_ssse3-no-sse4.1") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-velu_x64_standalone") {
+-    cflags = []
++source_set("qu8-gemm_avx-no-avx2-no-f16c-no-fma") {
 +  cflags = [
-+    "-mno-sse4.1",
-+    "-mssse3"
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u4.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rdsum/gen/qu8-rdsum-7p7x-ssse3-c64.c"
++    "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-avx-ld128.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-2x4c8-minmax-fp32-avx-ld128.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -31099,15 +31004,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-rdsum_ssse3-no-sse4.1_standalone") {
++source_set("qu8-gemm_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
-+    "-mno-sse4.1",
-+    "-mssse3"
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rdsum/gen/qu8-rdsum-7p7x-ssse3-c64.c"
++    "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-avx-ld128.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-2x4c8-minmax-fp32-avx-ld128.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31120,24 +31028,52 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-rdsum_x64") {
+ 
+-  source_set("f32-vexp_avx-no-avx2-no-f16c-no-fma") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
++source_set("qu8-gemm_f16c-fma-avx2") {
 +  cflags = [
-+
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vexp/gen/f32-vexp-avx-rational-3-2-div.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rdsum/gen/qu8-rdsum-scalar.c"
++    "src/src/qu8-gemm/gen/qu8-gemm-1x8c8-minmax-fp32-avx2.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-3x8c8-minmax-fp32-avx2.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -31153,14 +31089,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-rdsum_x64_standalone") {
++source_set("qu8-gemm_f16c-fma-avx2_standalone") {
 +  cflags = [
-+
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rdsum/gen/qu8-rdsum-scalar.c"
++    "src/src/qu8-gemm/gen/qu8-gemm-1x8c8-minmax-fp32-avx2.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-3x8c8-minmax-fp32-avx2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31175,24 +31114,52 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
 +
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-rsum_f16c-fma-avx2") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vexp_avx-no-avx2-no-f16c-no-fma_standalone") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
++source_set("qu8-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 +  cflags = [
-+    "-mavx2",
++    "-mavx512bw",
++    "-mavx512cd",
++    "-mavx512dq",
++    "-mavx512f",
++    "-mavx512vl",
 +    "-mf16c",
 +    "-mfma"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vexp/gen/f32-vexp-avx-rational-3-2-div.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rsum/gen/qu8-rsum-avx2-u64-acc2.c"
++    "src/src/qu8-gemm/gen/qu8-gemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-7x16c8-minmax-fp32-avx512skx-prfm.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -31208,16 +31175,21 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-rsum_f16c-fma-avx2_standalone") {
++source_set("qu8-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 +  cflags = [
-+    "-mavx2",
++    "-mavx512bw",
++    "-mavx512cd",
++    "-mavx512dq",
++    "-mavx512f",
++    "-mavx512vl",
 +    "-mf16c",
 +    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rsum/gen/qu8-rsum-avx2-u64-acc2.c"
++    "src/src/qu8-gemm/gen/qu8-gemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-7x16c8-minmax-fp32-avx512skx-prfm.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31230,25 +31202,46 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-rsum_sse2-no-sse3") {
+ 
+-  source_set("f32-vexp_avx512f") {
+-    cflags = [ "-mavx512f" ]
++source_set("qu8-gemm_sse2-no-sse3") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vexp/gen/f32-vexp-avx512f-rational-3-2-div.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rsum/gen/qu8-rsum-sse2-u32-acc2.c"
++    "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-sse2-ld64.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-3x4c8-minmax-fp32-sse2-ld64.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -31264,7 +31257,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-rsum_sse2-no-sse3_standalone") {
++source_set("qu8-gemm_sse2-no-sse3_standalone") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
@@ -31272,7 +31265,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rsum/gen/qu8-rsum-sse2-u32-acc2.c"
++    "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-sse2-ld64.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-3x4c8-minmax-fp32-sse2-ld64.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31287,20 +31281,120 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
 +
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
++}
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vexp_avx512f_standalone") {
+-    cflags = [ "-mavx512f" ]
++source_set("qu8-gemm_sse4.1-no-sse4.2") {
++  cflags = [
++    "-mno-sse4.2",
++    "-msse4.1"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vexp/gen/f32-vexp-avx512f-rational-3-2-div.c",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-sse41-ld64.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-3x4c8-minmax-fp32-sse41-ld64.c"
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
 +}
 +
-+source_set("qu8-rsum_x64") {
++# This is a target that cannot depend on //base.
++source_set("qu8-gemm_sse4.1-no-sse4.2_standalone") {
 +  cflags = [
++    "-mno-sse4.2",
++    "-msse4.1"
++  ]
 +
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-sse41-ld64.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-3x4c8-minmax-fp32-sse41-ld64.c"
 +  ]
 +
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  source_set("f32-vexp_f16c-fma-no-avx2") {
+-    cflags = [
+-      "-mf16c",
+-      "-mfma",
+-      "-mno-avx2",
+-    ]
++source_set("qu8-gemm_x64") {
++  cflags = [
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vexp/gen/f32-vexp-fma3-rational-3-2-div.c",
+-    ]
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rsum/gen/qu8-rsum-scalar-u4.c"
++    "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31318,14 +31412,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-rsum_x64_standalone") {
++source_set("qu8-gemm_x64_standalone") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rsum/gen/qu8-rsum-scalar-u4.c"
++    "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31340,25 +31435,48 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
 +
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vadd_avx-no-avx2-no-f16c-no-fma") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vexp_f16c-fma-no-avx2_standalone") {
+-    cflags = [
+-      "-mf16c",
+-      "-mfma",
+-      "-mno-avx2",
+-    ]
++source_set("qu8-igemm_avx-no-avx2-no-f16c-no-fma") {
 +  cflags = [
 +    "-mavx",
 +    "-mno-avx2",
 +    "-mno-f16c",
 +    "-mno-fma"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vexp/gen/f32-vexp-fma3-rational-3-2-div.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx-mul32-ld32-u8.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-avx-ld128.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-2x4c8-minmax-fp32-avx-ld128.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -31372,9 +31490,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  public_configs = [ ":xnnpack_config" ]
 +}
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vadd_avx-no-avx2-no-f16c-no-fma_standalone") {
++source_set("qu8-igemm_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
 +    "-mavx",
 +    "-mno-avx2",
@@ -31384,7 +31503,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx-mul32-ld32-u8.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-avx-ld128.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-2x4c8-minmax-fp32-avx-ld128.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31399,24 +31519,47 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
 +
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vadd_f16c-fma-avx2") {
+ 
+-  source_set("f32-vexp_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
++source_set("qu8-igemm_f16c-fma-avx2") {
 +  cflags = [
 +    "-mavx2",
 +    "-mf16c",
 +    "-mfma"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vexp/gen/f32-vexp-sse2-rational-3-2-div.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx2-mul32-ld64-u16.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-1x8c8-minmax-fp32-avx2.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-3x8c8-minmax-fp32-avx2.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -31432,7 +31575,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vadd_f16c-fma-avx2_standalone") {
++source_set("qu8-igemm_f16c-fma-avx2_standalone") {
 +  cflags = [
 +    "-mavx2",
 +    "-mf16c",
@@ -31441,7 +31584,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx2-mul32-ld64-u16.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-1x8c8-minmax-fp32-avx2.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-3x8c8-minmax-fp32-avx2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31454,15 +31598,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
 +
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vadd_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vexp_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
++source_set("qu8-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 +  cflags = [
 +    "-mavx512bw",
 +    "-mavx512cd",
@@ -31472,16 +31623,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "-mf16c",
 +    "-mfma"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vexp/gen/f32-vexp-sse2-rational-3-2-div.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx512skx-mul32-ld128-u16.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-7x16c8-minmax-fp32-avx512skx-prfm.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -31493,7 +31660,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vadd_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
++source_set("qu8-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 +  cflags = [
 +    "-mavx512bw",
 +    "-mavx512cd",
@@ -31506,13 +31673,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx512skx-mul32-ld128-u16.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-7x16c8-minmax-fp32-avx512skx-prfm.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -31524,20 +31695,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vadd_sse2-no-sse3") {
+ 
+-  source_set("f32-vexp_x64") {
+-    cflags = []
++source_set("qu8-igemm_sse2-no-sse3") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vexp/gen/f32-vexp-scalar-rational-3-2-div.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-sse2-mul16-ld64-u8.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-sse2-ld64.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-3x4c8-minmax-fp32-sse2-ld64.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -31553,7 +31741,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vadd_sse2-no-sse3_standalone") {
++source_set("qu8-igemm_sse2-no-sse3_standalone") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
@@ -31561,7 +31749,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-sse2-mul16-ld64-u8.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-sse2-ld64.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-3x4c8-minmax-fp32-sse2-ld64.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31574,25 +31763,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
 +
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vadd_sse4.1-no-sse4.2") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vexp_x64_standalone") {
+-    cflags = []
++source_set("qu8-igemm_sse4.1-no-sse4.2") {
 +  cflags = [
 +    "-mno-sse4.2",
 +    "-msse4.1"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vexp/gen/f32-vexp-scalar-rational-3-2-div.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-sse41-mul16-ld64-u8.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-sse41-ld64.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-3x4c8-minmax-fp32-sse41-ld64.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -31608,7 +31816,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vadd_sse4.1-no-sse4.2_standalone") {
++source_set("qu8-igemm_sse4.1-no-sse4.2_standalone") {
 +  cflags = [
 +    "-mno-sse4.2",
 +    "-msse4.1"
@@ -31616,9 +31824,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-sse41-mul16-ld64-u8.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-sse41-ld64.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-3x4c8-minmax-fp32-sse41-ld64.c"
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -31631,21 +31841,45 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
 +
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vadd_x64") {
+ 
+-  source_set("f32-vgelu_avx-no-avx2-no-f16c-no-fma") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
++source_set("qu8-igemm_x64") {
 +  cflags = [
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vgelu/gen/f32-vgelu-avx-rational-12-10-div.c",
+-    ]
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u1.c",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u4.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31663,15 +31897,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vadd_x64_standalone") {
++source_set("qu8-igemm_x64_standalone") {
 +  cflags = [
 +
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u1.c",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u4.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31689,22 +31924,42 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vaddc_avx-no-avx2-no-f16c-no-fma") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vgelu_avx-no-avx2-no-f16c-no-fma_standalone") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
++source_set("qu8-rdsum_ssse3-no-sse4.1") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mno-sse4.1",
++    "-mssse3"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vgelu/gen/f32-vgelu-avx-rational-12-10-div.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx-mul32-ld32-u8.c"
++    "src/src/qu8-rdsum/gen/qu8-rdsum-7p7x-ssse3-c64.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -31720,17 +31975,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vaddc_avx-no-avx2-no-f16c-no-fma_standalone") {
++source_set("qu8-rdsum_ssse3-no-sse4.1_standalone") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mno-sse4.1",
++    "-mssse3"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx-mul32-ld32-u8.c"
++    "src/src/qu8-rdsum/gen/qu8-rdsum-7p7x-ssse3-c64.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31743,24 +31996,42 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vaddc_f16c-fma-avx2") {
+ 
+-  source_set("f32-vgelu_avx512f") {
+-    cflags = [ "-mavx512f" ]
++source_set("qu8-rdsum_x64") {
 +  cflags = [
-+    "-mavx2",
-+    "-mf16c",
-+    "-mfma"
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vgelu/gen/f32-vgelu-avx512f-rational-12-10-div.c",
+-    ]
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx2-mul32-ld64-u16.c"
++    "src/src/qu8-rdsum/gen/qu8-rdsum-scalar.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31778,16 +32049,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vaddc_f16c-fma-avx2_standalone") {
++source_set("qu8-rdsum_x64_standalone") {
 +  cflags = [
-+    "-mavx2",
-+    "-mf16c",
-+    "-mfma"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx2-mul32-ld64-u16.c"
++    "src/src/qu8-rdsum/gen/qu8-rdsum-scalar.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31800,30 +32069,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
 +
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vaddc_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vgelu_avx512f_standalone") {
+-    cflags = [ "-mavx512f" ]
++source_set("qu8-rsum_f16c-fma-avx2") {
 +  cflags = [
-+    "-mavx512bw",
-+    "-mavx512cd",
-+    "-mavx512dq",
-+    "-mavx512f",
-+    "-mavx512vl",
++    "-mavx2",
 +    "-mf16c",
 +    "-mfma"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vgelu/gen/f32-vgelu-avx512f-rational-12-10-div.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx512skx-mul32-ld128-u16.c"
++    "src/src/qu8-rsum/gen/qu8-rsum-avx2-u64-acc2.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -31839,26 +32122,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vaddc_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
++source_set("qu8-rsum_f16c-fma-avx2_standalone") {
 +  cflags = [
-+    "-mavx512bw",
-+    "-mavx512cd",
-+    "-mavx512dq",
-+    "-mavx512f",
-+    "-mavx512vl",
++    "-mavx2",
 +    "-mf16c",
 +    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx512skx-mul32-ld128-u16.c"
++    "src/src/qu8-rsum/gen/qu8-rsum-avx2-u64-acc2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -31867,23 +32147,46 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
 +
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vaddc_sse2-no-sse3") {
+ 
+-  source_set("f32-vgelu_f16c-fma-no-avx2") {
+-    cflags = [
+-      "-mf16c",
+-      "-mfma",
+-      "-mno-avx2",
+-    ]
++source_set("qu8-rsum_sse2-no-sse3") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vgelu/gen/f32-vgelu-fma3-rational-12-10-div.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-sse2-mul16-ld64-u8.c"
++    "src/src/qu8-rsum/gen/qu8-rsum-sse2-u32-acc2.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -31899,7 +32202,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vaddc_sse2-no-sse3_standalone") {
++source_set("qu8-rsum_sse2-no-sse3_standalone") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
@@ -31907,7 +32210,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-sse2-mul16-ld64-u8.c"
++    "src/src/qu8-rsum/gen/qu8-rsum-sse2-u32-acc2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31920,23 +32223,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
 +
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vaddc_sse4.1-no-sse4.2") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vgelu_f16c-fma-no-avx2_standalone") {
+-    cflags = [
+-      "-mf16c",
+-      "-mfma",
+-      "-mno-avx2",
+-    ]
++source_set("qu8-rsum_x64") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vgelu/gen/f32-vgelu-fma3-rational-12-10-div.c",
+-    ]
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-sse41-mul16-ld64-u8.c"
++    "src/src/qu8-rsum/gen/qu8-rsum-scalar-u4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31954,15 +32278,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vaddc_sse4.1-no-sse4.2_standalone") {
++source_set("qu8-rsum_x64_standalone") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-sse41-mul16-ld64-u8.c"
++    "src/src/qu8-rsum/gen/qu8-rsum-scalar-u4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -31975,25 +32298,50 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vaddc_x64") {
+ 
+-  source_set("f32-vgelu_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
++source_set("qu8-vadd_avx-no-avx2-no-f16c-no-fma") {
 +  cflags = [
-+
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vgelu/gen/f32-vgelu-sse2-rational-12-10-div.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c"
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx-mul32-ld32-u8.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -32009,15 +32357,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vaddc_x64_standalone") {
++source_set("qu8-vadd_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
-+
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c"
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx-mul32-ld32-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32032,25 +32382,45 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
 +
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vcvt_avx-no-avx2-no-f16c-no-fma") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vgelu_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
++source_set("qu8-vadd_f16c-fma-avx2") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vgelu/gen/f32-vgelu-sse2-rational-12-10-div.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-avx-u32.c"
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx2-mul32-ld64-u16.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -32064,19 +32434,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  public_configs = [ ":xnnpack_config" ]
 +}
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vcvt_avx-no-avx2-no-f16c-no-fma_standalone") {
++source_set("qu8-vadd_f16c-fma-avx2_standalone") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-avx-u32.c"
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx2-mul32-ld64-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32091,24 +32461,47 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
 +
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vcvt_f16c-fma-avx2") {
+ 
+-  source_set("f32-vgelu_x64") {
+-    cflags = []
++source_set("qu8-vadd_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 +  cflags = [
-+    "-mavx2",
++    "-mavx512bw",
++    "-mavx512cd",
++    "-mavx512dq",
++    "-mavx512f",
++    "-mavx512vl",
 +    "-mf16c",
 +    "-mfma"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vgelu/gen/f32-vgelu-scalar-rational-12-10-div.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-avx2-u32.c"
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx512skx-mul32-ld128-u16.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -32124,16 +32517,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vcvt_f16c-fma-avx2_standalone") {
++source_set("qu8-vadd_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 +  cflags = [
-+    "-mavx2",
++    "-mavx512bw",
++    "-mavx512cd",
++    "-mavx512dq",
++    "-mavx512f",
++    "-mavx512vl",
 +    "-mf16c",
 +    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-avx2-u32.c"
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx512skx-mul32-ld128-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32148,23 +32545,41 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
 +
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vcvt_sse2-no-sse3") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vgelu_x64_standalone") {
+-    cflags = []
++source_set("qu8-vadd_sse2-no-sse3") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vgelu/gen/f32-vgelu-scalar-rational-12-10-div.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-sse2-u32.c"
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-sse2-mul16-ld64-u8.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -32175,12 +32590,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vcvt_sse2-no-sse3_standalone") {
++source_set("qu8-vadd_sse2-no-sse3_standalone") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
@@ -32188,7 +32604,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-sse2-u32.c"
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-sse2-mul16-ld64-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32201,29 +32617,54 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  public_configs = [ ":xnnpack_config" ]
 +
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vcvt_sse4.1-no-sse4.2") {
+ 
+-  source_set("f32-vhswish_avx-no-avx2-no-f16c-no-fma") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
++source_set("qu8-vadd_sse4.1-no-sse4.2") {
 +  cflags = [
 +    "-mno-sse4.2",
 +    "-msse4.1"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vhswish/gen/f32-vhswish-avx.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-sse41-u32.c"
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-sse41-mul16-ld64-u8.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -32235,7 +32676,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vcvt_sse4.1-no-sse4.2_standalone") {
++source_set("qu8-vadd_sse4.1-no-sse4.2_standalone") {
 +  cflags = [
 +    "-mno-sse4.2",
 +    "-msse4.1"
@@ -32243,7 +32684,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-sse41-u32.c"
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-sse41-mul16-ld64-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32261,20 +32702,43 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vcvt_ssse3-no-sse4.1") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vhswish_avx-no-avx2-no-f16c-no-fma_standalone") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
++source_set("qu8-vadd_x64") {
 +  cflags = [
-+    "-mno-sse4.1",
-+    "-mssse3"
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vhswish/gen/f32-vhswish-avx.c",
+-    ]
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-ssse3-u32.c"
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u1.c",
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u4.c"
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -32290,15 +32754,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vcvt_ssse3-no-sse4.1_standalone") {
++source_set("qu8-vadd_x64_standalone") {
 +  cflags = [
-+    "-mno-sse4.1",
-+    "-mssse3"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-ssse3-u32.c"
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u1.c",
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32313,23 +32777,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
 +
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vcvt_x64") {
+ 
+-  source_set("f32-vhswish_avx512f") {
+-    cflags = [ "-mavx512f" ]
++source_set("qu8-vaddc_avx-no-avx2-no-f16c-no-fma") {
 +  cflags = [
-+
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vhswish/gen/f32-vhswish-avx512f.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u1.c",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c"
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx-mul32-ld32-u8.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -32345,15 +32830,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vcvt_x64_standalone") {
++source_set("qu8-vaddc_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
-+
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u1.c",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c"
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx-mul32-ld32-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32368,29 +32855,46 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
 +
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
 +
-+source_set("qu8-vlrelu_avx-no-avx2-no-f16c-no-fma") {
++source_set("qu8-vaddc_f16c-fma-avx2") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
-+
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vhswish_avx512f_standalone") {
+-    cflags = [ "-mavx512f" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-avx-u32.c"
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx2-mul32-ld64-u16.c"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vhswish/gen/f32-vhswish-avx512f.c",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -32402,17 +32906,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vlrelu_avx-no-avx2-no-f16c-no-fma_standalone") {
++source_set("qu8-vaddc_f16c-fma-avx2_standalone") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-avx-u32.c"
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx2-mul32-ld64-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32425,26 +32928,54 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vlrelu_f16c-fma-avx2") {
+ 
+-  source_set("f32-vhswish_f16c-fma-no-avx2") {
+-    cflags = [
+-      "-mf16c",
+-      "-mfma",
+-      "-mno-avx2",
+-    ]
++source_set("qu8-vaddc_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 +  cflags = [
-+    "-mavx2",
++    "-mavx512bw",
++    "-mavx512cd",
++    "-mavx512dq",
++    "-mavx512f",
++    "-mavx512vl",
 +    "-mf16c",
 +    "-mfma"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vhswish/gen/f32-vhswish-fma3.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-avx2-u32.c"
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx512skx-mul32-ld128-u16.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -32460,16 +32991,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vlrelu_f16c-fma-avx2_standalone") {
++source_set("qu8-vaddc_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 +  cflags = [
-+    "-mavx2",
++    "-mavx512bw",
++    "-mavx512cd",
++    "-mavx512dq",
++    "-mavx512f",
++    "-mavx512vl",
 +    "-mf16c",
 +    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-avx2-u32.c"
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx512skx-mul32-ld128-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32484,23 +33019,45 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
 +
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vlrelu_sse2-no-sse3") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vhswish_f16c-fma-no-avx2_standalone") {
+-    cflags = [
+-      "-mf16c",
+-      "-mfma",
+-      "-mno-avx2",
+-    ]
++source_set("qu8-vaddc_sse2-no-sse3") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vhswish/gen/f32-vhswish-fma3.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-sse2-u32.c"
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-sse2-mul16-ld64-u8.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -32516,7 +33073,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vlrelu_sse2-no-sse3_standalone") {
++source_set("qu8-vaddc_sse2-no-sse3_standalone") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
@@ -32524,7 +33081,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-sse2-u32.c"
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-sse2-mul16-ld64-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32537,25 +33094,48 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vlrelu_sse4.1-no-sse4.2") {
+ 
+-  source_set("f32-vhswish_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
++source_set("qu8-vaddc_sse4.1-no-sse4.2") {
 +  cflags = [
 +    "-mno-sse4.2",
 +    "-msse4.1"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vhswish/gen/f32-vhswish-sse2.c",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-sse41-u32.c"
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-sse41-mul16-ld64-u8.c"
 +  ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -32571,7 +33151,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vlrelu_sse4.1-no-sse4.2_standalone") {
++source_set("qu8-vaddc_sse4.1-no-sse4.2_standalone") {
 +  cflags = [
 +    "-mno-sse4.2",
 +    "-msse4.1"
@@ -32579,13 +33159,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-sse41-u32.c"
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-sse41-mul16-ld64-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -32597,18 +33178,24 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
-+
-+source_set("qu8-vlrelu_ssse3-no-sse4.1") {
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vhswish_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
++source_set("qu8-vaddc_x64") {
 +  cflags = [
-+    "-mno-sse4.1",
-+    "-mssse3"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-ssse3-u32.c"
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32626,15 +33213,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vlrelu_ssse3-no-sse4.1_standalone") {
++source_set("qu8-vaddc_x64_standalone") {
 +  cflags = [
-+    "-mno-sse4.1",
-+    "-mssse3"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-ssse3-u32.c"
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32655,15 +33242,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-vlrelu_x64") {
++source_set("qu8-vcvt_avx-no-avx2-no-f16c-no-fma") {
 +  cflags = [
-+
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-select-u4.c"
++    "src/src/qu8-vcvt/gen/qu8-vcvt-avx-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32681,15 +33270,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vlrelu_x64_standalone") {
++source_set("qu8-vcvt_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
-+
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-select-u4.c"
++    "src/src/qu8-vcvt/gen/qu8-vcvt-avx-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32710,17 +33301,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-vmul_avx-no-avx2-no-f16c-no-fma") {
++source_set("qu8-vcvt_f16c-fma-avx2") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-avx-mul16-ld64-u16.c"
++    "src/src/qu8-vcvt/gen/qu8-vcvt-avx2-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32738,17 +33328,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vmul_avx-no-avx2-no-f16c-no-fma_standalone") {
++source_set("qu8-vcvt_f16c-fma-avx2_standalone") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-avx-mul16-ld64-u16.c"
++    "src/src/qu8-vcvt/gen/qu8-vcvt-avx2-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32769,7 +33358,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-vmul_sse2-no-sse3") {
++source_set("qu8-vcvt_sse2-no-sse3") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
@@ -32777,7 +33366,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-sse2-mul16-ld64-u8.c"
++    "src/src/qu8-vcvt/gen/qu8-vcvt-sse2-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32795,7 +33384,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vmul_sse2-no-sse3_standalone") {
++source_set("qu8-vcvt_sse2-no-sse3_standalone") {
 +  cflags = [
 +    "-mno-sse3",
 +    "-msse2"
@@ -32803,7 +33392,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-sse2-mul16-ld64-u8.c"
++    "src/src/qu8-vcvt/gen/qu8-vcvt-sse2-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32824,7 +33413,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-vmul_sse4.1-no-sse4.2") {
++source_set("qu8-vcvt_sse4.1-no-sse4.2") {
 +  cflags = [
 +    "-mno-sse4.2",
 +    "-msse4.1"
@@ -32832,7 +33421,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-sse41-mul16-ld64-u16.c"
++    "src/src/qu8-vcvt/gen/qu8-vcvt-sse41-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32850,7 +33439,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vmul_sse4.1-no-sse4.2_standalone") {
++source_set("qu8-vcvt_sse4.1-no-sse4.2_standalone") {
 +  cflags = [
 +    "-mno-sse4.2",
 +    "-msse4.1"
@@ -32858,7 +33447,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-sse41-mul16-ld64-u16.c"
++    "src/src/qu8-vcvt/gen/qu8-vcvt-sse41-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32879,14 +33468,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-vmul_x64") {
++source_set("qu8-vcvt_ssse3-no-sse4.1") {
 +  cflags = [
-+
++    "-mno-sse4.1",
++    "-mssse3"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c"
++    "src/src/qu8-vcvt/gen/qu8-vcvt-ssse3-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32904,14 +33494,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vmul_x64_standalone") {
++source_set("qu8-vcvt_ssse3-no-sse4.1_standalone") {
 +  cflags = [
-+
++    "-mno-sse4.1",
++    "-mssse3"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c"
++    "src/src/qu8-vcvt/gen/qu8-vcvt-ssse3-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32932,17 +33523,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-vmulc_avx-no-avx2-no-f16c-no-fma") {
++source_set("qu8-vcvt_x64") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-avx-mul16-ld64-u16.c"
++    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32960,17 +33548,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vmulc_avx-no-avx2-no-f16c-no-fma_standalone") {
++source_set("qu8-vcvt_x64_standalone") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-avx-mul16-ld64-u16.c"
++    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -32991,15 +33576,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-vmulc_sse2-no-sse3") {
++source_set("qu8-vlrelu_avx-no-avx2-no-f16c-no-fma") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-sse2-mul16-ld64-u8.c"
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-avx-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33017,15 +33604,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vmulc_sse2-no-sse3_standalone") {
++source_set("qu8-vlrelu_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-sse2-mul16-ld64-u8.c"
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-avx-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33046,15 +33635,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-vmulc_sse4.1-no-sse4.2") {
++source_set("qu8-vlrelu_f16c-fma-avx2") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-sse41-mul16-ld64-u16.c"
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-avx2-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33072,15 +33662,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vmulc_sse4.1-no-sse4.2_standalone") {
++source_set("qu8-vlrelu_f16c-fma-avx2_standalone") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-sse41-mul16-ld64-u16.c"
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-avx2-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33101,14 +33692,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("qu8-vmulc_x64") {
++source_set("qu8-vlrelu_sse2-no-sse3") {
 +  cflags = [
-+
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c"
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-sse2-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33126,14 +33718,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vmulc_x64_standalone") {
++source_set("qu8-vlrelu_sse2-no-sse3_standalone") {
 +  cflags = [
-+
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c"
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-sse2-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33154,16 +33747,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("reference_x64") {
++source_set("qu8-vlrelu_sse4.1-no-sse4.2") {
 +  cflags = [
-+
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/reference/binary-elementwise.cc",
-+    "src/src/reference/packing.cc",
-+    "src/src/reference/unary-elementwise.cc"
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-sse41-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33181,16 +33773,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("reference_x64_standalone") {
++source_set("qu8-vlrelu_sse4.1-no-sse4.2_standalone") {
 +  cflags = [
-+
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/reference/binary-elementwise.cc",
-+    "src/src/reference/packing.cc",
-+    "src/src/reference/unary-elementwise.cc"
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-sse41-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33211,15 +33802,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("s8-ibilinear_sse2-no-sse3") {
++source_set("qu8-vlrelu_ssse3-no-sse4.1") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-mno-sse4.1",
++    "-mssse3"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-ibilinear/gen/s8-ibilinear-sse2-c8.c"
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-ssse3-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33237,15 +33828,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("s8-ibilinear_sse2-no-sse3_standalone") {
++source_set("qu8-vlrelu_ssse3-no-sse4.1_standalone") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-mno-sse4.1",
++    "-mssse3"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-ibilinear/gen/s8-ibilinear-sse2-c8.c"
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-ssse3-u32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33266,15 +33857,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("s8-ibilinear_sse4.1-no-sse4.2") {
++source_set("qu8-vlrelu_x64") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-ibilinear/gen/s8-ibilinear-sse41-c16.c"
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33292,15 +33882,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("s8-ibilinear_sse4.1-no-sse4.2_standalone") {
++source_set("qu8-vlrelu_x64_standalone") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-ibilinear/gen/s8-ibilinear-sse41-c16.c"
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33321,14 +33910,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("s8-ibilinear_x64") {
++source_set("qu8-vmul_avx-no-avx2-no-f16c-no-fma") {
 +  cflags = [
-+
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-ibilinear/gen/s8-ibilinear-scalar-c1.c"
++    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-avx-mul16-ld64-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33346,14 +33938,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("s8-ibilinear_x64_standalone") {
++source_set("qu8-vmul_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
-+
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-ibilinear/gen/s8-ibilinear-scalar-c1.c"
++    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-avx-mul16-ld64-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33374,15 +33969,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("s8-maxpool_sse4.1-no-sse4.2") {
++source_set("qu8-vmul_sse2-no-sse3") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-sse41-u16.c"
++    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-sse2-mul16-ld64-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33400,15 +33995,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("s8-maxpool_sse4.1-no-sse4.2_standalone") {
++source_set("qu8-vmul_sse2-no-sse3_standalone") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-sse41-u16.c"
++    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-sse2-mul16-ld64-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33429,14 +34024,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("s8-maxpool_x64") {
++source_set("qu8-vmul_sse4.1-no-sse4.2") {
 +  cflags = [
-+
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-scalar-u1.c"
++    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-sse41-mul16-ld64-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33454,14 +34050,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("s8-maxpool_x64_standalone") {
++source_set("qu8-vmul_sse4.1-no-sse4.2_standalone") {
 +  cflags = [
-+
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-scalar-u1.c"
++    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-sse41-mul16-ld64-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33482,16 +34079,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("s8-rdminmax_sse4.1-no-sse4.2") {
++source_set("qu8-vmul_x64") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-sse41-c32.c",
-+    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-sse41-c32.c"
++    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33509,16 +34104,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("s8-rdminmax_sse4.1-no-sse4.2_standalone") {
++source_set("qu8-vmul_x64_standalone") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-sse41-c32.c",
-+    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-sse41-c32.c"
++    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33539,15 +34132,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("s8-rdminmax_x64") {
++source_set("qu8-vmulc_avx-no-avx2-no-f16c-no-fma") {
 +  cflags = [
-+
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-scalar-c2.c",
-+    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-scalar-c2.c"
++    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-avx-mul16-ld64-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33565,15 +34160,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("s8-rdminmax_x64_standalone") {
++source_set("qu8-vmulc_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
-+
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-scalar-c2.c",
-+    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-scalar-c2.c"
++    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-avx-mul16-ld64-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33594,17 +34191,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("s8-rminmax_sse4.1-no-sse4.2") {
++source_set("qu8-vmulc_sse2-no-sse3") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-rminmax/gen/s8-rmax-sse41-u32-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rmin-sse41-u32-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rminmax-sse41-u32-acc2.c"
++    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-sse2-mul16-ld64-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33622,17 +34217,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("s8-rminmax_sse4.1-no-sse4.2_standalone") {
++source_set("qu8-vmulc_sse2-no-sse3_standalone") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-rminmax/gen/s8-rmax-sse41-u32-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rmin-sse41-u32-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rminmax-sse41-u32-acc2.c"
++    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-sse2-mul16-ld64-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33653,16 +34246,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("s8-rminmax_x64") {
++source_set("qu8-vmulc_sse4.1-no-sse4.2") {
 +  cflags = [
-+
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-rminmax/gen/s8-rmax-scalar-u2-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rmin-scalar-u2-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rminmax-scalar-u2-acc2.c"
++    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-sse41-mul16-ld64-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33680,16 +34272,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("s8-rminmax_x64_standalone") {
++source_set("qu8-vmulc_sse4.1-no-sse4.2_standalone") {
 +  cflags = [
-+
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-rminmax/gen/s8-rmax-scalar-u2-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rmin-scalar-u2-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rminmax-scalar-u2-acc2.c"
++    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-sse41-mul16-ld64-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33710,16 +34301,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("s8-vclamp_f16c-fma-avx2") {
++source_set("qu8-vmulc_x64") {
 +  cflags = [
-+    "-mavx2",
-+    "-mf16c",
-+    "-mfma"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-vclamp/s8-vclamp-avx2-u128.c"
++    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33737,16 +34326,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("s8-vclamp_f16c-fma-avx2_standalone") {
++source_set("qu8-vmulc_x64_standalone") {
 +  cflags = [
-+    "-mavx2",
-+    "-mf16c",
-+    "-mfma"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-vclamp/s8-vclamp-avx2-u128.c"
++    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33767,20 +34354,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("s8-vclamp_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
++source_set("qu8-vprelu_f16c-fma-avx2") {
 +  cflags = [
-+    "-mavx512bw",
-+    "-mavx512cd",
-+    "-mavx512dq",
-+    "-mavx512f",
-+    "-mavx512vl",
++    "-mavx2",
 +    "-mf16c",
 +    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-vclamp/s8-vclamp-avx512skx-u256.c"
++    "src/src/qu8-vprelu/gen/qu8-vprelu-avx2-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33798,20 +34381,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("s8-vclamp_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
++source_set("qu8-vprelu_f16c-fma-avx2_standalone") {
 +  cflags = [
-+    "-mavx512bw",
-+    "-mavx512cd",
-+    "-mavx512dq",
-+    "-mavx512f",
-+    "-mavx512vl",
++    "-mavx2",
 +    "-mf16c",
 +    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-vclamp/s8-vclamp-avx512skx-u256.c"
++    "src/src/qu8-vprelu/gen/qu8-vprelu-avx2-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33832,15 +34411,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("s8-vclamp_sse2-no-sse3") {
++source_set("qu8-vprelu_x64") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-vclamp/s8-vclamp-sse2-u64.c"
++    "src/src/qu8-vprelu/gen/qu8-vprelu-scalar-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33858,15 +34436,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("s8-vclamp_sse2-no-sse3_standalone") {
++source_set("qu8-vprelu_x64_standalone") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-vclamp/s8-vclamp-sse2-u64.c"
++    "src/src/qu8-vprelu/gen/qu8-vprelu-scalar-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33887,15 +34464,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("s8-vclamp_sse4.1-no-sse4.2") {
++source_set("qu8-vpreluc_f16c-fma-avx2") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-vclamp/s8-vclamp-sse41-u64.c"
++    "src/src/qu8-vpreluc/gen/qu8-vpreluc-avx2-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33913,15 +34491,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("s8-vclamp_sse4.1-no-sse4.2_standalone") {
++source_set("qu8-vpreluc_f16c-fma-avx2_standalone") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-vclamp/s8-vclamp-sse41-u64.c"
++    "src/src/qu8-vpreluc/gen/qu8-vpreluc-avx2-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33942,14 +34521,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("s8-vclamp_x64") {
++source_set("qu8-vpreluc_x64") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-vclamp/s8-vclamp-scalar-u4.c"
++    "src/src/qu8-vpreluc/gen/qu8-vpreluc-scalar-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33967,14 +34546,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("s8-vclamp_x64_standalone") {
++source_set("qu8-vpreluc_x64_standalone") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-vclamp/s8-vclamp-scalar-u4.c"
++    "src/src/qu8-vpreluc/gen/qu8-vpreluc-scalar-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -33995,41 +34574,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("subgraph_x64") {
++source_set("qu8-vrpreluc_f16c-fma-avx2") {
 +  cflags = [
-+
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/subgraph/argmax-pooling-2d.c",
-+    "src/src/subgraph/average-pooling-2d.c",
-+    "src/src/subgraph/batch-matrix-multiply.c",
-+    "src/src/subgraph/binary.c",
-+    "src/src/subgraph/concatenate.c",
-+    "src/src/subgraph/convolution-2d.c",
-+    "src/src/subgraph/copy.c",
-+    "src/src/subgraph/deconvolution-2d.c",
-+    "src/src/subgraph/deprecated.c",
-+    "src/src/subgraph/depth-to-space-2d.c",
-+    "src/src/subgraph/depthwise-convolution-2d.c",
-+    "src/src/subgraph/even-split.c",
-+    "src/src/subgraph/fully-connected-sparse.c",
-+    "src/src/subgraph/fully-connected.c",
-+    "src/src/subgraph/max-pooling-2d.c",
-+    "src/src/subgraph/pack-lh.c",
-+    "src/src/subgraph/reshape-helpers.c",
-+    "src/src/subgraph/rope.c",
-+    "src/src/subgraph/softmax.c",
-+    "src/src/subgraph/space-to-depth-2d.c",
-+    "src/src/subgraph/static-constant-pad.c",
-+    "src/src/subgraph/static-reduce.c",
-+    "src/src/subgraph/static-resize-bilinear-2d.c",
-+    "src/src/subgraph/static-slice.c",
-+    "src/src/subgraph/static-transpose.c",
-+    "src/src/subgraph/unary.c",
-+    "src/src/subgraph/unpooling-2d.c",
-+    "src/src/subgraph/validation.c"
++    "src/src/qu8-vrpreluc/gen/qu8-vrpreluc-avx2-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34047,41 +34601,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("subgraph_x64_standalone") {
++source_set("qu8-vrpreluc_f16c-fma-avx2_standalone") {
 +  cflags = [
-+
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/subgraph/argmax-pooling-2d.c",
-+    "src/src/subgraph/average-pooling-2d.c",
-+    "src/src/subgraph/batch-matrix-multiply.c",
-+    "src/src/subgraph/binary.c",
-+    "src/src/subgraph/concatenate.c",
-+    "src/src/subgraph/convolution-2d.c",
-+    "src/src/subgraph/copy.c",
-+    "src/src/subgraph/deconvolution-2d.c",
-+    "src/src/subgraph/deprecated.c",
-+    "src/src/subgraph/depth-to-space-2d.c",
-+    "src/src/subgraph/depthwise-convolution-2d.c",
-+    "src/src/subgraph/even-split.c",
-+    "src/src/subgraph/fully-connected-sparse.c",
-+    "src/src/subgraph/fully-connected.c",
-+    "src/src/subgraph/max-pooling-2d.c",
-+    "src/src/subgraph/pack-lh.c",
-+    "src/src/subgraph/reshape-helpers.c",
-+    "src/src/subgraph/rope.c",
-+    "src/src/subgraph/softmax.c",
-+    "src/src/subgraph/space-to-depth-2d.c",
-+    "src/src/subgraph/static-constant-pad.c",
-+    "src/src/subgraph/static-reduce.c",
-+    "src/src/subgraph/static-resize-bilinear-2d.c",
-+    "src/src/subgraph/static-slice.c",
-+    "src/src/subgraph/static-transpose.c",
-+    "src/src/subgraph/unary.c",
-+    "src/src/subgraph/unpooling-2d.c",
-+    "src/src/subgraph/validation.c"
++    "src/src/qu8-vrpreluc/gen/qu8-vrpreluc-avx2-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34102,22 +34631,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("tables_x64") {
++source_set("qu8-vrpreluc_x64") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/tables/exp2-k-over-2048.c",
-+    "src/src/tables/exp2-k-over-64.c",
-+    "src/src/tables/exp2minus-k-over-16.c",
-+    "src/src/tables/exp2minus-k-over-2048.c",
-+    "src/src/tables/exp2minus-k-over-32.c",
-+    "src/src/tables/exp2minus-k-over-4.c",
-+    "src/src/tables/exp2minus-k-over-64.c",
-+    "src/src/tables/exp2minus-k-over-8.c",
-+    "src/src/tables/vlog.c"
++    "src/src/qu8-vrpreluc/gen/qu8-vrpreluc-scalar-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34135,22 +34656,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("tables_x64_standalone") {
++source_set("qu8-vrpreluc_x64_standalone") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/tables/exp2-k-over-2048.c",
-+    "src/src/tables/exp2-k-over-64.c",
-+    "src/src/tables/exp2minus-k-over-16.c",
-+    "src/src/tables/exp2minus-k-over-2048.c",
-+    "src/src/tables/exp2minus-k-over-32.c",
-+    "src/src/tables/exp2minus-k-over-4.c",
-+    "src/src/tables/exp2minus-k-over-64.c",
-+    "src/src/tables/exp2minus-k-over-8.c",
-+    "src/src/tables/vlog.c"
++    "src/src/qu8-vrpreluc/gen/qu8-vrpreluc-scalar-u8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34171,15 +34684,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("u8-ibilinear_sse2-no-sse3") {
++source_set("reference_x64") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-ibilinear/gen/u8-ibilinear-sse2-c8.c"
++    "src/src/reference/binary-elementwise.cc",
++    "src/src/reference/packing.cc",
++    "src/src/reference/unary-elementwise.cc"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34197,15 +34711,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("u8-ibilinear_sse2-no-sse3_standalone") {
++source_set("reference_x64_standalone") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-ibilinear/gen/u8-ibilinear-sse2-c8.c"
++    "src/src/reference/binary-elementwise.cc",
++    "src/src/reference/packing.cc",
++    "src/src/reference/unary-elementwise.cc"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34226,15 +34741,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("u8-ibilinear_sse4.1-no-sse4.2") {
++source_set("s8-ibilinear_sse2-no-sse3") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-ibilinear/gen/u8-ibilinear-sse41-c16.c"
++    "src/src/s8-ibilinear/gen/s8-ibilinear-sse2-c8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34252,15 +34767,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("u8-ibilinear_sse4.1-no-sse4.2_standalone") {
++source_set("s8-ibilinear_sse2-no-sse3_standalone") {
 +  cflags = [
-+    "-mno-sse4.2",
-+    "-msse4.1"
++    "-mno-sse3",
++    "-msse2"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-ibilinear/gen/u8-ibilinear-sse41-c16.c"
++    "src/src/s8-ibilinear/gen/s8-ibilinear-sse2-c8.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34281,14 +34796,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("u8-ibilinear_x64") {
++source_set("s8-ibilinear_sse4.1-no-sse4.2") {
 +  cflags = [
-+
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-ibilinear/gen/u8-ibilinear-scalar-c1.c"
++    "src/src/s8-ibilinear/gen/s8-ibilinear-sse41-c16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34306,14 +34822,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("u8-ibilinear_x64_standalone") {
++source_set("s8-ibilinear_sse4.1-no-sse4.2_standalone") {
 +  cflags = [
-+
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-ibilinear/gen/u8-ibilinear-scalar-c1.c"
++    "src/src/s8-ibilinear/gen/s8-ibilinear-sse41-c16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34334,14 +34851,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("u8-lut32norm_x64") {
++source_set("s8-ibilinear_x64") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-lut32norm/u8-lut32norm-scalar.c"
++    "src/src/s8-ibilinear/gen/s8-ibilinear-scalar-c1.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34359,14 +34876,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("u8-lut32norm_x64_standalone") {
++source_set("s8-ibilinear_x64_standalone") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-lut32norm/u8-lut32norm-scalar.c"
++    "src/src/s8-ibilinear/gen/s8-ibilinear-scalar-c1.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34387,15 +34904,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("u8-maxpool_sse2-no-sse3") {
++source_set("s8-maxpool_sse4.1-no-sse4.2") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-sse2-u16.c"
++    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-sse41-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34413,15 +34930,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("u8-maxpool_sse2-no-sse3_standalone") {
++source_set("s8-maxpool_sse4.1-no-sse4.2_standalone") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-sse2-u16.c"
++    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-sse41-u16.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34442,14 +34959,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("u8-maxpool_x64") {
++source_set("s8-maxpool_x64") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-scalar-u1.c"
++    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-scalar-u1.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34467,14 +34984,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("u8-maxpool_x64_standalone") {
++source_set("s8-maxpool_x64_standalone") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-scalar-u1.c"
++    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-scalar-u1.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34495,16 +35012,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("u8-rdminmax_sse2-no-sse3") {
++source_set("s8-rdminmax_sse4.1-no-sse4.2") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-sse2-c32.c",
-+    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-sse2-c32.c"
++    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-sse41-c32.c",
++    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-sse41-c32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34522,16 +35039,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("u8-rdminmax_sse2-no-sse3_standalone") {
++source_set("s8-rdminmax_sse4.1-no-sse4.2_standalone") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-sse2-c32.c",
-+    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-sse2-c32.c"
++    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-sse41-c32.c",
++    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-sse41-c32.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34552,15 +35069,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("u8-rdminmax_x64") {
++source_set("s8-rdminmax_x64") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-scalar-c2.c",
-+    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-scalar-c2.c"
++    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-scalar-c2.c",
++    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-scalar-c2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34578,15 +35095,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("u8-rdminmax_x64_standalone") {
++source_set("s8-rdminmax_x64_standalone") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-scalar-c2.c",
-+    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-scalar-c2.c"
++    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-scalar-c2.c",
++    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-scalar-c2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34607,17 +35124,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("u8-rminmax_sse2-no-sse3") {
++source_set("s8-rminmax_sse4.1-no-sse4.2") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-rminmax/gen/u8-rmax-sse2-u32-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rmin-sse2-u32-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rminmax-sse2-u32-acc2.c"
++    "src/src/s8-rminmax/gen/s8-rmax-sse41-u32-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rmin-sse41-u32-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rminmax-sse41-u32-acc2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34635,17 +35152,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("u8-rminmax_sse2-no-sse3_standalone") {
++source_set("s8-rminmax_sse4.1-no-sse4.2_standalone") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-mno-sse4.2",
++    "-msse4.1"
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-rminmax/gen/u8-rmax-sse2-u32-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rmin-sse2-u32-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rminmax-sse2-u32-acc2.c"
++    "src/src/s8-rminmax/gen/s8-rmax-sse41-u32-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rmin-sse41-u32-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rminmax-sse41-u32-acc2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34666,16 +35183,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  }
 +}
 +
-+source_set("u8-rminmax_x64") {
++source_set("s8-rminmax_x64") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-rminmax/gen/u8-rmax-scalar-u2-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rmin-scalar-u2-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rminmax-scalar-u2-acc2.c"
++    "src/src/s8-rminmax/gen/s8-rmax-scalar-u2-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rmin-scalar-u2-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rminmax-scalar-u2-acc2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34693,16 +35210,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("u8-rminmax_x64_standalone") {
++source_set("s8-rminmax_x64_standalone") {
 +  cflags = [
 +
 +  ]
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-rminmax/gen/u8-rmax-scalar-u2-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rmin-scalar-u2-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rminmax-scalar-u2-acc2.c"
++    "src/src/s8-rminmax/gen/s8-rmax-scalar-u2-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rmin-scalar-u2-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rminmax-scalar-u2-acc2.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34720,10 +35237,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-   }
++  }
 +}
 +
-+source_set("u8-vclamp_f16c-fma-avx2") {
++source_set("s8-vclamp_f16c-fma-avx2") {
 +  cflags = [
 +    "-mavx2",
 +    "-mf16c",
@@ -34732,7 +35249,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-vclamp/u8-vclamp-avx2-u128.c"
++    "src/src/s8-vclamp/s8-vclamp-avx2-u128.c"
 +  ]
 +
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34750,67 +35267,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
 +
 +# This is a target that cannot depend on //base.
-+source_set("u8-vclamp_f16c-fma-avx2_standalone") {
++source_set("s8-vclamp_f16c-fma-avx2_standalone") {
 +  cflags = [
 +    "-mavx2",
 +    "-mf16c",
 +    "-mfma"
 +  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("f32-velu_f16c-fma-avx2_standalone") {
--    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
--    ]
++
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-vclamp/u8-vclamp-avx2-u128.c"
++    "src/src/s8-vclamp/s8-vclamp-avx2-u128.c"
 +  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-velu/gen/f32-velu-avx2-rr1-lut4-p4-perm-u32.c",
--    ]
++
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
++
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
- 
--    public_configs = [ ":xnnpack_config" ]
++
 +  public_configs = [ ":xnnpack_config" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
++
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-   }
++  }
 +}
- 
--  source_set("f32-velu_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
-+source_set("u8-vclamp_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
++
++source_set("s8-vclamp_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 +  cflags = [
 +    "-mavx512bw",
 +    "-mavx512cd",
@@ -34820,16 +35307,2378 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "-mf16c",
 +    "-mfma"
 +  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-velu/gen/f32-velu-sse2-rr2-lut16-p3-u12.c",
--    ]
++
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-vclamp/u8-vclamp-avx512skx-u256.c"
++    "src/src/s8-vclamp/s8-vclamp-avx512skx-u256.c"
 +  ]
- 
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("s8-vclamp_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
++  cflags = [
++    "-mavx512bw",
++    "-mavx512cd",
++    "-mavx512dq",
++    "-mavx512f",
++    "-mavx512vl",
++    "-mf16c",
++    "-mfma"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/s8-vclamp/s8-vclamp-avx512skx-u256.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("s8-vclamp_sse2-no-sse3") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/s8-vclamp/s8-vclamp-sse2-u64.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("s8-vclamp_sse2-no-sse3_standalone") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/s8-vclamp/s8-vclamp-sse2-u64.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("s8-vclamp_sse4.1-no-sse4.2") {
++  cflags = [
++    "-mno-sse4.2",
++    "-msse4.1"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/s8-vclamp/s8-vclamp-sse41-u64.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("s8-vclamp_sse4.1-no-sse4.2_standalone") {
++  cflags = [
++    "-mno-sse4.2",
++    "-msse4.1"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/s8-vclamp/s8-vclamp-sse41-u64.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("s8-vclamp_x64") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/s8-vclamp/s8-vclamp-scalar-u4.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("s8-vclamp_x64_standalone") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/s8-vclamp/s8-vclamp-scalar-u4.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("subgraph_x64") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/subgraph/argmax-pooling-2d.c",
++    "src/src/subgraph/average-pooling-2d.c",
++    "src/src/subgraph/batch-matrix-multiply.c",
++    "src/src/subgraph/binary.c",
++    "src/src/subgraph/concatenate.c",
++    "src/src/subgraph/convolution-2d.c",
++    "src/src/subgraph/copy.c",
++    "src/src/subgraph/deconvolution-2d.c",
++    "src/src/subgraph/deprecated.c",
++    "src/src/subgraph/depth-to-space-2d.c",
++    "src/src/subgraph/depthwise-convolution-2d.c",
++    "src/src/subgraph/even-split.c",
++    "src/src/subgraph/fully-connected-sparse.c",
++    "src/src/subgraph/fully-connected.c",
++    "src/src/subgraph/max-pooling-2d.c",
++    "src/src/subgraph/pack-lh.c",
++    "src/src/subgraph/reshape-helpers.c",
++    "src/src/subgraph/rope.c",
++    "src/src/subgraph/softmax.c",
++    "src/src/subgraph/space-to-depth-2d.c",
++    "src/src/subgraph/static-constant-pad.c",
++    "src/src/subgraph/static-reduce.c",
++    "src/src/subgraph/static-resize-bilinear-2d.c",
++    "src/src/subgraph/static-slice.c",
++    "src/src/subgraph/static-transpose.c",
++    "src/src/subgraph/subgraph-utils.c",
++    "src/src/subgraph/unary.c",
++    "src/src/subgraph/unpooling-2d.c",
++    "src/src/subgraph/validation.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("subgraph_x64_standalone") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/subgraph/argmax-pooling-2d.c",
++    "src/src/subgraph/average-pooling-2d.c",
++    "src/src/subgraph/batch-matrix-multiply.c",
++    "src/src/subgraph/binary.c",
++    "src/src/subgraph/concatenate.c",
++    "src/src/subgraph/convolution-2d.c",
++    "src/src/subgraph/copy.c",
++    "src/src/subgraph/deconvolution-2d.c",
++    "src/src/subgraph/deprecated.c",
++    "src/src/subgraph/depth-to-space-2d.c",
++    "src/src/subgraph/depthwise-convolution-2d.c",
++    "src/src/subgraph/even-split.c",
++    "src/src/subgraph/fully-connected-sparse.c",
++    "src/src/subgraph/fully-connected.c",
++    "src/src/subgraph/max-pooling-2d.c",
++    "src/src/subgraph/pack-lh.c",
++    "src/src/subgraph/reshape-helpers.c",
++    "src/src/subgraph/rope.c",
++    "src/src/subgraph/softmax.c",
++    "src/src/subgraph/space-to-depth-2d.c",
++    "src/src/subgraph/static-constant-pad.c",
++    "src/src/subgraph/static-reduce.c",
++    "src/src/subgraph/static-resize-bilinear-2d.c",
++    "src/src/subgraph/static-slice.c",
++    "src/src/subgraph/static-transpose.c",
++    "src/src/subgraph/subgraph-utils.c",
++    "src/src/subgraph/unary.c",
++    "src/src/subgraph/unpooling-2d.c",
++    "src/src/subgraph/validation.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("tables_x64") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/tables/exp2-k-over-2048.c",
++    "src/src/tables/exp2-k-over-64.c",
++    "src/src/tables/exp2minus-k-over-16.c",
++    "src/src/tables/exp2minus-k-over-2048.c",
++    "src/src/tables/exp2minus-k-over-32.c",
++    "src/src/tables/exp2minus-k-over-4.c",
++    "src/src/tables/exp2minus-k-over-64.c",
++    "src/src/tables/exp2minus-k-over-8.c",
++    "src/src/tables/vlog.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("tables_x64_standalone") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/tables/exp2-k-over-2048.c",
++    "src/src/tables/exp2-k-over-64.c",
++    "src/src/tables/exp2minus-k-over-16.c",
++    "src/src/tables/exp2minus-k-over-2048.c",
++    "src/src/tables/exp2minus-k-over-32.c",
++    "src/src/tables/exp2minus-k-over-4.c",
++    "src/src/tables/exp2minus-k-over-64.c",
++    "src/src/tables/exp2minus-k-over-8.c",
++    "src/src/tables/vlog.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("u8-ibilinear_sse2-no-sse3") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-ibilinear/gen/u8-ibilinear-sse2-c8.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("u8-ibilinear_sse2-no-sse3_standalone") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-ibilinear/gen/u8-ibilinear-sse2-c8.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("u8-ibilinear_sse4.1-no-sse4.2") {
++  cflags = [
++    "-mno-sse4.2",
++    "-msse4.1"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-ibilinear/gen/u8-ibilinear-sse41-c16.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("u8-ibilinear_sse4.1-no-sse4.2_standalone") {
++  cflags = [
++    "-mno-sse4.2",
++    "-msse4.1"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-ibilinear/gen/u8-ibilinear-sse41-c16.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("u8-ibilinear_x64") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-ibilinear/gen/u8-ibilinear-scalar-c1.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("u8-ibilinear_x64_standalone") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-ibilinear/gen/u8-ibilinear-scalar-c1.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("u8-lut32norm_x64") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-lut32norm/u8-lut32norm-scalar.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("u8-lut32norm_x64_standalone") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-lut32norm/u8-lut32norm-scalar.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("u8-maxpool_sse2-no-sse3") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-sse2-u16.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("u8-maxpool_sse2-no-sse3_standalone") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-sse2-u16.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("u8-maxpool_x64") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-scalar-u1.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("u8-maxpool_x64_standalone") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-scalar-u1.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("u8-rdminmax_sse2-no-sse3") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-sse2-c32.c",
++    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-sse2-c32.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("u8-rdminmax_sse2-no-sse3_standalone") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-sse2-c32.c",
++    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-sse2-c32.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("u8-rdminmax_x64") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-scalar-c2.c",
++    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-scalar-c2.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("u8-rdminmax_x64_standalone") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-scalar-c2.c",
++    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-scalar-c2.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("u8-rminmax_sse2-no-sse3") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-rminmax/gen/u8-rmax-sse2-u32-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rmin-sse2-u32-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rminmax-sse2-u32-acc2.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("u8-rminmax_sse2-no-sse3_standalone") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-rminmax/gen/u8-rmax-sse2-u32-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rmin-sse2-u32-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rminmax-sse2-u32-acc2.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("u8-rminmax_x64") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-rminmax/gen/u8-rmax-scalar-u2-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rmin-scalar-u2-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rminmax-scalar-u2-acc2.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("u8-rminmax_x64_standalone") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-rminmax/gen/u8-rmax-scalar-u2-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rmin-scalar-u2-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rminmax-scalar-u2-acc2.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("u8-vclamp_f16c-fma-avx2") {
++  cflags = [
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-vclamp/u8-vclamp-avx2-u128.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("u8-vclamp_f16c-fma-avx2_standalone") {
++  cflags = [
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-vclamp/u8-vclamp-avx2-u128.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("u8-vclamp_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
++  cflags = [
++    "-mavx512bw",
++    "-mavx512cd",
++    "-mavx512dq",
++    "-mavx512f",
++    "-mavx512vl",
++    "-mf16c",
++    "-mfma"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-vclamp/u8-vclamp-avx512skx-u256.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("u8-vclamp_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
++  cflags = [
++    "-mavx512bw",
++    "-mavx512cd",
++    "-mavx512dq",
++    "-mavx512f",
++    "-mavx512vl",
++    "-mf16c",
++    "-mfma"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-vclamp/u8-vclamp-avx512skx-u256.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("u8-vclamp_sse2-no-sse3") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-vclamp/u8-vclamp-sse2-u64.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("u8-vclamp_sse2-no-sse3_standalone") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-vclamp/u8-vclamp-sse2-u64.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("u8-vclamp_x64") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-vclamp/u8-vclamp-scalar-u4.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("u8-vclamp_x64_standalone") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-vclamp/u8-vclamp-scalar-u4.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x16-packw_f16c-fma-avx2") {
++  cflags = [
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x16-packw/gen/x16-packw-x16-gemm-goi-avx2-u16-prfm.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x16-packw_f16c-fma-avx2_standalone") {
++  cflags = [
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x16-packw/gen/x16-packw-x16-gemm-goi-avx2-u16-prfm.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x16-transposec_f16c-fma-avx2") {
++  cflags = [
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x16-transposec/gen/x16-transposec-16x16-reuse-switch-avx2.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x16-transposec_f16c-fma-avx2_standalone") {
++  cflags = [
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x16-transposec/gen/x16-transposec-16x16-reuse-switch-avx2.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x16-transposec_sse2-no-sse3") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x16-transposec/gen/x16-transposec-8x8-reuse-multi-sse2.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x16-transposec_sse2-no-sse3_standalone") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x16-transposec/gen/x16-transposec-8x8-reuse-multi-sse2.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x16-transposec_x64") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x16-transposec_x64_standalone") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x16-x32-packw_x64") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-gio-scalar.c",
++    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-goi-scalar.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x16-x32-packw_x64_standalone") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-gio-scalar.c",
++    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-goi-scalar.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x24-transposec_ssse3-no-sse4.1") {
++  cflags = [
++    "-mno-sse4.1",
++    "-mssse3"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x24-transposec/x24-transposec-4x4-ssse3.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x24-transposec_ssse3-no-sse4.1_standalone") {
++  cflags = [
++    "-mno-sse4.1",
++    "-mssse3"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x24-transposec/x24-transposec-4x4-ssse3.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x24-transposec_x64") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x24-transposec/gen/x24-transposec-1x2-scalar.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x24-transposec_x64_standalone") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x24-transposec/gen/x24-transposec-1x2-scalar.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x32-packw_avx-no-avx2-no-f16c-no-fma") {
++  cflags = [
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-packw/gen/x32-packw-x16-gemm-gio-avx-u8.c",
++    "src/src/x32-packw/gen/x32-packw-x16-gemm-goi-avx-u4.c",
++    "src/src/x32-packw/gen/x32-packw-x16s4-gemm-goi-avx-u4.c",
++    "src/src/x32-packw/gen/x32-packw-x8-gemm-gio-avx-u8.c",
++    "src/src/x32-packw/gen/x32-packw-x8-gemm-goi-avx-u4.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x32-packw_avx-no-avx2-no-f16c-no-fma_standalone") {
++  cflags = [
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-packw/gen/x32-packw-x16-gemm-gio-avx-u8.c",
++    "src/src/x32-packw/gen/x32-packw-x16-gemm-goi-avx-u4.c",
++    "src/src/x32-packw/gen/x32-packw-x16s4-gemm-goi-avx-u4.c",
++    "src/src/x32-packw/gen/x32-packw-x8-gemm-gio-avx-u8.c",
++    "src/src/x32-packw/gen/x32-packw-x8-gemm-goi-avx-u4.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x32-packw_avx512f") {
++  cflags = [
++    "-mavx512f"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-packw/gen/x32-packw-x16-gemm-gio-avx512f-u8.c",
++    "src/src/x32-packw/gen/x32-packw-x16-gemm-goi-avx512f-u4-prfm.c",
++    "src/src/x32-packw/gen/x32-packw-x32-gemm-gio-avx512f-u8.c",
++    "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-avx512f-u4-prfm.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x32-packw_avx512f_standalone") {
++  cflags = [
++    "-mavx512f"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-packw/gen/x32-packw-x16-gemm-gio-avx512f-u8.c",
++    "src/src/x32-packw/gen/x32-packw-x16-gemm-goi-avx512f-u4-prfm.c",
++    "src/src/x32-packw/gen/x32-packw-x32-gemm-gio-avx512f-u8.c",
++    "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-avx512f-u4-prfm.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x32-packw_sse2-no-sse3") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-packw/gen/x32-packw-x2c4-gemm-goi-sse2-u4.c",
++    "src/src/x32-packw/gen/x32-packw-x8-gemm-goi-sse2-u4.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x32-packw_sse2-no-sse3_standalone") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-packw/gen/x32-packw-x2c4-gemm-goi-sse2-u4.c",
++    "src/src/x32-packw/gen/x32-packw-x8-gemm-goi-sse2-u4.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x32-packw_x64") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-packw/gen/x32-packw-x2-gemm-gio-scalar.c",
++    "src/src/x32-packw/gen/x32-packw-x2-gemm-goi-scalar-float-u4.c",
++    "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-scalar-int-u2.c",
++    "src/src/x32-packw/gen/x32-packw-x4-gemm-gio-scalar.c",
++    "src/src/x32-packw/gen/x32-packw-x4-gemm-goi-scalar-float-u4.c",
++    "src/src/x32-packw/gen/x32-packw-x64-gemm-goi-scalar-int-u2.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x32-packw_x64_standalone") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-packw/gen/x32-packw-x2-gemm-gio-scalar.c",
++    "src/src/x32-packw/gen/x32-packw-x2-gemm-goi-scalar-float-u4.c",
++    "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-scalar-int-u2.c",
++    "src/src/x32-packw/gen/x32-packw-x4-gemm-gio-scalar.c",
++    "src/src/x32-packw/gen/x32-packw-x4-gemm-goi-scalar-float-u4.c",
++    "src/src/x32-packw/gen/x32-packw-x64-gemm-goi-scalar-int-u2.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x32-transposec_avx-no-avx2-no-f16c-no-fma") {
++  cflags = [
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-transposec/gen/x32-transposec-8x8-reuse-multi-avx.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x32-transposec_avx-no-avx2-no-f16c-no-fma_standalone") {
++  cflags = [
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-transposec/gen/x32-transposec-8x8-reuse-multi-avx.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x32-transposec_sse-no-sse2") {
++  cflags = [
++    "-mno-sse2",
++    "-msse"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-transposec/x32-transposec-4x4-sse.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x32-transposec_sse-no-sse2_standalone") {
++  cflags = [
++    "-mno-sse2",
++    "-msse"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-transposec/x32-transposec-4x4-sse.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x32-transposec_x64") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-transposec/gen/x32-transposec-2x4-scalar-int.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x32-transposec_x64_standalone") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-transposec/gen/x32-transposec-2x4-scalar-int.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x32-unpool_sse2-no-sse3") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-unpool/x32-unpool-sse2.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x32-unpool_sse2-no-sse3_standalone") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-unpool/x32-unpool-sse2.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x32-unpool_x64") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-unpool/x32-unpool-scalar.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x32-unpool_x64_standalone") {
++  cflags = [
++
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-unpool/x32-unpool-scalar.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x64-transposec_avx-no-avx2-no-f16c-no-fma") {
++  cflags = [
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x64-transposec/gen/x64-transposec-4x4-reuse-multi-avx.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++}
++
++# This is a target that cannot depend on //base.
++source_set("x64-transposec_avx-no-avx2-no-f16c-no-fma_standalone") {
++  cflags = [
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x64-transposec/gen/x64-transposec-4x4-reuse-multi-avx.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
++  public_configs = [ ":xnnpack_config" ]
++
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
++
++source_set("x64-transposec_sse2-no-sse3") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
++
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x64-transposec/gen/x64-transposec-2x2-multi-mov-sse2.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vhswish/gen/f32-vhswish-sse2.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    public_configs = [ ":xnnpack_config" ]
++# This is a target that cannot depend on //base.
++source_set("x64-transposec_sse2-no-sse3_standalone") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x64-transposec/gen/x64-transposec-2x2-multi-mov-sse2.c"
++  ]
+ 
+-  source_set("f32-vhswish_x64") {
+-    cflags = []
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vhswish/gen/f32-vhswish-scalar.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vhswish_x64_standalone") {
+-    cflags = []
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vhswish/gen/f32-vhswish-scalar.c",
+-    ]
++source_set("x64-transposec_x64") {
++  cflags = [
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c"
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  source_set("f32-vlog_avx512f") {
+-    cflags = [ "-mavx512f" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vlog/gen/f32-vlog-avx512f-rational-3-3-div.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("x64-transposec_x64_standalone") {
++  cflags = [
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vlog_avx512f_standalone") {
+-    cflags = [ "-mavx512f" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vlog/gen/f32-vlog-avx512f-rational-3-3-div.c",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  source_set("f32-vlog_f16c-fma-avx2") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++source_set("x8-lut_avx-no-avx2-no-f16c-no-fma") {
++  cflags = [
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vlog/gen/f32-vlog-avx2-rational-3-3-div.c",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-lut/gen/x8-lut-avx-u64.c"
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -34854,28 +37703,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-velu_sse2-no-sse3_standalone") {
+-  source_set("f32-vlog_f16c-fma-avx2_standalone") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-velu/gen/f32-velu-sse2-rr2-lut16-p3-u12.c",
+-      "src/src/f32-vlog/gen/f32-vlog-avx2-rational-3-3-div.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("u8-vclamp_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
++source_set("x8-lut_avx-no-avx2-no-f16c-no-fma_standalone") {
 +  cflags = [
-+    "-mavx512bw",
-+    "-mavx512cd",
-+    "-mavx512dq",
-+    "-mavx512f",
-+    "-mavx512vl",
-+    "-mf16c",
-+    "-mfma"
++    "-mavx",
++    "-mno-avx2",
++    "-mno-f16c",
++    "-mno-fma"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -34890,7 +37737,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-vclamp/u8-vclamp-avx512skx-u256.c"
++    "src/src/x8-lut/gen/x8-lut-avx-u64.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -34909,14 +37756,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--  source_set("f32-velu_x64") {
--    cflags = []
+-  source_set("f32-vlog_f16c-fma-no-avx2") {
+-    cflags = [
+-      "-mf16c",
+-      "-mfma",
+-      "-mno-avx2",
+-    ]
 +  public_configs = [ ":xnnpack_config" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u2.c",
--      "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u4.c",
+-      "src/src/f32-vlog/gen/f32-vlog-fma3-rational-3-3-div.c",
 -    ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
@@ -34933,30 +37783,34 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("u8-vclamp_sse2-no-sse3") {
++source_set("x8-lut_f16c-fma-avx2") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-vclamp/u8-vclamp-sse2-u64.c"
++    "src/src/x8-lut/gen/x8-lut-avx2-u128.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-velu_x64_standalone") {
--    cflags = []
+-  source_set("f32-vlog_f16c-fma-no-avx2_standalone") {
+-    cflags = [
+-      "-mf16c",
+-      "-mfma",
+-      "-mno-avx2",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u2.c",
--      "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u4.c",
+-      "src/src/f32-vlog/gen/f32-vlog-fma3-rational-3-3-div.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -34980,10 +37834,11 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("u8-vclamp_sse2-no-sse3_standalone") {
++source_set("x8-lut_f16c-fma-avx2_standalone") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -34992,15 +37847,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-vclamp/u8-vclamp-sse2-u64.c"
++    "src/src/x8-lut/gen/x8-lut-avx2-u128.c"
 +  ]
  
--  source_set("f32-vexp_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("f32-vlog_sse2-no-sse3") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -35008,7 +37861,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vexp/gen/f32-vexp-avx-rational-3-2-div.c",
+-      "src/src/f32-vlog/gen/f32-vlog-sse2-rational-3-3-div.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -35036,20 +37889,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vexp_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("f32-vlog_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vexp/gen/f32-vexp-avx-rational-3-2-div.c",
+-      "src/src/f32-vlog/gen/f32-vlog-sse2-rational-3-3-div.c",
 -    ]
-+source_set("u8-vclamp_x64") {
++source_set("x8-lut_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 +  cflags = [
++    "-mavx512bw",
++    "-mavx512cd",
++    "-mavx512dq",
++    "-mavx512f",
++    "-mavx512vl",
++    "-mf16c",
++    "-mfma"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -35061,38 +37920,144 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-vclamp/u8-vclamp-scalar-u4.c"
++    "src/src/x8-lut/gen/x8-lut-avx512skx-vpshufb-u64.c"
 +  ]
  
+-    public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-  source_set("f32-vlog_x64") {
+-    cflags = []
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vlog/gen/f32-vlog-scalar-rational-3-3-div.c",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("x8-lut_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
++  cflags = [
++    "-mavx512bw",
++    "-mavx512cd",
++    "-mavx512dq",
++    "-mavx512f",
++    "-mavx512vl",
++    "-mf16c",
++    "-mfma"
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-lut/gen/x8-lut-avx512skx-vpshufb-u64.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("f32-vexp_avx512f") {
--    cflags = [ "-mavx512f" ]
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vlog_x64_standalone") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vexp/gen/f32-vexp-avx512f-rational-3-2-div.c",
+-      "src/src/f32-vlog/gen/f32-vlog-scalar-rational-3-3-div.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
 +}
  
+-    public_configs = [ ":xnnpack_config" ]
++source_set("x8-lut_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vbmi") {
++  cflags = [
++    "-mavx512bw",
++    "-mavx512cd",
++    "-mavx512dq",
++    "-mavx512f",
++    "-mavx512vbmi",
++    "-mavx512vl",
++    "-mf16c",
++    "-mfma"
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-lut/gen/x8-lut-avx512vbmi-vpermx2b-u128.c"
++  ]
+ 
+-  source_set("f32-vlrelu_avx-no-avx2-no-f16c-no-fma") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vlrelu/gen/f32-vlrelu-avx-u16.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -35103,25 +38068,40 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("u8-vclamp_x64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
++# This is a target that cannot depend on //base.
++source_set("x8-lut_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vbmi_standalone") {
++  cflags = [
++    "-mavx512bw",
++    "-mavx512cd",
++    "-mavx512dq",
++    "-mavx512f",
++    "-mavx512vbmi",
++    "-mavx512vl",
++    "-mf16c",
++    "-mfma"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vexp_avx512f_standalone") {
--    cflags = [ "-mavx512f" ]
+-  source_set("f32-vlrelu_avx-no-avx2-no-f16c-no-fma_standalone") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-vclamp/u8-vclamp-scalar-u4.c"
++    "src/src/x8-lut/gen/x8-lut-avx512vbmi-vpermx2b-u128.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vexp/gen/f32-vexp-avx512f-rational-3-2-div.c",
+-      "src/src/f32-vlrelu/gen/f32-vlrelu-avx-u16.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -35155,26 +38135,15 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("f32-vexp_f16c-fma-no-avx2") {
--    cflags = [
--      "-mf16c",
--      "-mfma",
--      "-mno-avx2",
--    ]
-+source_set("x16-packw_f16c-fma-avx2") {
+-  source_set("f32-vlrelu_avx512f") {
+-    cflags = [ "-mavx512f" ]
++source_set("x8-lut_x64") {
 +  cflags = [
-+    "-mavx2",
-+    "-mf16c",
-+    "-mfma"
-+  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vexp/gen/f32-vexp-fma3-rational-3-2-div.c",
+-      "src/src/f32-vlrelu/gen/f32-vlrelu-avx512f-u16.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x16-packw/gen/x16-packw-x16-gemm-goi-avx2-u16-prfm.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -35187,12 +38156,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-lut/gen/x8-lut-scalar-u4.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vlrelu_avx512f_standalone") {
+-    cflags = [ "-mavx512f" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -35200,27 +38177,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("f32-vexp_f16c-fma-no-avx2_standalone") {
--    cflags = [
--      "-mf16c",
--      "-mfma",
--      "-mno-avx2",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vexp/gen/f32-vexp-fma3-rational-3-2-div.c",
+-      "src/src/f32-vlrelu/gen/f32-vlrelu-avx512f-u16.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("x16-packw_f16c-fma-avx2_standalone") {
-+  cflags = [
-+    "-mavx2",
-+    "-mf16c",
-+    "-mfma"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -35232,42 +38194,41 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x16-packw/gen/x16-packw-x16-gemm-goi-avx2-u16-prfm.c"
-+  ]
++# This is a target that cannot depend on //base.
++source_set("x8-lut_x64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-lut/gen/x8-lut-scalar-u4.c"
 +  ]
  
--  source_set("f32-vexp_sse2-no-sse3") {
+-  source_set("f32-vlrelu_sse-no-sse2") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mno-sse2",
+-      "-msse",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vexp/gen/f32-vexp-sse2-rational-3-2-div.c",
+-      "src/src/f32-vlrelu/gen/f32-vlrelu-sse-u8.c",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -35279,40 +38240,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("x16-transposec_f16c-fma-avx2") {
-+  cflags = [
-+    "-mavx2",
-+    "-mf16c",
-+    "-mfma"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x16-transposec/gen/x16-transposec-16x16-reuse-switch-avx2.c"
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vexp_sse2-no-sse3_standalone") {
+-  source_set("f32-vlrelu_sse-no-sse2_standalone") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mno-sse2",
+-      "-msse",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vexp/gen/f32-vexp-sse2-rational-3-2-div.c",
+-      "src/src/f32-vlrelu/gen/f32-vlrelu-sse-u8.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++source_set("x8-packq_x64") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -35324,44 +38272,41 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+# This is a target that cannot depend on //base.
-+source_set("x16-transposec_f16c-fma-avx2_standalone") {
-+  cflags = [
-+    "-mavx2",
-+    "-mf16c",
-+    "-mfma"
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-packq/x8-packq-scalar-f32qp8-u1.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x16-transposec/gen/x16-transposec-16x16-reuse-switch-avx2.c"
-+  ]
- 
--  source_set("f32-vexp_x64") {
--    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vexp/gen/f32-vexp-scalar-rational-3-2-div.c",
+-  source_set("f32-vlrelu_sse4.1-no-sse4.2") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vlrelu/gen/f32-vlrelu-sse41-u8.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -35372,31 +38317,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++# This is a target that cannot depend on //base.
++source_set("x8-packq_x64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vexp_x64_standalone") {
--    cflags = []
-+source_set("x16-transposec_sse2-no-sse3") {
-+  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
+-  source_set("f32-vlrelu_sse4.1-no-sse4.2_standalone") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-packq/x8-packq-scalar-f32qp8-u1.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vexp/gen/f32-vexp-scalar-rational-3-2-div.c",
+-      "src/src/f32-vlrelu/gen/f32-vlrelu-sse41-u8.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x16-transposec/gen/x16-transposec-8x8-reuse-multi-sse2.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -35408,46 +38354,33 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
-+  public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
 +}
  
--  source_set("f32-vgelu_avx-no-avx2-no-f16c-no-fma") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
-+# This is a target that cannot depend on //base.
-+source_set("x16-transposec_sse2-no-sse3_standalone") {
+-  source_set("f32-vlrelu_x64") {
+-    cflags = []
++source_set("x8-packw_x64") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
-+  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vgelu/gen/f32-vgelu-avx-rational-12-10-div.c",
+-      "src/src/f32-vlrelu/gen/f32-vlrelu-scalar-u4.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x16-transposec/gen/x16-transposec-8x8-reuse-multi-sse2.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -35460,36 +38393,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-packw/gen/x8-packw-x16-gemm-goi-scalar-u2.c",
++    "src/src/x8-packw/gen/x8-packw-x32-gemm-goi-scalar-u2.c",
++    "src/src/x8-packw/gen/x8-packw-x4-gemm-goi-scalar-u2.c",
++    "src/src/x8-packw/gen/x8-packw-x8-gemm-goi-scalar-u2.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vlrelu_x64_standalone") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("f32-vgelu_avx-no-avx2-no-f16c-no-fma_standalone") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vgelu/gen/f32-vgelu-avx-rational-12-10-div.c",
+-      "src/src/f32-vlrelu/gen/f32-vlrelu-scalar-u4.c",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
++  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -35502,7 +38434,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("x16-transposec_x64") {
++# This is a target that cannot depend on //base.
++source_set("x8-packw_x64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -35514,24 +38447,30 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c"
++    "src/src/x8-packw/gen/x8-packw-x16-gemm-goi-scalar-u2.c",
++    "src/src/x8-packw/gen/x8-packw-x32-gemm-goi-scalar-u2.c",
++    "src/src/x8-packw/gen/x8-packw-x4-gemm-goi-scalar-u2.c",
++    "src/src/x8-packw/gen/x8-packw-x8-gemm-goi-scalar-u2.c"
 +  ]
  
--  source_set("f32-vgelu_avx512f") {
--    cflags = [ "-mavx512f" ]
+-  source_set("f32-vmulcaddc_sse-no-sse2") {
+-    cflags = [
+-      "-mno-sse2",
+-      "-msse",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vgelu/gen/f32-vgelu-avx512f-rational-12-10-nr.c",
+-      "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c4-minmax-sse-2x.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -35545,26 +38484,29 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
-+}
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+# This is a target that cannot depend on //base.
-+source_set("x16-transposec_x64_standalone") {
-+  cflags = [
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vgelu_avx512f_standalone") {
--    cflags = [ "-mavx512f" ]
-+  ]
- 
+-  source_set("f32-vmulcaddc_sse-no-sse2_standalone") {
+-    cflags = [
+-      "-mno-sse2",
+-      "-msse",
+-    ]
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vgelu/gen/f32-vgelu-avx512f-rational-12-10-nr.c",
+-      "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c4-minmax-sse-2x.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c"
++source_set("x8-transposec_f16c-fma-avx2") {
++  cflags = [
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -35577,41 +38519,42 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-transposec/gen/x8-transposec-32x32-reuse-switch-avx2.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
+-  source_set("f32-vmulcaddc_x64") {
+-    cflags = []
 +  public_configs = [ ":xnnpack_config" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
 +}
  
--  source_set("f32-vgelu_f16c-fma-no-avx2") {
--    cflags = [
--      "-mf16c",
--      "-mfma",
--      "-mno-avx2",
--    ]
-+source_set("x16-x32-packw_x64") {
-+  cflags = [
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vgelu/gen/f32-vgelu-fma3-rational-12-10-div.c",
+-      "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c1-minmax-scalar-2x.c",
 -    ]
++# This is a target that cannot depend on //base.
++source_set("x8-transposec_f16c-fma-avx2_standalone") {
++  cflags = [
++    "-mavx2",
++    "-mf16c",
++    "-mfma"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -35626,8 +38569,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-gio-scalar.c",
-+    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-goi-scalar.c"
++    "src/src/x8-transposec/gen/x8-transposec-32x32-reuse-switch-avx2.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -35637,25 +38579,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vgelu_f16c-fma-no-avx2_standalone") {
--    cflags = [
--      "-mf16c",
--      "-mfma",
--      "-mno-avx2",
--    ]
+-  source_set("f32-vmulcaddc_x64_standalone") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vgelu/gen/f32-vgelu-fma3-rational-12-10-div.c",
+-      "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c1-minmax-scalar-2x.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
-+}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -35667,11 +38604,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("x16-x32-packw_x64_standalone") {
-+  cflags = [
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    public_configs = [ ":xnnpack_config" ]
++source_set("x8-transposec_sse2-no-sse3") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -35680,14 +38622,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-gio-scalar.c",
-+    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-goi-scalar.c"
++    "src/src/x8-transposec/gen/x8-transposec-16x16-reuse-mov-sse2.c"
 +  ]
  
--  source_set("f32-vgelu_sse2-no-sse3") {
+-  source_set("f32-vrelu_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -35695,13 +38638,13 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vgelu/gen/f32-vgelu-sse2-rational-12-10-div.c",
+-      "src/src/f32-vrelu/gen/f32-vrelu-avx.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -35715,33 +38658,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++# This is a target that cannot depend on //base.
++source_set("x8-transposec_sse2-no-sse3_standalone") {
++  cflags = [
++    "-mno-sse3",
++    "-msse2"
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vgelu_sse2-no-sse3_standalone") {
+-  source_set("f32-vrelu_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
-+source_set("x24-transposec_ssse3-no-sse4.1") {
-+  cflags = [
-+    "-mno-sse4.1",
-+    "-mssse3"
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-transposec/gen/x8-transposec-16x16-reuse-mov-sse2.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vgelu/gen/f32-vgelu-sse2-rational-12-10-div.c",
+-      "src/src/f32-vrelu/gen/f32-vrelu-avx.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x24-transposec/x24-transposec-4x4-ssse3.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -35753,41 +38700,33 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
-+  public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
 +}
  
--  source_set("f32-vgelu_x64") {
--    cflags = []
-+# This is a target that cannot depend on //base.
-+source_set("x24-transposec_ssse3-no-sse4.1_standalone") {
+-  source_set("f32-vrelu_avx512f") {
+-    cflags = [ "-mavx512f" ]
++source_set("x8-transposec_x64") {
 +  cflags = [
-+    "-mno-sse4.1",
-+    "-mssse3"
-+  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vgelu/gen/f32-vgelu-scalar-rational-12-10-div.c",
+-      "src/src/f32-vrelu/gen/f32-vrelu-avx512f.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x24-transposec/x24-transposec-4x4-ssse3.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -35800,31 +38739,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-transposec/gen/x8-transposec-2x4-scalar-int.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vrelu_avx512f_standalone") {
+-    cflags = [ "-mavx512f" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("f32-vgelu_x64_standalone") {
--    cflags = []
-+  public_configs = [ ":xnnpack_config" ]
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vgelu/gen/f32-vgelu-scalar-rational-12-10-div.c",
+-      "src/src/f32-vrelu/gen/f32-vrelu-avx512f.c",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
++  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -35837,7 +38777,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("x24-transposec_x64") {
++# This is a target that cannot depend on //base.
++source_set("x8-transposec_x64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -35849,15 +38790,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x24-transposec/gen/x24-transposec-1x2-scalar.c"
++    "src/src/x8-transposec/gen/x8-transposec-2x4-scalar-int.c"
 +  ]
  
--  source_set("f32-vhswish_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("f32-vrelu_sse2-no-sse3") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -35865,13 +38804,13 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vhswish/gen/f32-vhswish-avx-u16.c",
+-      "src/src/f32-vrelu/gen/f32-vrelu-sse2.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -35885,32 +38824,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
-+}
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+# This is a target that cannot depend on //base.
-+source_set("x24-transposec_x64_standalone") {
-+  cflags = [
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vhswish_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("f32-vrelu_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+  ]
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vhswish/gen/f32-vhswish-avx-u16.c",
+-      "src/src/f32-vrelu/gen/f32-vrelu-sse2.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x24-transposec/gen/x24-transposec-1x2-scalar.c"
-+  ]
++source_set("xx-copy_x64") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -35922,50 +38855,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/xx-copy/xx-copy-scalar-memcpy.c"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("f32-vhswish_avx512f") {
--    cflags = [ "-mavx512f" ]
-+source_set("x32-packw_avx-no-avx2-no-f16c-no-fma") {
-+  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
+-  source_set("f32-vrelu_x64") {
+-    cflags = []
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vhswish/gen/f32-vhswish-avx512f-u16.c",
+-      "src/src/f32-vrelu/gen/f32-vrelu-scalar.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x32-packw/gen/x32-packw-x16-gemm-gio-avx-u8.c",
-+    "src/src/x32-packw/gen/x32-packw-x16-gemm-goi-avx-u4.c",
-+    "src/src/x32-packw/gen/x32-packw-x16s4-gemm-goi-avx-u4.c",
-+    "src/src/x32-packw/gen/x32-packw-x8-gemm-gio-avx-u8.c",
-+    "src/src/x32-packw/gen/x32-packw-x8-gemm-goi-avx-u4.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -35977,37 +38897,29 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++# This is a target that cannot depend on //base.
++source_set("xx-copy_x64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vhswish_avx512f_standalone") {
--    cflags = [ "-mavx512f" ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
+-  source_set("f32-vrelu_x64_standalone") {
+-    cflags = []
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/xx-copy/xx-copy-scalar-memcpy.c"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vhswish/gen/f32-vhswish-avx512f-u16.c",
+-      "src/src/f32-vrelu/gen/f32-vrelu-scalar.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("x32-packw_avx-no-avx2-no-f16c-no-fma_standalone") {
-+  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -36019,19 +38931,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x32-packw/gen/x32-packw-x16-gemm-gio-avx-u8.c",
-+    "src/src/x32-packw/gen/x32-packw-x16-gemm-goi-avx-u4.c",
-+    "src/src/x32-packw/gen/x32-packw-x16s4-gemm-goi-avx-u4.c",
-+    "src/src/x32-packw/gen/x32-packw-x8-gemm-gio-avx-u8.c",
-+    "src/src/x32-packw/gen/x32-packw-x8-gemm-goi-avx-u4.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -36050,27 +38949,29 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("f32-vhswish_f16c-fma-no-avx2") {
+-  source_set("f32-vrnd_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
--      "-mf16c",
--      "-mfma",
+-      "-mavx",
 -      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
-+source_set("x32-packw_avx512f") {
++source_set("xx-fill_sse2-no-sse3") {
 +  cflags = [
-+    "-mavx512f"
++    "-mno-sse3",
++    "-msse2"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vhswish/gen/f32-vhswish-fma3-u16.c",
+-      "src/src/f32-vrnd/gen/f32-vrndd-avx-u16.c",
+-      "src/src/f32-vrnd/gen/f32-vrndne-avx-u16.c",
+-      "src/src/f32-vrnd/gen/f32-vrndu-avx-u16.c",
+-      "src/src/f32-vrnd/gen/f32-vrndz-avx-u16.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x32-packw/gen/x32-packw-x16-gemm-gio-avx512f-u8.c",
-+    "src/src/x32-packw/gen/x32-packw-x16-gemm-goi-avx512f-u4-prfm.c",
-+    "src/src/x32-packw/gen/x32-packw-x32-gemm-gio-avx512f-u8.c",
-+    "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-avx512f-u4-prfm.c"
++    "src/src/xx-fill/xx-fill-sse2-u64.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -36097,23 +38998,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vhswish_f16c-fma-no-avx2_standalone") {
+-  source_set("f32-vrnd_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
--      "-mf16c",
--      "-mfma",
+-      "-mavx",
 -      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vhswish/gen/f32-vhswish-fma3-u16.c",
+-      "src/src/f32-vrnd/gen/f32-vrndd-avx-u16.c",
+-      "src/src/f32-vrnd/gen/f32-vrndne-avx-u16.c",
+-      "src/src/f32-vrnd/gen/f32-vrndu-avx-u16.c",
+-      "src/src/f32-vrnd/gen/f32-vrndz-avx-u16.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x32-packw_avx512f_standalone") {
++source_set("xx-fill_sse2-no-sse3_standalone") {
 +  cflags = [
-+    "-mavx512f"
++    "-mno-sse3",
++    "-msse2"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -36128,10 +39034,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x32-packw/gen/x32-packw-x16-gemm-gio-avx512f-u8.c",
-+    "src/src/x32-packw/gen/x32-packw-x16-gemm-goi-avx512f-u4-prfm.c",
-+    "src/src/x32-packw/gen/x32-packw-x32-gemm-gio-avx512f-u8.c",
-+    "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-avx512f-u4-prfm.c"
++    "src/src/xx-fill/xx-fill-sse2-u64.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -36143,114 +39046,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--  source_set("f32-vhswish_sse-no-sse2") {
--    cflags = [
--      "-mno-sse2",
--      "-msse",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vhswish/gen/f32-vhswish-sse-u8.c",
--    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
 -
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+source_set("x32-packw_sse2-no-sse3") {
-+  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x32-packw/gen/x32-packw-x2c4-gemm-goi-sse2-u4.c",
-+    "src/src/x32-packw/gen/x32-packw-x8-gemm-goi-sse2-u4.c"
-+  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("f32-vhswish_sse-no-sse2_standalone") {
--    cflags = [
--      "-mno-sse2",
--      "-msse",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vhswish/gen/f32-vhswish-sse-u8.c",
--    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-  source_set("f32-vrnd_avx512f") {
+-    cflags = [ "-mavx512f" ]
 -
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
-+# This is a target that cannot depend on //base.
-+source_set("x32-packw_sse2-no-sse3_standalone") {
-+  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
-+  ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x32-packw/gen/x32-packw-x2c4-gemm-goi-sse2-u4.c",
-+    "src/src/x32-packw/gen/x32-packw-x8-gemm-goi-sse2-u4.c"
-+  ]
- 
--  source_set("f32-vhswish_x64") {
--    cflags = []
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vhswish/gen/f32-vhswish-scalar-u4.c",
+-      "src/src/f32-vrnd/gen/f32-vrndd-avx512f-u16.c",
+-      "src/src/f32-vrnd/gen/f32-vrndne-avx512f-u16.c",
+-      "src/src/f32-vrnd/gen/f32-vrndu-avx512f-u16.c",
+-      "src/src/f32-vrnd/gen/f32-vrndz-avx512f-u16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -36278,14 +39083,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vhswish_x64_standalone") {
--    cflags = []
+-  source_set("f32-vrnd_avx512f_standalone") {
+-    cflags = [ "-mavx512f" ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vhswish/gen/f32-vhswish-scalar-u4.c",
+-      "src/src/f32-vrnd/gen/f32-vrndd-avx512f-u16.c",
+-      "src/src/f32-vrnd/gen/f32-vrndne-avx512f-u16.c",
+-      "src/src/f32-vrnd/gen/f32-vrndu-avx512f-u16.c",
+-      "src/src/f32-vrnd/gen/f32-vrndz-avx512f-u16.c",
 -    ]
-+source_set("x32-packw_x64") {
++source_set("xx-fill_x64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -36303,12 +39111,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x32-packw/gen/x32-packw-x2-gemm-gio-scalar.c",
-+    "src/src/x32-packw/gen/x32-packw-x2-gemm-goi-scalar-float-u4.c",
-+    "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-scalar-int-u2.c",
-+    "src/src/x32-packw/gen/x32-packw-x4-gemm-gio-scalar.c",
-+    "src/src/x32-packw/gen/x32-packw-x4-gemm-goi-scalar-float-u4.c",
-+    "src/src/x32-packw/gen/x32-packw-x64-gemm-goi-scalar-int-u2.c"
++    "src/src/xx-fill/xx-fill-scalar-u16.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -36319,8 +39122,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("f32-vlog_avx512f") {
--    cflags = [ "-mavx512f" ]
+-  source_set("f32-vrnd_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -36330,7 +39136,10 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vlog/gen/f32-vlog-avx512f-rational-3-3-div.c",
+-      "src/src/f32-vrnd/gen/f32-vrndd-sse2-u8.c",
+-      "src/src/f32-vrnd/gen/f32-vrndne-sse2-u8.c",
+-      "src/src/f32-vrnd/gen/f32-vrndu-sse2-u8.c",
+-      "src/src/f32-vrnd/gen/f32-vrndz-sse2-u8.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -36346,7 +39155,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x32-packw_x64_standalone") {
++source_set("xx-fill_x64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -36354,21 +39163,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vlog_avx512f_standalone") {
--    cflags = [ "-mavx512f" ]
+-  source_set("f32-vrnd_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x32-packw/gen/x32-packw-x2-gemm-gio-scalar.c",
-+    "src/src/x32-packw/gen/x32-packw-x2-gemm-goi-scalar-float-u4.c",
-+    "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-scalar-int-u2.c",
-+    "src/src/x32-packw/gen/x32-packw-x4-gemm-gio-scalar.c",
-+    "src/src/x32-packw/gen/x32-packw-x4-gemm-goi-scalar-float-u4.c",
-+    "src/src/x32-packw/gen/x32-packw-x64-gemm-goi-scalar-int-u2.c"
++    "src/src/xx-fill/xx-fill-scalar-u16.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vlog/gen/f32-vlog-avx512f-rational-3-3-div.c",
+-      "src/src/f32-vrnd/gen/f32-vrndd-sse2-u8.c",
+-      "src/src/f32-vrnd/gen/f32-vrndne-sse2-u8.c",
+-      "src/src/f32-vrnd/gen/f32-vrndu-sse2-u8.c",
+-      "src/src/f32-vrnd/gen/f32-vrndz-sse2-u8.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -36402,27 +39212,27 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("f32-vlog_f16c-fma-avx2") {
+-  source_set("f32-vrnd_sse4.1-no-sse4.2") {
 -    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
-+source_set("x32-transposec_avx-no-avx2-no-f16c-no-fma") {
++source_set("xx-pad_sse2-no-sse3") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mno-sse3",
++    "-msse2"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vlog/gen/f32-vlog-avx2-rational-3-3-div.c",
+-      "src/src/f32-vrnd/gen/f32-vrndd-sse41-u8.c",
+-      "src/src/f32-vrnd/gen/f32-vrndne-sse41-u8.c",
+-      "src/src/f32-vrnd/gen/f32-vrndu-sse41-u8.c",
+-      "src/src/f32-vrnd/gen/f32-vrndz-sse41-u8.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x32-transposec/gen/x32-transposec-8x8-reuse-multi-avx.c"
++    "src/src/xx-pad/xx-pad-p16-sse2-u16.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -36449,26 +39259,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vlog_f16c-fma-avx2_standalone") {
+-  source_set("f32-vrnd_sse4.1-no-sse4.2_standalone") {
 -    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vlog/gen/f32-vlog-avx2-rational-3-3-div.c",
+-      "src/src/f32-vrnd/gen/f32-vrndd-sse41-u8.c",
+-      "src/src/f32-vrnd/gen/f32-vrndne-sse41-u8.c",
+-      "src/src/f32-vrnd/gen/f32-vrndu-sse41-u8.c",
+-      "src/src/f32-vrnd/gen/f32-vrndz-sse41-u8.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x32-transposec_avx-no-avx2-no-f16c-no-fma_standalone") {
++source_set("xx-pad_sse2-no-sse3_standalone") {
 +  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++    "-mno-sse3",
++    "-msse2"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -36483,7 +39293,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x32-transposec/gen/x32-transposec-8x8-reuse-multi-avx.c"
++    "src/src/xx-pad/xx-pad-p16-sse2-u16.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -36495,6 +39305,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
+-
+-  source_set("f32-vrnd_x64") {
+-    cflags = []
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
+-      "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
+-      "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
+-      "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -36502,21 +39323,84 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--  source_set("f32-vlog_f16c-fma-no-avx2") {
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vrnd_x64_standalone") {
+-    cflags = []
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
+-      "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
+-      "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
+-      "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c",
+-    ]
++source_set("xx-pad_x64") {
++  cflags = [
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/xx-pad/xx-pad-p4-scalar-u16.c"
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  source_set("f32-vrsqrt_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
--      "-mf16c",
--      "-mfma",
+-      "-mavx",
 -      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vlog/gen/f32-vlog-fma3-rational-3-3-div.c",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx-rsqrt.c",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx-sqrt.c",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
++  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -36529,34 +39413,100 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("x32-transposec_sse-no-sse2") {
++# This is a target that cannot depend on //base.
++source_set("xx-pad_x64_standalone") {
 +  cflags = [
-+    "-mno-sse2",
-+    "-msse"
-+  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x32-transposec/x32-transposec-4x4-sse.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vlog_f16c-fma-no-avx2_standalone") {
+-  source_set("f32-vrsqrt_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
--      "-mf16c",
--      "-mfma",
+-      "-mavx",
 -      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/xx-pad/xx-pad-p4-scalar-u16.c"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx-rsqrt.c",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx-sqrt.c",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  source_set("f32-vrsqrt_avx512f") {
+-    cflags = [ "-mavx512f" ]
++source_set("xx-transposev_x64") {
++  cflags = [
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx512f-rsqrt.c",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx512f-sqrt.c",
+-    ]
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/xx-transposev/xx-transposev-1x1-scalar-memcpy.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vlog/gen/f32-vlog-fma3-rational-3-3-div.c",
--    ]
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vrsqrt_avx512f_standalone") {
+-    cflags = [ "-mavx512f" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -36564,6 +39514,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx512f-rsqrt.c",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx512f-sqrt.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -36574,15 +39532,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("x32-transposec_sse-no-sse2_standalone") {
++source_set("xx-transposev_x64_standalone") {
 +  cflags = [
-+    "-mno-sse2",
-+    "-msse"
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -36591,10 +39545,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x32-transposec/x32-transposec-4x4-sse.c"
++    "src/src/xx-transposev/xx-transposev-1x1-scalar-memcpy.c"
 +  ]
  
--  source_set("f32-vlog_sse2-no-sse3") {
+-  source_set("f32-vrsqrt_sse2-no-sse3") {
 -    cflags = [
 -      "-mno-sse3",
 -      "-msse2",
@@ -36605,7 +39559,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vlog/gen/f32-vlog-sse2-rational-3-3-div.c",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-sse2-sqrt.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -36633,17 +39587,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vlog_sse2-no-sse3_standalone") {
+-  source_set("f32-vrsqrt_sse2-no-sse3_standalone") {
 -    cflags = [
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
--
++}
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vlog/gen/f32-vlog-sse2-rational-3-3-div.c",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-sse2-sqrt.c",
 -    ]
-+source_set("x32-transposec_x64") {
++if (current_cpu == "arm64") {
++source_set("configs_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -36661,7 +39617,30 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x32-transposec/gen/x32-transposec-2x4-scalar-int.c"
++    "src/src/configs/argmaxpool-config.c",
++    "src/src/configs/avgpool-config.c",
++    "src/src/configs/binary-elementwise-config.c",
++    "src/src/configs/cmul-config.c",
++    "src/src/configs/conv-hwc2chw-config.c",
++    "src/src/configs/dwconv-config.c",
++    "src/src/configs/dwconv2d-chw-config.c",
++    "src/src/configs/gemm-config.c",
++    "src/src/configs/hardware-config.c",
++    "src/src/configs/ibilinear-chw-config.c",
++    "src/src/configs/ibilinear-config.c",
++    "src/src/configs/lut32norm-config.c",
++    "src/src/configs/maxpool-config.c",
++    "src/src/configs/pack-lh-config.c",
++    "src/src/configs/raddstoreexpminusmax-config.c",
++    "src/src/configs/reduce-config.c",
++    "src/src/configs/spmm-config.c",
++    "src/src/configs/transpose-config.c",
++    "src/src/configs/unary-elementwise-config.c",
++    "src/src/configs/unpool-config.c",
++    "src/src/configs/vmulcaddc-config.c",
++    "src/src/configs/x8-lut-config.c",
++    "src/src/configs/xx-fill-config.c",
++    "src/src/configs/xx-pad-config.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -36672,7 +39651,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("f32-vlog_x64") {
+-  source_set("f32-vrsqrt_x64") {
 -    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -36683,7 +39662,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vlog/gen/f32-vlog-scalar-rational-3-3-div.c",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-sqrt.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -36699,7 +39679,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x32-transposec_x64_standalone") {
++source_set("configs_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -36707,16 +39687,40 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vlog_x64_standalone") {
+-  source_set("f32-vrsqrt_x64_standalone") {
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x32-transposec/gen/x32-transposec-2x4-scalar-int.c"
++    "src/src/configs/argmaxpool-config.c",
++    "src/src/configs/avgpool-config.c",
++    "src/src/configs/binary-elementwise-config.c",
++    "src/src/configs/cmul-config.c",
++    "src/src/configs/conv-hwc2chw-config.c",
++    "src/src/configs/dwconv-config.c",
++    "src/src/configs/dwconv2d-chw-config.c",
++    "src/src/configs/gemm-config.c",
++    "src/src/configs/hardware-config.c",
++    "src/src/configs/ibilinear-chw-config.c",
++    "src/src/configs/ibilinear-config.c",
++    "src/src/configs/lut32norm-config.c",
++    "src/src/configs/maxpool-config.c",
++    "src/src/configs/pack-lh-config.c",
++    "src/src/configs/raddstoreexpminusmax-config.c",
++    "src/src/configs/reduce-config.c",
++    "src/src/configs/spmm-config.c",
++    "src/src/configs/transpose-config.c",
++    "src/src/configs/unary-elementwise-config.c",
++    "src/src/configs/unpool-config.c",
++    "src/src/configs/vmulcaddc-config.c",
++    "src/src/configs/x8-lut-config.c",
++    "src/src/configs/xx-fill-config.c",
++    "src/src/configs/xx-pad-config.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vlog/gen/f32-vlog-scalar-rational-3-3-div.c",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-sqrt.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -36750,118 +39754,19 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("f32-vlrelu_avx-no-avx2-no-f16c-no-fma") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
-+source_set("x32-unpool_sse2-no-sse3") {
-+  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vlrelu/gen/f32-vlrelu-avx-u16.c",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x32-unpool/x32-unpool-sse2.c"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("f32-vlrelu_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("f32-vsigmoid_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vlrelu/gen/f32-vlrelu-avx-u16.c",
--    ]
-+# This is a target that cannot depend on //base.
-+source_set("x32-unpool_sse2-no-sse3_standalone") {
-+  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x32-unpool/x32-unpool-sse2.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
- 
--  source_set("f32-vlrelu_avx512f") {
--    cflags = [ "-mavx512f" ]
-+source_set("x32-unpool_x64") {
++source_set("enums_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vlrelu/gen/f32-vlrelu-avx512f-u16.c",
+-      "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx-rr2-p5-nr2-u16.c",
 -    ]
 +  ]
  
@@ -36877,7 +39782,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x32-unpool/x32-unpool-scalar.c"
++    "src/src/enums/allocation-type.c",
++    "src/src/enums/datatype-strings.c",
++    "src/src/enums/microkernel-type.c",
++    "src/src/enums/node-type.c",
++    "src/src/enums/operator-type.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -36887,8 +39796,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vlrelu_avx512f_standalone") {
--    cflags = [ "-mavx512f" ]
+-  source_set("f32-vsigmoid_avx-no-avx2-no-f16c-no-fma_standalone") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -36898,7 +39812,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vlrelu/gen/f32-vlrelu-avx512f-u16.c",
+-      "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx-rr2-p5-nr2-u16.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -36914,7 +39828,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x32-unpool_x64_standalone") {
++source_set("enums_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -36926,21 +39840,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x32-unpool/x32-unpool-scalar.c"
++    "src/src/enums/allocation-type.c",
++    "src/src/enums/datatype-strings.c",
++    "src/src/enums/microkernel-type.c",
++    "src/src/enums/node-type.c",
++    "src/src/enums/operator-type.c"
 +  ]
  
--  source_set("f32-vlrelu_sse-no-sse2") {
--    cflags = [
--      "-mno-sse2",
--      "-msse",
--    ]
+-  source_set("f32-vsigmoid_avx512f") {
+-    cflags = [ "-mavx512f" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vlrelu/gen/f32-vlrelu-sse-u8.c",
+-      "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx512f-rr2-lut32-p2-perm2-scalef-div-u64.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -36968,116 +39883,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vlrelu_sse-no-sse2_standalone") {
--    cflags = [
--      "-mno-sse2",
--      "-msse",
--    ]
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vlrelu/gen/f32-vlrelu-sse-u8.c",
--    ]
-+source_set("x64-transposec_avx-no-avx2-no-f16c-no-fma") {
-+  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-  source_set("f32-vsigmoid_avx512f_standalone") {
+-    cflags = [ "-mavx512f" ]
 -
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x64-transposec/gen/x64-transposec-4x4-reuse-multi-avx.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--  source_set("f32-vlrelu_sse4.1-no-sse4.2") {
--    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vlrelu/gen/f32-vlrelu-sse41-u8.c",
+-      "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx512f-rr2-lut32-p2-perm2-scalef-div-u64.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("x64-transposec_avx-no-avx2-no-f16c-no-fma_standalone") {
-+  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
 -
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x64-transposec/gen/x64-transposec-4x4-reuse-multi-avx.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("f32-vlrelu_sse4.1-no-sse4.2_standalone") {
--    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
--    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vlrelu/gen/f32-vlrelu-sse41-u8.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -37088,16 +39901,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
- 
+-
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("x64-transposec_sse2-no-sse3") {
++source_set("f16-avgpool_arch=armv8.2-a+fp16") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -37106,18 +39914,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x64-transposec/gen/x64-transposec-2x2-multi-mov-sse2.c"
++    "src/src/f16-avgpool/gen/f16-avgpool-9p-minmax-neonfp16arith.c"
 +  ]
  
--  source_set("f32-vlrelu_x64") {
--    cflags = []
+-  source_set("f32-vsigmoid_f16c-fma-avx2") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vlrelu/gen/f32-vlrelu-scalar-u4.c",
+-      "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx2-rr1-p5-div-u16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -37142,23 +39954,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("x64-transposec_sse2-no-sse3_standalone") {
++source_set("f16-avgpool_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vlrelu_x64_standalone") {
--    cflags = []
+-  source_set("f32-vsigmoid_f16c-fma-avx2_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x64-transposec/gen/x64-transposec-2x2-multi-mov-sse2.c"
++    "src/src/f16-avgpool/gen/f16-avgpool-9p-minmax-neonfp16arith.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vlrelu/gen/f32-vlrelu-scalar-u4.c",
+-      "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx2-rr1-p5-div-u16.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -37192,20 +40007,17 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("f32-vmulcaddc_sse-no-sse2") {
+-  source_set("f32-vsigmoid_sse2-no-sse3") {
 -    cflags = [
--      "-mno-sse2",
--      "-msse",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+source_set("x64-transposec_x64") {
-+  cflags = [
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c4-minmax-sse-2x.c",
+-      "src/src/f32-vsigmoid/gen/f32-vsigmoid-sse2-rr2-lut64-p2-div-u8.c",
 -    ]
-+  ]
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -37216,22 +40028,31 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c"
++source_set("f16-conv-hwc2chw_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-conv-hwc2chw/f16-conv-hwc2chw-3x3s2p1c3x4-neonfp16arith-2x2.c"
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vsigmoid_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("f32-vmulcaddc_sse-no-sse2_standalone") {
--    cflags = [
--      "-mno-sse2",
--      "-msse",
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vsigmoid/gen/f32-vsigmoid-sse2-rr2-lut64-p2-div-u8.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -37240,13 +40061,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c4-minmax-sse-2x.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -37257,11 +40071,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("x64-transposec_x64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
++# This is a target that cannot depend on //base.
++source_set("f16-conv-hwc2chw_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -37270,18 +40087,21 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c"
++    "src/src/f16-conv-hwc2chw/f16-conv-hwc2chw-3x3s2p1c3x4-neonfp16arith-2x2.c"
 +  ]
  
--  source_set("f32-vmulcaddc_x64") {
--    cflags = []
+-  source_set("f32-vsigmoid_sse4.1-no-sse4.2") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c1-minmax-scalar-2x.c",
+-      "src/src/f32-vsigmoid/gen/f32-vsigmoid-sse41-rr2-lut64-p2-div-u8.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -37309,23 +40129,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vmulcaddc_x64_standalone") {
--    cflags = []
-+source_set("x8-lut_avx-no-avx2-no-f16c-no-fma") {
-+  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
-+  ]
- 
+-  source_set("f32-vsigmoid_sse4.1-no-sse4.2_standalone") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
+-    ]
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c1-minmax-scalar-2x.c",
+-      "src/src/f32-vsigmoid/gen/f32-vsigmoid-sse41-rr2-lut64-p2-div-u8.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x8-lut/gen/x8-lut-avx-u64.c"
++source_set("f16-dwconv2d-chw_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -37338,48 +40154,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-dwconv2d-chw/gen/f16-dwconv2d-chw-3x3p1-minmax-neonfp16arith-2x8.c",
++    "src/src/f16-dwconv2d-chw/gen/f16-dwconv2d-chw-3x3s2p1-minmax-neonfp16arith-1x8.c",
++    "src/src/f16-dwconv2d-chw/gen/f16-dwconv2d-chw-5x5p2-minmax-neonfp16arith-1x8.c",
++    "src/src/f16-dwconv2d-chw/gen/f16-dwconv2d-chw-5x5s2p2-minmax-neonfp16arith-1x8.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
-+# This is a target that cannot depend on //base.
-+source_set("x8-lut_avx-no-avx2-no-f16c-no-fma_standalone") {
-+  cflags = [
-+    "-mavx",
-+    "-mno-avx2",
-+    "-mno-f16c",
-+    "-mno-fma"
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
--  source_set("f32-vrelu_x64") {
+-  source_set("f32-vsigmoid_x64") {
 -    cflags = []
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x8-lut/gen/x8-lut-avx-u64.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrelu/gen/f32-vrelu-scalar-u8.c",
+-      "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++# This is a target that cannot depend on //base.
++source_set("f16-dwconv2d-chw_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -37391,39 +40203,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-dwconv2d-chw/gen/f16-dwconv2d-chw-3x3p1-minmax-neonfp16arith-2x8.c",
++    "src/src/f16-dwconv2d-chw/gen/f16-dwconv2d-chw-3x3s2p1-minmax-neonfp16arith-1x8.c",
++    "src/src/f16-dwconv2d-chw/gen/f16-dwconv2d-chw-5x5p2-minmax-neonfp16arith-1x8.c",
++    "src/src/f16-dwconv2d-chw/gen/f16-dwconv2d-chw-5x5s2p2-minmax-neonfp16arith-1x8.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
-+
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vrelu_x64_standalone") {
+-  source_set("f32-vsigmoid_x64_standalone") {
 -    cflags = []
-+source_set("x8-lut_f16c-fma-avx2") {
-+  cflags = [
-+    "-mavx2",
-+    "-mf16c",
-+    "-mfma"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrelu/gen/f32-vrelu-scalar-u8.c",
+-      "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x8-lut/gen/x8-lut-avx2-u128.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -37435,10 +40243,45 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
+ 
+-    public_configs = [ ":xnnpack_config" ]
++source_set("f16-dwconv_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-dwconv/gen/f16-dwconv-25p8c-minmax-neonfp16arith-acc2.c",
++    "src/src/f16-dwconv/gen/f16-dwconv-3p16c-minmax-neonfp16arith.c",
++    "src/src/f16-dwconv/gen/f16-dwconv-4p16c-minmax-neonfp16arith.c",
++    "src/src/f16-dwconv/gen/f16-dwconv-9p16c-minmax-neonfp16arith.c",
++    "src/src/f16-dwconv/gen/f16-dwconv-9p8c-minmax-neonfp16arith.c"
++  ]
+ 
+-  source_set("f32-vsin_avx-no-avx2-no-f16c-no-fma") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vsin/gen/f32-vsin-avx-rational-5-4-div.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -37446,23 +40289,29 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
+-    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("x8-lut_f16c-fma-avx2_standalone") {
++source_set("f16-dwconv_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
-+    "-mavx2",
-+    "-mf16c",
-+    "-mfma"
++    "-march=armv8.2-a+fp16"
 +  ]
  
--  source_set("f32-vrnd_avx-no-avx2-no-f16c-no-fma") {
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vsin_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
@@ -37471,15 +40320,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x8-lut/gen/x8-lut-avx2-u128.c"
++    "src/src/f16-dwconv/gen/f16-dwconv-25p8c-minmax-neonfp16arith-acc2.c",
++    "src/src/f16-dwconv/gen/f16-dwconv-3p16c-minmax-neonfp16arith.c",
++    "src/src/f16-dwconv/gen/f16-dwconv-4p16c-minmax-neonfp16arith.c",
++    "src/src/f16-dwconv/gen/f16-dwconv-9p16c-minmax-neonfp16arith.c",
++    "src/src/f16-dwconv/gen/f16-dwconv-9p8c-minmax-neonfp16arith.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrnd/gen/f32-vrndd-avx-u16.c",
--      "src/src/f32-vrnd/gen/f32-vrndne-avx-u16.c",
--      "src/src/f32-vrnd/gen/f32-vrndu-avx-u16.c",
--      "src/src/f32-vrnd/gen/f32-vrndz-avx-u16.c",
+-      "src/src/f32-vsin/gen/f32-vsin-avx-rational-5-4-div.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -37493,7 +40343,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -37501,44 +40351,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
    }
 +}
  
--  # This is a target that cannot depend on //base.
--  source_set("f32-vrnd_avx-no-avx2-no-f16c-no-fma_standalone") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
-+source_set("x8-lut_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-  source_set("f32-vsin_avx512f") {
+-    cflags = [ "-mavx512f" ]
++source_set("f16-f32-vcvt_arm64") {
 +  cflags = [
-+    "-mavx512bw",
-+    "-mavx512cd",
-+    "-mavx512dq",
-+    "-mavx512f",
-+    "-mavx512vl",
-+    "-mf16c",
-+    "-mfma"
-+  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrnd/gen/f32-vrndd-avx-u16.c",
--      "src/src/f32-vrnd/gen/f32-vrndne-avx-u16.c",
--      "src/src/f32-vrnd/gen/f32-vrndu-avx-u16.c",
--      "src/src/f32-vrnd/gen/f32-vrndz-avx-u16.c",
+-      "src/src/f32-vsin/gen/f32-vsin-avx512f-rational-5-4-div.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x8-lut/gen/x8-lut-avx512skx-vpshufb-u64.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -37549,13 +40382,24 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-f32-vcvt/gen/f16-f32-vcvt-neon-int16-u16.c",
++    "src/src/f16-f32-vcvt/gen/f16-f32-vcvt-neonfp16-u16.c",
++    "src/src/f16-f32-vcvt/gen/f16-f32-vcvt-scalar-u4.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vsin_avx512f_standalone") {
+-    cflags = [ "-mavx512f" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -37563,20 +40407,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
--
--  source_set("f32-vrnd_avx512f") {
--    cflags = [ "-mavx512f" ]
--
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrnd/gen/f32-vrndd-avx512f-u16.c",
--      "src/src/f32-vrnd/gen/f32-vrndne-avx512f-u16.c",
--      "src/src/f32-vrnd/gen/f32-vrndu-avx512f-u16.c",
--      "src/src/f32-vrnd/gen/f32-vrndz-avx512f-u16.c",
+-      "src/src/f32-vsin/gen/f32-vsin-avx512f-rational-5-4-div.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -37589,51 +40422,39 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x8-lut_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
++source_set("f16-f32-vcvt_arm64_standalone") {
 +  cflags = [
-+    "-mavx512bw",
-+    "-mavx512cd",
-+    "-mavx512dq",
-+    "-mavx512f",
-+    "-mavx512vl",
-+    "-mf16c",
-+    "-mfma"
-+  ]
  
 -    public_configs = [ ":xnnpack_config" ]
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x8-lut/gen/x8-lut-avx512skx-vpshufb-u64.c"
++    "src/src/f16-f32-vcvt/gen/f16-f32-vcvt-neon-int16-u16.c",
++    "src/src/f16-f32-vcvt/gen/f16-f32-vcvt-neonfp16-u16.c",
++    "src/src/f16-f32-vcvt/gen/f16-f32-vcvt-scalar-u4.c"
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("f32-vrnd_avx512f_standalone") {
--    cflags = [ "-mavx512f" ]
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vrnd/gen/f32-vrndd-avx512f-u16.c",
--      "src/src/f32-vrnd/gen/f32-vrndne-avx512f-u16.c",
--      "src/src/f32-vrnd/gen/f32-vrndu-avx512f-u16.c",
--      "src/src/f32-vrnd/gen/f32-vrndz-avx512f-u16.c",
+-  source_set("f32-vsin_f16c-fma-no-avx2") {
+-    cflags = [
+-      "-mf16c",
+-      "-mfma",
+-      "-mno-avx2",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vsin/gen/f32-vsin-fma3-rational-5-4-div.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -37642,46 +40463,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
- 
--  source_set("f32-vrnd_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
-+source_set("x8-lut_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vbmi") {
-+  cflags = [
-+    "-mavx512bw",
-+    "-mavx512cd",
-+    "-mavx512dq",
-+    "-mavx512f",
-+    "-mavx512vbmi",
-+    "-mavx512vl",
-+    "-mf16c",
-+    "-mfma"
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vrnd/gen/f32-vrndd-sse2-u8.c",
--      "src/src/f32-vrnd/gen/f32-vrndne-sse2-u8.c",
--      "src/src/f32-vrnd/gen/f32-vrndu-sse2-u8.c",
--      "src/src/f32-vrnd/gen/f32-vrndz-sse2-u8.c",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x8-lut/gen/x8-lut-avx512vbmi-vpermx2b-u128.c"
-+  ]
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -37692,48 +40473,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vrnd_sse2-no-sse3_standalone") {
+-  source_set("f32-vsin_f16c-fma-no-avx2_standalone") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mf16c",
+-      "-mfma",
+-      "-mno-avx2",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrnd/gen/f32-vrndd-sse2-u8.c",
--      "src/src/f32-vrnd/gen/f32-vrndne-sse2-u8.c",
--      "src/src/f32-vrnd/gen/f32-vrndu-sse2-u8.c",
--      "src/src/f32-vrnd/gen/f32-vrndz-sse2-u8.c",
+-      "src/src/f32-vsin/gen/f32-vsin-fma3-rational-5-4-div.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("x8-lut_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vbmi_standalone") {
-+  cflags = [
-+    "-mavx512bw",
-+    "-mavx512cd",
-+    "-mavx512dq",
-+    "-mavx512f",
-+    "-mavx512vbmi",
-+    "-mavx512vl",
-+    "-mf16c",
-+    "-mfma"
-+  ]
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -37744,45 +40504,41 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x8-lut/gen/x8-lut-avx512vbmi-vpermx2b-u128.c"
-+  ]
- 
+-
 -    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++source_set("f16-f32acc-rdsum_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-f32acc-rdsum/gen/f16-f32acc-rdsum-7p7x-minmax-neonfp16arith-c16.c"
 +  ]
  
--  source_set("f32-vrnd_sse4.1-no-sse4.2") {
+-  source_set("f32-vsin_sse2-no-sse3") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrnd/gen/f32-vrndd-sse41-u8.c",
--      "src/src/f32-vrnd/gen/f32-vrndne-sse41-u8.c",
--      "src/src/f32-vrnd/gen/f32-vrndu-sse41-u8.c",
--      "src/src/f32-vrnd/gen/f32-vrndz-sse41-u8.c",
+-      "src/src/f32-vsin/gen/f32-vsin-sse2-rational-5-4-div.c",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -37794,30 +40550,31 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("x8-lut_x64") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
++# This is a target that cannot depend on //base.
++source_set("f16-f32acc-rdsum_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vrnd_sse4.1-no-sse4.2_standalone") {
+-  source_set("f32-vsin_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x8-lut/gen/x8-lut-scalar-u4.c"
++    "src/src/f16-f32acc-rdsum/gen/f16-f32acc-rdsum-7p7x-minmax-neonfp16arith-c16.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrnd/gen/f32-vrndd-sse41-u8.c",
--      "src/src/f32-vrnd/gen/f32-vrndne-sse41-u8.c",
--      "src/src/f32-vrnd/gen/f32-vrndu-sse41-u8.c",
--      "src/src/f32-vrnd/gen/f32-vrndz-sse41-u8.c",
+-      "src/src/f32-vsin/gen/f32-vsin-sse2-rational-5-4-div.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -37837,35 +40594,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
-+}
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
--
--  source_set("f32-vrnd_x64") {
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  source_set("f32-vsin_x64") {
 -    cflags = []
--
++source_set("f16-f32acc-rsum_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
--      "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u4.c",
--      "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
--      "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u4.c",
--      "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
--      "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u4.c",
--      "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c",
--      "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u4.c",
+-      "src/src/f32-vsin/gen/f32-vsin-scalar-rational-5-4-div.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("x8-lut_x64_standalone") {
-+  cflags = [
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-f32acc-rsum/gen/f16-f32acc-rsum-neonfp16arith-u32-acc4.c"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -37877,33 +40634,34 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x8-lut/gen/x8-lut-scalar-u4.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vrnd_x64_standalone") {
+-  source_set("f32-vsin_x64_standalone") {
 -    cflags = []
--
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
--      "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u4.c",
--      "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
--      "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u4.c",
--      "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
--      "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u4.c",
--      "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c",
--      "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u4.c",
+-      "src/src/f32-vsin/gen/f32-vsin-scalar-rational-5-4-div.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++# This is a target that cannot depend on //base.
++source_set("f16-f32acc-rsum_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -37915,39 +40673,45 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-f32acc-rsum/gen/f16-f32acc-rsum-neonfp16arith-u32-acc4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
--  source_set("f32-vrsqrt_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("f32-vsqrt_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
-+source_set("x8-packq_x64") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx-rsqrt-u16.c",
+-      "src/src/f32-vsqrt/gen/f32-vsqrt-avx-rsqrt.c",
+-      "src/src/f32-vsqrt/gen/f32-vsqrt-avx-sqrt.c",
 -    ]
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -37959,25 +40723,38 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x8-packq/x8-packq-scalar-f32qp8-u1.c"
++source_set("f16-gemm_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-gemm/gen/f16-gemm-1x16-minmax-neonfp16arith-ld64.c",
++    "src/src/f16-gemm/gen/f16-gemm-1x8-minmax-neonfp16arith-ld64.c",
++    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-neonfp16arith-ld64.c",
++    "src/src/f16-gemm/gen/f16-gemm-6x8-minmax-neonfp16arith-ld64.c"
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vrsqrt_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("f32-vsqrt_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vsqrt/gen/f32-vsqrt-avx-rsqrt.c",
+-      "src/src/f32-vsqrt/gen/f32-vsqrt-avx-sqrt.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -37985,13 +40762,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx-rsqrt-u16.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -38002,11 +40772,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("x8-packq_x64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
++# This is a target that cannot depend on //base.
++source_set("f16-gemm_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -38015,10 +40788,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x8-packq/x8-packq-scalar-f32qp8-u1.c"
++    "src/src/f16-gemm/gen/f16-gemm-1x16-minmax-neonfp16arith-ld64.c",
++    "src/src/f16-gemm/gen/f16-gemm-1x8-minmax-neonfp16arith-ld64.c",
++    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-neonfp16arith-ld64.c",
++    "src/src/f16-gemm/gen/f16-gemm-6x8-minmax-neonfp16arith-ld64.c"
 +  ]
  
--  source_set("f32-vrsqrt_avx512f") {
+-  source_set("f32-vsqrt_avx512f") {
 -    cflags = [ "-mavx512f" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -38026,7 +40802,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx512f-rsqrt-u32.c",
+-      "src/src/f32-vsqrt/gen/f32-vsqrt-avx512f-rsqrt.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -38054,15 +40830,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vrsqrt_avx512f_standalone") {
+-  source_set("f32-vsqrt_avx512f_standalone") {
 -    cflags = [ "-mavx512f" ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrsqrt/gen/f32-vrsqrt-avx512f-rsqrt-u32.c",
+-      "src/src/f32-vsqrt/gen/f32-vsqrt-avx512f-rsqrt.c",
 -    ]
-+source_set("x8-packw_x64") {
++source_set("f16-gemm_arch=armv8.2-a+fp16+dotprod") {
 +  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -38074,15 +40852,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
++  asmflags = cflags
  
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x8-packw/gen/x8-packw-x16-gemm-goi-scalar-u2.c",
-+    "src/src/x8-packw/gen/x8-packw-x32-gemm-goi-scalar-u2.c",
-+    "src/src/x8-packw/gen/x8-packw-x4-gemm-goi-scalar-u2.c",
-+    "src/src/x8-packw/gen/x8-packw-x8-gemm-goi-scalar-u2.c"
++    "src/src/f16-gemm/gen/f16-gemm-1x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
++    "src/src/f16-gemm/gen/f16-gemm-4x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
++    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a55.S",
++    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a55r0.S",
++    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a75.S",
++    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-ld64.S"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -38093,11 +40873,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("f32-vrsqrt_f16c-fma-no-avx2") {
+-  source_set("f32-vsqrt_sse2-no-sse3") {
 -    cflags = [
--      "-mf16c",
--      "-mfma",
--      "-mno-avx2",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -38108,7 +40887,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrsqrt/gen/f32-vrsqrt-fma3-rsqrt-u16.c",
+-      "src/src/f32-vsqrt/gen/f32-vsqrt-sse2-sqrt.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -38124,31 +40903,34 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x8-packw_x64_standalone") {
++source_set("f16-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
 +  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  ]
++  asmflags = cflags
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vrsqrt_f16c-fma-no-avx2_standalone") {
+-  source_set("f32-vsqrt_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mf16c",
--      "-mfma",
--      "-mno-avx2",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x8-packw/gen/x8-packw-x16-gemm-goi-scalar-u2.c",
-+    "src/src/x8-packw/gen/x8-packw-x32-gemm-goi-scalar-u2.c",
-+    "src/src/x8-packw/gen/x8-packw-x4-gemm-goi-scalar-u2.c",
-+    "src/src/x8-packw/gen/x8-packw-x8-gemm-goi-scalar-u2.c"
++    "src/src/f16-gemm/gen/f16-gemm-1x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
++    "src/src/f16-gemm/gen/f16-gemm-4x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
++    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a55.S",
++    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a55r0.S",
++    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a75.S",
++    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-ld64.S"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrsqrt/gen/f32-vrsqrt-fma3-rsqrt-u16.c",
+-      "src/src/f32-vsqrt/gen/f32-vsqrt-sse2-sqrt.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -38182,27 +40964,14 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("f32-vrsqrt_sse-no-sse2") {
--    cflags = [
--      "-mno-sse2",
--      "-msse",
--    ]
-+source_set("x8-transposec_f16c-fma-avx2") {
-+  cflags = [
-+    "-mavx2",
-+    "-mf16c",
-+    "-mfma"
-+  ]
- 
+-  source_set("f32-vsqrt_x64") {
+-    cflags = []
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrsqrt/gen/f32-vrsqrt-sse-rsqrt-u8.c",
+-      "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x8-transposec/gen/x8-transposec-32x32-reuse-switch-avx2.c"
-+  ]
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -38213,38 +40982,34 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++source_set("f16-ibilinear-chw_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-ibilinear-chw/gen/f16-ibilinear-chw-neonfp16arith-p8.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vrsqrt_sse-no-sse2_standalone") {
--    cflags = [
--      "-mno-sse2",
--      "-msse",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
+-  source_set("f32-vsqrt_x64_standalone") {
+-    cflags = []
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrsqrt/gen/f32-vrsqrt-sse-rsqrt-u8.c",
+-      "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("x8-transposec_f16c-fma-avx2_standalone") {
-+  cflags = [
-+    "-mavx2",
-+    "-mf16c",
-+    "-mfma"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -38257,90 +41022,45 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x8-transposec/gen/x8-transposec-32x32-reuse-switch-avx2.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
 +}
  
--  source_set("f32-vrsqrt_x64") {
--    cflags = []
-+source_set("x8-transposec_sse2-no-sse3") {
+-    public_configs = [ ":xnnpack_config" ]
++# This is a target that cannot depend on //base.
++source_set("f16-ibilinear-chw_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-march=armv8.2-a+fp16"
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
--      "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u4.c",
--    ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x8-transposec/gen/x8-transposec-16x16-reuse-mov-sse2.c"
++    "src/src/f16-ibilinear-chw/gen/f16-ibilinear-chw-neonfp16arith-p8.c"
 +  ]
  
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-  source_set("f32-vtanh_avx-no-avx2-no-f16c-no-fma") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vtanh/gen/f32-vtanh-avx-rational-9-8-div.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("f32-vrsqrt_x64_standalone") {
--    cflags = []
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
--      "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u4.c",
--    ]
-+# This is a target that cannot depend on //base.
-+source_set("x8-transposec_sse2-no-sse3_standalone") {
-+  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -38351,49 +41071,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x8-transposec/gen/x8-transposec-16x16-reuse-mov-sse2.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
    }
 +}
  
--  source_set("f32-vsigmoid_avx-no-avx2-no-f16c-no-fma") {
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vtanh_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
-+source_set("x8-transposec_x64") {
-+  cflags = [
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx-rr2-p5-nr2-u16.c",
+-      "src/src/f32-vtanh/gen/f32-vtanh-avx-rational-9-8-div.c",
 -    ]
++source_set("f16-ibilinear_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -38404,27 +41107,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x8-transposec/gen/x8-transposec-2x4-scalar-int.c"
++    "src/src/f16-ibilinear/gen/f16-ibilinear-neonfp16arith-c8.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("f32-vsigmoid_avx-no-avx2-no-f16c-no-fma_standalone") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -38432,12 +41130,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-  source_set("f32-vtanh_avx512f") {
+-    cflags = [ "-mavx512f" ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx-rr2-p5-nr2-u16.c",
+-      "src/src/f32-vtanh/gen/f32-vtanh-avx512f-rational-9-8-div.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++# This is a target that cannot depend on //base.
++source_set("f16-ibilinear_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -38447,34 +41153,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("x8-transposec_x64_standalone") {
-+  cflags = [
- 
--    public_configs = [ ":xnnpack_config" ]
-+  ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x8-transposec/gen/x8-transposec-2x4-scalar-int.c"
++    "src/src/f16-ibilinear/gen/f16-ibilinear-neonfp16arith-c8.c"
 +  ]
  
--  source_set("f32-vsigmoid_avx512f") {
--    cflags = [ "-mavx512f" ]
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx512f-rr2-lut32-p2-perm2-scalef-div-u64.c",
--    ]
+-  # This is a target that cannot depend on //base.
+-  source_set("f32-vtanh_avx512f_standalone") {
+-    cflags = [ "-mavx512f" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -38482,34 +41176,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
- 
--  # This is a target that cannot depend on //base.
--  source_set("f32-vsigmoid_avx512f_standalone") {
--    cflags = [ "-mavx512f" ]
--
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx512f-rr2-lut32-p2-perm2-scalef-div-u64.c",
+-      "src/src/f32-vtanh/gen/f32-vtanh-avx512f-rational-9-8-div.c",
 -    ]
-+source_set("xx-copy_x64") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -38521,27 +41192,42 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/xx-copy/xx-copy-scalar-memcpy.c"
++source_set("f16-igemm_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-igemm/gen/f16-igemm-1x16-minmax-neonfp16arith-ld64.c",
++    "src/src/f16-igemm/gen/f16-igemm-1x8-minmax-neonfp16arith-ld64.c",
++    "src/src/f16-igemm/gen/f16-igemm-6x16-minmax-neonfp16arith-ld64.c",
++    "src/src/f16-igemm/gen/f16-igemm-6x8-minmax-neonfp16arith-ld64.c"
++  ]
  
--  source_set("f32-vsigmoid_f16c-fma-avx2") {
+-  source_set("f32-vtanh_f16c-fma-no-avx2") {
 -    cflags = [
--      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
+-      "-mno-avx2",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vtanh/gen/f32-vtanh-fma3-rational-9-8-div.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -38550,13 +41236,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx2-rr1-p5-div-u16.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -38567,29 +41246,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("xx-copy_x64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
++# This is a target that cannot depend on //base.
++source_set("f16-igemm_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vsigmoid_f16c-fma-avx2_standalone") {
+-  source_set("f32-vtanh_f16c-fma-no-avx2_standalone") {
 -    cflags = [
--      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
+-      "-mno-avx2",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/xx-copy/xx-copy-scalar-memcpy.c"
++    "src/src/f16-igemm/gen/f16-igemm-1x16-minmax-neonfp16arith-ld64.c",
++    "src/src/f16-igemm/gen/f16-igemm-1x8-minmax-neonfp16arith-ld64.c",
++    "src/src/f16-igemm/gen/f16-igemm-6x16-minmax-neonfp16arith-ld64.c",
++    "src/src/f16-igemm/gen/f16-igemm-6x8-minmax-neonfp16arith-ld64.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsigmoid/gen/f32-vsigmoid-avx2-rr1-p5-div-u16.c",
+-      "src/src/f32-vtanh/gen/f32-vtanh-fma3-rational-9-8-div.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -38623,119 +41308,21 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("f32-vsigmoid_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
-+source_set("xx-fill_sse2-no-sse3") {
-+  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vsigmoid/gen/f32-vsigmoid-sse2-rr2-lut64-p2-div-u8.c",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/xx-fill/xx-fill-sse2-u64.c"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("f32-vsigmoid_sse2-no-sse3_standalone") {
+-  source_set("f32-vtanh_sse2-no-sse3") {
 -    cflags = [
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vsigmoid/gen/f32-vsigmoid-sse2-rr2-lut64-p2-div-u8.c",
--    ]
-+# This is a target that cannot depend on //base.
-+source_set("xx-fill_sse2-no-sse3_standalone") {
++source_set("f16-igemm_arch=armv8.2-a+fp16+dotprod") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/xx-fill/xx-fill-sse2-u64.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "-march=armv8.2-a+fp16+dotprod"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
- 
--  source_set("f32-vsigmoid_sse4.1-no-sse4.2") {
--    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
--    ]
-+source_set("xx-fill_x64") {
-+  cflags = [
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsigmoid/gen/f32-vsigmoid-sse41-rr2-lut64-p2-div-u8.c",
+-      "src/src/f32-vtanh/gen/f32-vtanh-sse2-rational-9-8-div.c",
 -    ]
-+  ]
++  asmflags = cflags
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -38749,7 +41336,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/xx-fill/xx-fill-scalar-u16.c"
++    "src/src/f16-igemm/f16-igemm-1x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
++    "src/src/f16-igemm/f16-igemm-4x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
++    "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a55.S",
++    "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a55r0.S",
++    "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a75.S",
++    "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-ld64.S"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -38759,10 +41351,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vsigmoid_sse4.1-no-sse4.2_standalone") {
+-  source_set("f32-vtanh_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -38773,7 +41365,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsigmoid/gen/f32-vsigmoid-sse41-rr2-lut64-p2-div-u8.c",
+-      "src/src/f32-vtanh/gen/f32-vtanh-sse2-rational-9-8-div.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -38789,11 +41381,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("xx-fill_x64_standalone") {
++source_set("f16-igemm_arch=armv8.2-a+fp16+dotprod_standalone") {
 +  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  ]
++  asmflags = cflags
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
@@ -38801,10 +41395,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/xx-fill/xx-fill-scalar-u16.c"
++    "src/src/f16-igemm/f16-igemm-1x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
++    "src/src/f16-igemm/f16-igemm-4x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
++    "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a55.S",
++    "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a55r0.S",
++    "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a75.S",
++    "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-ld64.S"
 +  ]
  
--  source_set("f32-vsigmoid_x64") {
+-  source_set("f32-vtanh_x64") {
 -    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -38812,7 +41411,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c",
+-      "src/src/f32-vtanh/gen/f32-vtanh-scalar-rational-9-8-div.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -38840,21 +41439,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vsigmoid_x64_standalone") {
+-  source_set("f32-vtanh_x64_standalone") {
 -    cflags = []
-+source_set("xx-pad_sse2-no-sse3") {
++source_set("f16-maxpool_arch=armv8.2-a+fp16") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c",
+-      "src/src/f32-vtanh/gen/f32-vtanh-scalar-rational-9-8-div.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/xx-pad/xx-pad-p16-sse2-u16.c"
++    "src/src/f16-maxpool/gen/f16-maxpool-9p-minmax-neonfp16arith-u8.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -38886,7 +41484,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
--  source_set("f32-vsin_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("f32-vunary_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
@@ -38894,19 +41492,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-fma",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("xx-pad_sse2-no-sse3_standalone") {
++source_set("f16-maxpool_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
-+    "-mno-sse3",
-+    "-msse2"
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsin/gen/f32-vsin-avx-rational-5-4-div.c",
+-      "src/src/f32-vunary/gen/f32-vabs-avx.c",
+-      "src/src/f32-vunary/gen/f32-vneg-avx.c",
+-      "src/src/f32-vunary/gen/f32-vsqr-avx.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/xx-pad/xx-pad-p16-sse2-u16.c"
++    "src/src/f16-maxpool/gen/f16-maxpool-9p-minmax-neonfp16arith-u8.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -38933,7 +41532,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vsin_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("f32-vunary_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
@@ -38944,7 +41543,9 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsin/gen/f32-vsin-avx-rational-5-4-div.c",
+-      "src/src/f32-vunary/gen/f32-vabs-avx.c",
+-      "src/src/f32-vunary/gen/f32-vneg-avx.c",
+-      "src/src/f32-vunary/gen/f32-vsqr-avx.c",
 -    ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
@@ -38961,31 +41562,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("xx-pad_x64") {
++source_set("f16-qs8-vcvt_arch=armv8.2-a+fp16") {
 +  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-qs8-vcvt/gen/f16-qs8-vcvt-neonfp16arith-u32.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/xx-pad/xx-pad-p4-scalar-u16.c"
-+  ]
- 
--  source_set("f32-vsin_avx512f") {
--    cflags = [ "-mavx512f" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vsin/gen/f32-vsin-avx512f-rational-5-4-nr.c",
--    ]
+-  source_set("f32-vunary_avx512f") {
+-    cflags = [ "-mavx512f" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -38993,6 +41590,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/f32-vunary/gen/f32-vabs-avx512f.c",
+-      "src/src/f32-vunary/gen/f32-vneg-avx512f.c",
+-      "src/src/f32-vunary/gen/f32-vsqr-avx512f.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -39003,28 +41609,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++# This is a target that cannot depend on //base.
++source_set("f16-qs8-vcvt_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+# This is a target that cannot depend on //base.
-+source_set("xx-pad_x64_standalone") {
-+  cflags = [
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-qs8-vcvt/gen/f16-qs8-vcvt-neonfp16arith-u32.c"
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vsin_avx512f_standalone") {
+-  source_set("f32-vunary_avx512f_standalone") {
 -    cflags = [ "-mavx512f" ]
-+  ]
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsin/gen/f32-vsin-avx512f-rational-5-4-nr.c",
+-      "src/src/f32-vunary/gen/f32-vabs-avx512f.c",
+-      "src/src/f32-vunary/gen/f32-vneg-avx512f.c",
+-      "src/src/f32-vunary/gen/f32-vsqr-avx512f.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/xx-pad/xx-pad-p4-scalar-u16.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -39036,10 +41646,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -39058,18 +41664,19 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("f32-vsin_f16c-fma-no-avx2") {
+-  source_set("f32-vunary_sse2-no-sse3") {
 -    cflags = [
--      "-mf16c",
--      "-mfma",
--      "-mno-avx2",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+source_set("xx-transposev_x64") {
++source_set("f16-qs8-vcvt_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsin/gen/f32-vsin-fma3-rational-5-4-div.c",
+-      "src/src/f32-vunary/gen/f32-vabs-sse2.c",
+-      "src/src/f32-vunary/gen/f32-vneg-sse2.c",
+-      "src/src/f32-vunary/gen/f32-vsqr-sse2.c",
 -    ]
 +  ]
  
@@ -39085,7 +41692,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/xx-transposev/xx-transposev-1x1-scalar-memcpy.c"
++    "src/src/f16-qs8-vcvt/gen/f16-qs8-vcvt-scalar-imagic-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -39095,11 +41702,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vsin_f16c-fma-no-avx2_standalone") {
+-  source_set("f32-vunary_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mf16c",
--      "-mfma",
--      "-mno-avx2",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -39110,7 +41716,9 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsin/gen/f32-vsin-fma3-rational-5-4-div.c",
+-      "src/src/f32-vunary/gen/f32-vabs-sse2.c",
+-      "src/src/f32-vunary/gen/f32-vneg-sse2.c",
+-      "src/src/f32-vunary/gen/f32-vsqr-sse2.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -39126,7 +41734,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("xx-transposev_x64_standalone") {
++source_set("f16-qs8-vcvt_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -39138,21 +41746,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/xx-transposev/xx-transposev-1x1-scalar-memcpy.c"
++    "src/src/f16-qs8-vcvt/gen/f16-qs8-vcvt-scalar-imagic-u4.c"
 +  ]
  
--  source_set("f32-vsin_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("f32-vunary_x64") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsin/gen/f32-vsin-sse2-rational-5-4-div.c",
+-      "src/src/f32-vunary/gen/f32-vabs-scalar.c",
+-      "src/src/f32-vunary/gen/f32-vneg-scalar.c",
+-      "src/src/f32-vunary/gen/f32-vsqr-scalar.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -39180,21 +41787,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vsin_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
-+}
- 
+-  source_set("f32-vunary_x64_standalone") {
+-    cflags = []
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsin/gen/f32-vsin-sse2-rational-5-4-div.c",
+-      "src/src/f32-vunary/gen/f32-vabs-scalar.c",
+-      "src/src/f32-vunary/gen/f32-vneg-scalar.c",
+-      "src/src/f32-vunary/gen/f32-vsqr-scalar.c",
 -    ]
-+if (current_cpu == "arm64") {
-+source_set("configs_arm64") {
-+  cflags = [
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -39205,160 +41807,47 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
++source_set("f16-qu8-vcvt_arm64") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/configs/argmaxpool-config.c",
-+    "src/src/configs/avgpool-config.c",
-+    "src/src/configs/binary-elementwise-config.c",
-+    "src/src/configs/cmul-config.c",
-+    "src/src/configs/conv-hwc2chw-config.c",
-+    "src/src/configs/dwconv-config.c",
-+    "src/src/configs/dwconv2d-chw-config.c",
-+    "src/src/configs/gemm-config.c",
-+    "src/src/configs/hardware-config.c",
-+    "src/src/configs/ibilinear-chw-config.c",
-+    "src/src/configs/ibilinear-config.c",
-+    "src/src/configs/lut32norm-config.c",
-+    "src/src/configs/maxpool-config.c",
-+    "src/src/configs/pack-lh-config.c",
-+    "src/src/configs/raddstoreexpminusmax-config.c",
-+    "src/src/configs/reduce-config.c",
-+    "src/src/configs/spmm-config.c",
-+    "src/src/configs/transpose-config.c",
-+    "src/src/configs/unary-elementwise-config.c",
-+    "src/src/configs/unpool-config.c",
-+    "src/src/configs/vmulcaddc-config.c",
-+    "src/src/configs/x8-lut-config.c",
-+    "src/src/configs/xx-fill-config.c",
-+    "src/src/configs/xx-pad-config.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--  source_set("f32-vsin_x64") {
--    cflags = []
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vsin/gen/f32-vsin-scalar-rational-5-4-div.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
 -
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+# This is a target that cannot depend on //base.
-+source_set("configs_arm64_standalone") {
-+  cflags = [
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
-+  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("f32-vsin_x64_standalone") {
+-  source_set("operators_x64") {
 -    cflags = []
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/configs/argmaxpool-config.c",
-+    "src/src/configs/avgpool-config.c",
-+    "src/src/configs/binary-elementwise-config.c",
-+    "src/src/configs/cmul-config.c",
-+    "src/src/configs/conv-hwc2chw-config.c",
-+    "src/src/configs/dwconv-config.c",
-+    "src/src/configs/dwconv2d-chw-config.c",
-+    "src/src/configs/gemm-config.c",
-+    "src/src/configs/hardware-config.c",
-+    "src/src/configs/ibilinear-chw-config.c",
-+    "src/src/configs/ibilinear-config.c",
-+    "src/src/configs/lut32norm-config.c",
-+    "src/src/configs/maxpool-config.c",
-+    "src/src/configs/pack-lh-config.c",
-+    "src/src/configs/raddstoreexpminusmax-config.c",
-+    "src/src/configs/reduce-config.c",
-+    "src/src/configs/spmm-config.c",
-+    "src/src/configs/transpose-config.c",
-+    "src/src/configs/unary-elementwise-config.c",
-+    "src/src/configs/unpool-config.c",
-+    "src/src/configs/vmulcaddc-config.c",
-+    "src/src/configs/x8-lut-config.c",
-+    "src/src/configs/xx-fill-config.c",
-+    "src/src/configs/xx-pad-config.c"
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vsin/gen/f32-vsin-scalar-rational-5-4-div.c",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
 -
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
- 
--  source_set("f32-vsqrt_avx-no-avx2-no-f16c-no-fma") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
-+source_set("enums_arm64") {
-+  cflags = [
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsqrt/gen/f32-vsqrt-avx-rsqrt-u16.c",
+-      "src/src/operators/argmax-pooling-nhwc.c",
+-      "src/src/operators/average-pooling-nhwc.c",
+-      "src/src/operators/batch-matrix-multiply-nc.c",
+-      "src/src/operators/binary-elementwise-nd.c",
+-      "src/src/operators/constant-pad-nd.c",
+-      "src/src/operators/convolution-nchw.c",
+-      "src/src/operators/convolution-nhwc.c",
+-      "src/src/operators/deconvolution-nhwc.c",
+-      "src/src/operators/dynamic-fully-connected-nc.c",
+-      "src/src/operators/fully-connected-nc.c",
+-      "src/src/operators/max-pooling-nhwc.c",
+-      "src/src/operators/pack-lh.c",
+-      "src/src/operators/reduce-nd.c",
+-      "src/src/operators/resize-bilinear-nchw.c",
+-      "src/src/operators/resize-bilinear-nhwc.c",
+-      "src/src/operators/rope-nthc.c",
+-      "src/src/operators/slice-nd.c",
+-      "src/src/operators/softmax-nc.c",
+-      "src/src/operators/transpose-nd.c",
+-      "src/src/operators/unary-elementwise-nc.c",
+-      "src/src/operators/unpooling-nhwc.c",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-qu8-vcvt/gen/f16-qu8-vcvt-scalar-imagic-u4.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -39371,29 +41860,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/enums/allocation-type.c",
-+    "src/src/enums/datatype-strings.c",
-+    "src/src/enums/microkernel-type.c",
-+    "src/src/enums/node-type.c",
-+    "src/src/enums/operator-type.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("f32-vsqrt_avx-no-avx2-no-f16c-no-fma_standalone") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -39401,9 +41873,33 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-  # This is a target that cannot depend on //base.
+-  source_set("operators_x64_standalone") {
+-    cflags = []
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsqrt/gen/f32-vsqrt-avx-rsqrt-u16.c",
+-      "src/src/operators/argmax-pooling-nhwc.c",
+-      "src/src/operators/average-pooling-nhwc.c",
+-      "src/src/operators/batch-matrix-multiply-nc.c",
+-      "src/src/operators/binary-elementwise-nd.c",
+-      "src/src/operators/constant-pad-nd.c",
+-      "src/src/operators/convolution-nchw.c",
+-      "src/src/operators/convolution-nhwc.c",
+-      "src/src/operators/deconvolution-nhwc.c",
+-      "src/src/operators/dynamic-fully-connected-nc.c",
+-      "src/src/operators/fully-connected-nc.c",
+-      "src/src/operators/max-pooling-nhwc.c",
+-      "src/src/operators/pack-lh.c",
+-      "src/src/operators/reduce-nd.c",
+-      "src/src/operators/resize-bilinear-nchw.c",
+-      "src/src/operators/resize-bilinear-nhwc.c",
+-      "src/src/operators/rope-nthc.c",
+-      "src/src/operators/slice-nd.c",
+-      "src/src/operators/softmax-nc.c",
+-      "src/src/operators/transpose-nd.c",
+-      "src/src/operators/unary-elementwise-nc.c",
+-      "src/src/operators/unpooling-nhwc.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -39419,7 +41915,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("enums_arm64_standalone") {
++source_set("f16-qu8-vcvt_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -39431,22 +41927,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/enums/allocation-type.c",
-+    "src/src/enums/datatype-strings.c",
-+    "src/src/enums/microkernel-type.c",
-+    "src/src/enums/node-type.c",
-+    "src/src/enums/operator-type.c"
++    "src/src/f16-qu8-vcvt/gen/f16-qu8-vcvt-scalar-imagic-u4.c"
 +  ]
  
--  source_set("f32-vsqrt_avx512f") {
--    cflags = [ "-mavx512f" ]
+-  source_set("qd8-f16-qb4w-gemm_f16c-fma-avx2") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsqrt/gen/f32-vsqrt-avx512f-rsqrt-u16.c",
+-      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x8c8-minmax-avx2.c",
+-      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-3x8c8-minmax-avx2.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -39474,109 +41971,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vsqrt_avx512f_standalone") {
--    cflags = [ "-mavx512f" ]
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vsqrt/gen/f32-vsqrt-avx512f-rsqrt-u16.c",
--    ]
-+source_set("f16-avgpool_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-avgpool/gen/f16-avgpool-9p-minmax-neonfp16arith.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--  source_set("f32-vsqrt_f16c-fma-no-avx2") {
+-  source_set("qd8-f16-qb4w-gemm_f16c-fma-avx2_standalone") {
 -    cflags = [
+-      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
--      "-mno-avx2",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vsqrt/gen/f32-vsqrt-fma3-rsqrt-u16.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-avgpool_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
 -
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-avgpool/gen/f16-avgpool-9p-minmax-neonfp16arith.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("f32-vsqrt_f16c-fma-no-avx2_standalone") {
--    cflags = [
--      "-mf16c",
--      "-mfma",
--      "-mno-avx2",
--    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsqrt/gen/f32-vsqrt-fma3-rsqrt-u16.c",
+-      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x8c8-minmax-avx2.c",
+-      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-3x8c8-minmax-avx2.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -39587,13 +41994,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
- 
+-
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("f16-conv-hwc2chw_arch=armv8.2-a+fp16") {
++source_set("f16-raddstoreexpminusmax_arch=armv8.2-a+fp16") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
@@ -39604,13 +42007,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-conv-hwc2chw/f16-conv-hwc2chw-3x3s2p1c3x4-neonfp16arith-2x2.c"
++    "src/src/f16-raddstoreexpminusmax/gen/f16-raddstoreexpminusmax-neonfp16arith-rr2-p2-u32.c"
 +  ]
  
--  source_set("f32-vsqrt_sse-no-sse2") {
+-  source_set("qd8-f16-qc4w-gemm_avx2-avxvnni-f16c-fma") {
 -    cflags = [
--      "-mno-sse2",
--      "-msse",
+-      "-mavx2",
+-      "-mavxvnni",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -39618,7 +42023,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsqrt/gen/f32-vsqrt-sse-rsqrt-u12.c",
+-      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-5x8c8-minmax-avxvnni-prfm.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -39643,25 +42049,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("f16-conv-hwc2chw_arch=armv8.2-a+fp16_standalone") {
++source_set("f16-raddstoreexpminusmax_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vsqrt_sse-no-sse2_standalone") {
+-  source_set("qd8-f16-qc4w-gemm_avx2-avxvnni-f16c-fma_standalone") {
 -    cflags = [
--      "-mno-sse2",
--      "-msse",
+-      "-mavx2",
+-      "-mavxvnni",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-conv-hwc2chw/f16-conv-hwc2chw-3x3s2p1c3x4-neonfp16arith-2x2.c"
++    "src/src/f16-raddstoreexpminusmax/gen/f16-raddstoreexpminusmax-neonfp16arith-rr2-p2-u32.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsqrt/gen/f32-vsqrt-sse-rsqrt-u12.c",
+-      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-5x8c8-minmax-avxvnni-prfm.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -39695,60 +42104,19 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("f32-vsqrt_x64") {
--    cflags = []
-+source_set("f16-dwconv2d-chw_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt-u1.c",
+-  source_set("qd8-f16-qc4w-gemm_f16c-fma-avx2") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-dwconv2d-chw/gen/f16-dwconv2d-chw-3x3p1-minmax-neonfp16arith-2x8.c",
-+    "src/src/f16-dwconv2d-chw/gen/f16-dwconv2d-chw-3x3s2p1-minmax-neonfp16arith-1x8.c",
-+    "src/src/f16-dwconv2d-chw/gen/f16-dwconv2d-chw-5x5p2-minmax-neonfp16arith-1x8.c",
-+    "src/src/f16-dwconv2d-chw/gen/f16-dwconv2d-chw-5x5s2p2-minmax-neonfp16arith-1x8.c"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
 -
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("f32-vsqrt_x64_standalone") {
--    cflags = []
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt-u1.c",
+-      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x8c8-minmax-avx2-madd-prfm.c",
+-      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-4x8c8-minmax-avx2-madd-prfm.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-dwconv2d-chw_arch=armv8.2-a+fp16_standalone") {
++source_set("f16-rdminmax_arch=armv8.2-a+fp16") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
@@ -39761,48 +42129,40 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-dwconv2d-chw/gen/f16-dwconv2d-chw-3x3p1-minmax-neonfp16arith-2x8.c",
-+    "src/src/f16-dwconv2d-chw/gen/f16-dwconv2d-chw-3x3s2p1-minmax-neonfp16arith-1x8.c",
-+    "src/src/f16-dwconv2d-chw/gen/f16-dwconv2d-chw-5x5p2-minmax-neonfp16arith-1x8.c",
-+    "src/src/f16-dwconv2d-chw/gen/f16-dwconv2d-chw-5x5s2p2-minmax-neonfp16arith-1x8.c"
++    "src/src/f16-rdminmax/gen/f16-rdmax-2p2x-neonfp16arith-c32.c",
++    "src/src/f16-rdminmax/gen/f16-rdmin-2p2x-neonfp16arith-c32.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f16-qc4w-gemm_f16c-fma-avx2_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
--  source_set("f32-vtanh_avx-no-avx2-no-f16c-no-fma") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vtanh/gen/f32-vtanh-avx-rational-9-8-div.c",
+-      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x8c8-minmax-avx2-madd-prfm.c",
+-      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-4x8c8-minmax-avx2-madd-prfm.c",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
++  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -39813,89 +42173,45 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("f16-dwconv_arch=armv8.2-a+fp16") {
++# This is a target that cannot depend on //base.
++source_set("f16-rdminmax_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-dwconv/gen/f16-dwconv-25p8c-minmax-neonfp16arith-acc2.c",
-+    "src/src/f16-dwconv/gen/f16-dwconv-3p16c-minmax-neonfp16arith.c",
-+    "src/src/f16-dwconv/gen/f16-dwconv-4p16c-minmax-neonfp16arith.c",
-+    "src/src/f16-dwconv/gen/f16-dwconv-9p16c-minmax-neonfp16arith.c",
-+    "src/src/f16-dwconv/gen/f16-dwconv-9p8c-minmax-neonfp16arith.c"
-+  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("f32-vtanh_avx-no-avx2-no-f16c-no-fma_standalone") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vtanh/gen/f32-vtanh-avx-rational-9-8-div.c",
--    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-dwconv_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
++    "src/src/f16-rdminmax/gen/f16-rdmax-2p2x-neonfp16arith-c32.c",
++    "src/src/f16-rdminmax/gen/f16-rdmin-2p2x-neonfp16arith-c32.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-dwconv/gen/f16-dwconv-25p8c-minmax-neonfp16arith-acc2.c",
-+    "src/src/f16-dwconv/gen/f16-dwconv-3p16c-minmax-neonfp16arith.c",
-+    "src/src/f16-dwconv/gen/f16-dwconv-4p16c-minmax-neonfp16arith.c",
-+    "src/src/f16-dwconv/gen/f16-dwconv-9p16c-minmax-neonfp16arith.c",
-+    "src/src/f16-dwconv/gen/f16-dwconv-9p8c-minmax-neonfp16arith.c"
-+  ]
- 
--  source_set("f32-vtanh_avx512f") {
--    cflags = [ "-mavx512f" ]
+-
+-  source_set(
+-      "qd8-f16-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vtanh/gen/f32-vtanh-avx512f-rational-9-8-nr.c",
+-      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x8c8-minmax-avx256skx-madd-prfm.c",
+-      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-8x8c8-minmax-avx256skx-madd-prfm.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -39923,15 +42239,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vtanh_avx512f_standalone") {
--    cflags = [ "-mavx512f" ]
--
+-  source_set(
+-      "qd8-f16-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++source_set("f16-rdminmax_arm64") {
++  cflags = [
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vtanh/gen/f32-vtanh-avx512f-rational-9-8-nr.c",
+-      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x8c8-minmax-avx256skx-madd-prfm.c",
+-      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-8x8c8-minmax-avx256skx-madd-prfm.c",
 -    ]
-+source_set("f16-f32-vcvt_arm64") {
-+  cflags = [
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -39943,29 +42270,33 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-f32-vcvt/gen/f16-f32-vcvt-neon-int16-u16.c",
-+    "src/src/f16-f32-vcvt/gen/f16-f32-vcvt-neonfp16-u16.c",
-+    "src/src/f16-f32-vcvt/gen/f16-f32-vcvt-scalar-u4.c"
++    "src/src/f16-rdminmax/gen/f16-rdmax-2p2x-scalar-c2.c",
++    "src/src/f16-rdminmax/gen/f16-rdmin-2p2x-scalar-c2.c"
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("f32-vtanh_f16c-fma-no-avx2") {
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
+-
+-  source_set(
+-      "qd8-f16-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
 -    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
--      "-mno-avx2",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -39976,7 +42307,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vtanh/gen/f32-vtanh-fma3-rational-9-8-div.c",
+-      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x8c8-minmax-avx256vnni.c",
+-      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-8x8c8-minmax-avx256vnni.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -39992,7 +42324,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-f32-vcvt_arm64_standalone") {
++source_set("f16-rdminmax_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -40000,22 +42332,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vtanh_f16c-fma-no-avx2_standalone") {
+-  source_set(
+-      "qd8-f16-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
 -    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
--      "-mno-avx2",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-f32-vcvt/gen/f16-f32-vcvt-neon-int16-u16.c",
-+    "src/src/f16-f32-vcvt/gen/f16-f32-vcvt-neonfp16-u16.c",
-+    "src/src/f16-f32-vcvt/gen/f16-f32-vcvt-scalar-u4.c"
++    "src/src/f16-rdminmax/gen/f16-rdmax-2p2x-scalar-c2.c",
++    "src/src/f16-rdminmax/gen/f16-rdmin-2p2x-scalar-c2.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vtanh/gen/f32-vtanh-fma3-rational-9-8-div.c",
+-      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x8c8-minmax-avx256vnni.c",
+-      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-8x8c8-minmax-avx256vnni.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -40049,111 +42387,20 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("f32-vtanh_sse2-no-sse3") {
+-  source_set("qd8-f16-qc8w-gemm_avx2-avxvnni-f16c-fma") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
-+source_set("f16-f32acc-rdsum_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vtanh/gen/f32-vtanh-sse2-rational-9-8-div.c",
+-      "-mavx2",
+-      "-mavxvnni",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-f32acc-rdsum/gen/f16-f32acc-rdsum-7p7x-minmax-neonfp16arith-c16.c"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
 -
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("f32-vtanh_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vtanh/gen/f32-vtanh-sse2-rational-9-8-div.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-5x8c8-minmax-avxvnni-prfm.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-f32acc-rdsum_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
 -
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-f32acc-rdsum/gen/f16-f32acc-rdsum-7p7x-minmax-neonfp16arith-c16.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--  source_set("f32-vtanh_x64") {
--    cflags = []
-+  public_configs = [ ":xnnpack_config" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vtanh/gen/f32-vtanh-scalar-rational-9-8-div.c",
--    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -40164,7 +42411,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("f16-f32acc-rsum_arch=armv8.2-a+fp16") {
++source_set("f16-rminmax_arch=armv8.2-a+fp16") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
@@ -40173,19 +42420,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-f32acc-rsum/gen/f16-f32acc-rsum-neonfp16arith-u32-acc4.c"
++    "src/src/f16-rminmax/gen/f16-rmax-neonfp16arith-u32-acc4.c",
++    "src/src/f16-rminmax/gen/f16-rmin-neonfp16arith-u32-acc4.c",
++    "src/src/f16-rminmax/gen/f16-rminmax-neonfp16arith-u32-acc4.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vtanh_x64_standalone") {
--    cflags = []
+-  source_set("qd8-f16-qc8w-gemm_avx2-avxvnni-f16c-fma_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mavxvnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vtanh/gen/f32-vtanh-scalar-rational-9-8-div.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-5x8c8-minmax-avxvnni-prfm.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -40209,7 +42464,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-f32acc-rsum_arch=armv8.2-a+fp16_standalone") {
++source_set("f16-rminmax_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
@@ -40220,15 +42475,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-f32acc-rsum/gen/f16-f32acc-rsum-neonfp16arith-u32-acc4.c"
++    "src/src/f16-rminmax/gen/f16-rmax-neonfp16arith-u32-acc4.c",
++    "src/src/f16-rminmax/gen/f16-rmin-neonfp16arith-u32-acc4.c",
++    "src/src/f16-rminmax/gen/f16-rminmax-neonfp16arith-u32-acc4.c"
 +  ]
  
--  source_set("f32-vunary_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qd8-f16-qc8w-gemm_f16c-fma-avx2") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -40236,9 +42492,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vunary/gen/f32-vabs-avx.c",
--      "src/src/f32-vunary/gen/f32-vneg-avx.c",
--      "src/src/f32-vunary/gen/f32-vsqr-avx.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c8-minmax-avx2.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-3x8c8-minmax-avx2.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -40266,125 +42521,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vunary_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("qd8-f16-qc8w-gemm_f16c-fma-avx2_standalone") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
-+source_set("f16-gemm_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vunary/gen/f32-vabs-avx.c",
--      "src/src/f32-vunary/gen/f32-vneg-avx.c",
--      "src/src/f32-vunary/gen/f32-vsqr-avx.c",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-gemm/gen/f16-gemm-1x16-minmax-neonfp16arith-ld64.c",
-+    "src/src/f16-gemm/gen/f16-gemm-1x8-minmax-neonfp16arith-ld64.c",
-+    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-neonfp16arith-ld64.c",
-+    "src/src/f16-gemm/gen/f16-gemm-6x8-minmax-neonfp16arith-ld64.c"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--  source_set("f32-vunary_avx512f") {
--    cflags = [ "-mavx512f" ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-gemm_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/f32-vunary/gen/f32-vabs-avx512f.c",
--      "src/src/f32-vunary/gen/f32-vneg-avx512f.c",
--      "src/src/f32-vunary/gen/f32-vsqr-avx512f.c",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-gemm/gen/f16-gemm-1x16-minmax-neonfp16arith-ld64.c",
-+    "src/src/f16-gemm/gen/f16-gemm-1x8-minmax-neonfp16arith-ld64.c",
-+    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-neonfp16arith-ld64.c",
-+    "src/src/f16-gemm/gen/f16-gemm-6x8-minmax-neonfp16arith-ld64.c"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
++source_set("f16-rminmax_arm64") {
++  cflags = [
  
--  # This is a target that cannot depend on //base.
--  source_set("f32-vunary_avx512f_standalone") {
--    cflags = [ "-mavx512f" ]
--
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vunary/gen/f32-vabs-avx512f.c",
--      "src/src/f32-vunary/gen/f32-vneg-avx512f.c",
--      "src/src/f32-vunary/gen/f32-vsqr-avx512f.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c8-minmax-avx2.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-3x8c8-minmax-avx2.c",
 -    ]
-+source_set("f16-gemm_arch=armv8.2-a+fp16+dotprod") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -40397,31 +42547,33 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  asmflags = cflags
- 
--    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-gemm/gen/f16-gemm-1x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
-+    "src/src/f16-gemm/gen/f16-gemm-4x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
-+    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a55.S",
-+    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a55r0.S",
-+    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a75.S",
-+    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-ld64.S"
++    "src/src/f16-rminmax/gen/f16-rmax-scalar-u2-acc2.c",
++    "src/src/f16-rminmax/gen/f16-rmin-scalar-u2-acc2.c",
++    "src/src/f16-rminmax/gen/f16-rminmax-scalar-u2-acc2.c"
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("f32-vunary_sse2-no-sse3") {
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
+-
+-  source_set(
+-      "qd8-f16-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -40432,9 +42584,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vunary/gen/f32-vabs-sse2.c",
--      "src/src/f32-vunary/gen/f32-vneg-sse2.c",
--      "src/src/f32-vunary/gen/f32-vsqr-sse2.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c8-minmax-avx256skx.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-5x8c8-minmax-avx256skx.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -40450,36 +42601,36 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
++source_set("f16-rminmax_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  asmflags = cflags
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("f32-vunary_sse2-no-sse3_standalone") {
+-  source_set(
+-      "qd8-f16-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-gemm/gen/f16-gemm-1x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
-+    "src/src/f16-gemm/gen/f16-gemm-4x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
-+    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a55.S",
-+    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a55r0.S",
-+    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a75.S",
-+    "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-ld64.S"
++    "src/src/f16-rminmax/gen/f16-rmax-scalar-u2-acc2.c",
++    "src/src/f16-rminmax/gen/f16-rmin-scalar-u2-acc2.c",
++    "src/src/f16-rminmax/gen/f16-rminmax-scalar-u2-acc2.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vunary/gen/f32-vabs-sse2.c",
--      "src/src/f32-vunary/gen/f32-vneg-sse2.c",
--      "src/src/f32-vunary/gen/f32-vsqr-sse2.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c8-minmax-avx256skx.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-5x8c8-minmax-avx256skx.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -40508,27 +42659,33 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
+-  }
+-
+-  source_set(
+-      "qd8-f16-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-   }
++  }
 +}
  
--  source_set("f32-vunary_x64") {
--    cflags = []
-+source_set("f16-ibilinear-chw_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vunary/gen/f32-vabs-scalar.c",
--      "src/src/f32-vunary/gen/f32-vneg-scalar.c",
--      "src/src/f32-vunary/gen/f32-vsqr-scalar.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c8-minmax-avx256vnni.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-8x8c8-minmax-avx256vnni.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-ibilinear-chw/gen/f16-ibilinear-chw-neonfp16arith-p8.c"
++source_set("f16-spmm_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -40541,12 +42698,30 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-spmm/gen/f16-spmm-32x1-minmax-neonfp16arith-pipelined.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set(
+-      "qd8-f16-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -40554,15 +42729,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("f32-vunary_x64_standalone") {
--    cflags = []
--
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vunary/gen/f32-vabs-scalar.c",
--      "src/src/f32-vunary/gen/f32-vneg-scalar.c",
--      "src/src/f32-vunary/gen/f32-vsqr-scalar.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c8-minmax-avx256vnni.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-8x8c8-minmax-avx256vnni.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -40578,7 +42748,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-ibilinear-chw_arch=armv8.2-a+fp16_standalone") {
++source_set("f16-spmm_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
@@ -40586,7 +42756,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-ibilinear-chw/gen/f16-ibilinear-chw-neonfp16arith-p8.c"
++    "src/src/f16-spmm/gen/f16-spmm-32x1-minmax-neonfp16arith-pipelined.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -40594,37 +42764,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    }
 -  }
 -
--  source_set("operators_x64") {
--    cflags = []
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/operators/argmax-pooling-nhwc.c",
--      "src/src/operators/average-pooling-nhwc.c",
--      "src/src/operators/batch-matrix-multiply-nc.c",
--      "src/src/operators/binary-elementwise-nd.c",
--      "src/src/operators/constant-pad-nd.c",
--      "src/src/operators/convolution-nchw.c",
--      "src/src/operators/convolution-nhwc.c",
--      "src/src/operators/deconvolution-nhwc.c",
--      "src/src/operators/dynamic-fully-connected-nc.c",
--      "src/src/operators/fully-connected-nc.c",
--      "src/src/operators/max-pooling-nhwc.c",
--      "src/src/operators/pack-lh.c",
--      "src/src/operators/reduce-nd.c",
--      "src/src/operators/resize-bilinear-nchw.c",
--      "src/src/operators/resize-bilinear-nhwc.c",
--      "src/src/operators/rope-nthc.c",
--      "src/src/operators/slice-nd.c",
--      "src/src/operators/softmax-nc.c",
--      "src/src/operators/transpose-nd.c",
--      "src/src/operators/unary-elementwise-nc.c",
--      "src/src/operators/unpooling-nhwc.c",
+-  source_set(
+-      "qd8-f16-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8") {
+-    cflags = [
+-      "-mamx-int8",
+-      "-mamx-tile",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
+-      "-mgfni",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-16x64c4-minmax-avx512amx.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x64c4-minmax-avx512amx.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -40635,13 +42805,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
 +  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -40651,38 +42814,36 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("operators_x64_standalone") {
--    cflags = []
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/operators/argmax-pooling-nhwc.c",
--      "src/src/operators/average-pooling-nhwc.c",
--      "src/src/operators/batch-matrix-multiply-nc.c",
--      "src/src/operators/binary-elementwise-nd.c",
--      "src/src/operators/constant-pad-nd.c",
--      "src/src/operators/convolution-nchw.c",
--      "src/src/operators/convolution-nhwc.c",
--      "src/src/operators/deconvolution-nhwc.c",
--      "src/src/operators/dynamic-fully-connected-nc.c",
--      "src/src/operators/fully-connected-nc.c",
--      "src/src/operators/max-pooling-nhwc.c",
--      "src/src/operators/pack-lh.c",
--      "src/src/operators/reduce-nd.c",
--      "src/src/operators/resize-bilinear-nchw.c",
--      "src/src/operators/resize-bilinear-nhwc.c",
--      "src/src/operators/rope-nthc.c",
--      "src/src/operators/slice-nd.c",
--      "src/src/operators/softmax-nc.c",
--      "src/src/operators/transpose-nd.c",
--      "src/src/operators/unary-elementwise-nc.c",
--      "src/src/operators/unpooling-nhwc.c",
+-  source_set(
+-      "qd8-f16-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8_standalone") {
+-    cflags = [
+-      "-mamx-int8",
+-      "-mamx-tile",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
+-      "-mgfni",
 -    ]
-+source_set("f16-ibilinear_arch=armv8.2-a+fp16") {
++source_set("f16-vapproxgelu_arch=armv8.2-a+fp16") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-16x64c4-minmax-avx512amx.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x64c4-minmax-avx512amx.c",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vapproxgelu/gen/f16-vapproxgelu-neonfp16arith-rational-6-4-div.c"
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -40693,20 +42854,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-ibilinear/gen/f16-ibilinear-neonfp16arith-c8.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -40714,24 +42866,34 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  source_set("qd8-f16-qb4w-gemm_f16c-fma-avx2") {
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-  source_set("qd8-f16-qc8w-igemm_avx2-avxvnni-f16c-fma") {
 -    cflags = [
 -      "-mavx2",
+-      "-mavxvnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++# This is a target that cannot depend on //base.
++source_set("f16-vapproxgelu_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x8c8-minmax-avx2.c",
--      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-3x8c8-minmax-avx2.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-5x8c8-minmax-avxvnni-prfm.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-ibilinear_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vapproxgelu/gen/f16-vapproxgelu-neonfp16arith-rational-6-4-div.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -40744,24 +42906,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-ibilinear/gen/f16-ibilinear-neonfp16arith-c8.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qd8-f16-qb4w-gemm_f16c-fma-avx2_standalone") {
--    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -40769,12 +42919,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f16-qc8w-igemm_avx2-avxvnni-f16c-fma_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mavxvnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x8c8-minmax-avx2.c",
--      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-3x8c8-minmax-avx2.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-5x8c8-minmax-avxvnni-prfm.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -40786,15 +42949,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++source_set("f16-vapproxgelu_arm64") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("f16-igemm_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -40803,16 +42961,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-igemm/gen/f16-igemm-1x16-minmax-neonfp16arith-ld64.c",
-+    "src/src/f16-igemm/gen/f16-igemm-1x8-minmax-neonfp16arith-ld64.c",
-+    "src/src/f16-igemm/gen/f16-igemm-6x16-minmax-neonfp16arith-ld64.c",
-+    "src/src/f16-igemm/gen/f16-igemm-6x8-minmax-neonfp16arith-ld64.c"
++    "src/src/f16-vapproxgelu/gen/f16-vapproxgelu-scalar-rational-6-4-div.c"
 +  ]
  
--  source_set("qd8-f16-qc4w-gemm_avx2-avxvnni-f16c-fma") {
+-  source_set("qd8-f16-qc8w-igemm_f16c-fma-avx2") {
 -    cflags = [
 -      "-mavx2",
--      "-mavxvnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
@@ -40822,8 +42976,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x8c8-minmax-avxvnni-prfm.c",
--      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-5x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c8-minmax-avx2.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-3x8c8-minmax-avx2.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -40848,35 +43002,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("f16-igemm_arch=armv8.2-a+fp16_standalone") {
++source_set("f16-vapproxgelu_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f16-qc4w-gemm_avx2-avxvnni-f16c-fma_standalone") {
+-  source_set("qd8-f16-qc8w-igemm_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
--      "-mavxvnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-igemm/gen/f16-igemm-1x16-minmax-neonfp16arith-ld64.c",
-+    "src/src/f16-igemm/gen/f16-igemm-1x8-minmax-neonfp16arith-ld64.c",
-+    "src/src/f16-igemm/gen/f16-igemm-6x16-minmax-neonfp16arith-ld64.c",
-+    "src/src/f16-igemm/gen/f16-igemm-6x8-minmax-neonfp16arith-ld64.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x8c8-minmax-avxvnni-prfm.c",
--      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-5x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c8-minmax-avx2.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-3x8c8-minmax-avx2.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vapproxgelu/gen/f16-vapproxgelu-scalar-rational-6-4-div.c"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -40888,6 +43034,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -40895,34 +43046,33 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
- 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
- 
--  source_set("qd8-f16-qc4w-gemm_f16c-fma-avx2") {
+-  }
+-
+-  source_set(
+-      "qd8-f16-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 -    cflags = [
--      "-mavx2",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("f16-igemm_arch=armv8.2-a+fp16+dotprod") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x8c8-minmax-avx2-madd-prfm.c",
--      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-4x8c8-minmax-avx2-madd-prfm.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c8-minmax-avx256skx.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-5x8c8-minmax-avx256skx.c",
 -    ]
-+  asmflags = cflags
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -40934,29 +43084,60 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-igemm/f16-igemm-1x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
-+    "src/src/f16-igemm/f16-igemm-4x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
-+    "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a55.S",
-+    "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a55r0.S",
-+    "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a75.S",
-+    "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-ld64.S"
++source_set("f16-vbinary_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vbinary/gen/f16-vadd-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vaddc-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vdiv-aarch64-neonfp16arith-u8.c",
++    "src/src/f16-vbinary/gen/f16-vdiv-fp16arith-u2.c",
++    "src/src/f16-vbinary/gen/f16-vdivc-aarch64-neonfp16arith-u8.c",
++    "src/src/f16-vbinary/gen/f16-vdivc-fp16arith-u2.c",
++    "src/src/f16-vbinary/gen/f16-vmax-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vmaxc-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vmin-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vminc-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vmul-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vmulc-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vprelu-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vpreluc-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vrdivc-aarch64-neonfp16arith-u8.c",
++    "src/src/f16-vbinary/gen/f16-vrdivc-fp16arith-u2.c",
++    "src/src/f16-vbinary/gen/f16-vrpreluc-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vrsubc-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vsqrdiff-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vsqrdiffc-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vsub-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vsubc-neonfp16arith-u16.c"
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f16-qc4w-gemm_f16c-fma-avx2_standalone") {
+-  source_set(
+-      "qd8-f16-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
--      "-mavx2",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
 -      "-mf16c",
 -      "-mfma",
 -    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c8-minmax-avx256skx.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-5x8c8-minmax-avx256skx.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -40964,14 +43145,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x8c8-minmax-avx2-madd-prfm.c",
--      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-4x8c8-minmax-avx2-madd-prfm.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -40982,45 +43155,63 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-igemm_arch=armv8.2-a+fp16+dotprod_standalone") {
++source_set("f16-vbinary_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
++    "-march=armv8.2-a+fp16"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
-+  asmflags = cflags
- 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
 -
 -  source_set(
--      "qd8-f16-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-      "qd8-f16-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
 -      "-mavx512dq",
 -      "-mavx512f",
 -      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-igemm/f16-igemm-1x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
-+    "src/src/f16-igemm/f16-igemm-4x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
-+    "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a55.S",
-+    "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a55r0.S",
-+    "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a75.S",
-+    "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-ld64.S"
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vbinary/gen/f16-vadd-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vaddc-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vdiv-aarch64-neonfp16arith-u8.c",
++    "src/src/f16-vbinary/gen/f16-vdiv-fp16arith-u2.c",
++    "src/src/f16-vbinary/gen/f16-vdivc-aarch64-neonfp16arith-u8.c",
++    "src/src/f16-vbinary/gen/f16-vdivc-fp16arith-u2.c",
++    "src/src/f16-vbinary/gen/f16-vmax-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vmaxc-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vmin-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vminc-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vmul-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vmulc-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vprelu-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vpreluc-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vrdivc-aarch64-neonfp16arith-u8.c",
++    "src/src/f16-vbinary/gen/f16-vrdivc-fp16arith-u2.c",
++    "src/src/f16-vbinary/gen/f16-vrpreluc-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vrsubc-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vsqrdiff-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vsqrdiffc-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vsub-neonfp16arith-u16.c",
++    "src/src/f16-vbinary/gen/f16-vsubc-neonfp16arith-u16.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x8c8-minmax-avx256skx-madd-prfm.c",
--      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-8x8c8-minmax-avx256skx-madd-prfm.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c8-minmax-avx256vnni.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-8x8c8-minmax-avx256vnni.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -41042,10 +43233,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
++
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
-+
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
    }
@@ -41053,29 +43244,30 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
--      "qd8-f16-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-      "qd8-f16-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
 -      "-mavx512dq",
 -      "-mavx512f",
 -      "-mavx512vl",
+-      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("f16-maxpool_arch=armv8.2-a+fp16") {
++source_set("f16-vclamp_arch=armv8.2-a+fp16") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x8c8-minmax-avx256skx-madd-prfm.c",
--      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-8x8c8-minmax-avx256skx-madd-prfm.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c8-minmax-avx256vnni.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-8x8c8-minmax-avx256vnni.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-maxpool/gen/f16-maxpool-9p-minmax-neonfp16arith-u8.c"
++    "src/src/f16-vclamp/gen/f16-vclamp-neonfp16arith-u16.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -41106,8 +43298,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 -
 -  source_set(
--      "qd8-f16-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
+-      "qd8-f16-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8") {
 -    cflags = [
+-      "-mamx-int8",
+-      "-mamx-tile",
 -      "-mavx512bw",
 -      "-mavx512cd",
 -      "-mavx512dq",
@@ -41116,17 +43310,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
+-      "-mgfni",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x8c8-minmax-avx256vnni.c",
--      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-8x8c8-minmax-avx256vnni.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-16x64c4-minmax-avx512amx.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x64c4-minmax-avx512amx.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-maxpool_arch=armv8.2-a+fp16_standalone") {
++source_set("f16-vclamp_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
@@ -41143,7 +43338,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-maxpool/gen/f16-maxpool-9p-minmax-neonfp16arith-u8.c"
++    "src/src/f16-vclamp/gen/f16-vclamp-neonfp16arith-u16.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -41154,8 +43349,10 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
--      "qd8-f16-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
+-      "qd8-f16-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8_standalone") {
 -    cflags = [
+-      "-mamx-int8",
+-      "-mamx-tile",
 -      "-mavx512bw",
 -      "-mavx512cd",
 -      "-mavx512dq",
@@ -41164,6 +43361,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
+-      "-mgfni",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -41174,8 +43372,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x8c8-minmax-avx256vnni.c",
--      "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-8x8c8-minmax-avx256vnni.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-16x64c4-minmax-avx512amx.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x64c4-minmax-avx512amx.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
  
@@ -41195,7 +43393,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("f16-qs8-vcvt_arch=armv8.2-a+fp16") {
++source_set("f16-vcmul_arch=armv8.2-a+fp16") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
@@ -41206,15 +43404,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-qs8-vcvt/gen/f16-qs8-vcvt-neonfp16arith-u32.c"
++    "src/src/f16-vcmul/gen/f16-vcmul-neonfp16arith-u16.c"
 +  ]
  
--  source_set("qd8-f16-qc8w-gemm_avx2-avxvnni-f16c-fma") {
+-  source_set("qd8-f32-qb4w-gemm_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
--      "-mavx2",
--      "-mavxvnni",
--      "-mf16c",
--      "-mfma",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -41222,8 +43420,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c8-minmax-avxvnni-prfm.c",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-5x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4c8-minmax-avx-ld128.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4c8-minmax-avx-ld128.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -41248,28 +43446,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("f16-qs8-vcvt_arch=armv8.2-a+fp16_standalone") {
++source_set("f16-vcmul_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f16-qc8w-gemm_avx2-avxvnni-f16c-fma_standalone") {
+-  source_set("qd8-f32-qb4w-gemm_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
--      "-mavx2",
--      "-mavxvnni",
--      "-mf16c",
--      "-mfma",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-qs8-vcvt/gen/f16-qs8-vcvt-neonfp16arith-u32.c"
++    "src/src/f16-vcmul/gen/f16-vcmul-neonfp16arith-u16.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c8-minmax-avxvnni-prfm.c",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-5x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4c8-minmax-avx-ld128.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4c8-minmax-avx-ld128.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -41303,20 +43501,21 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qd8-f16-qc8w-gemm_f16c-fma-avx2") {
+-  source_set("qd8-f32-qb4w-gemm_f16c-fma-avx2") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("f16-qs8-vcvt_arm64") {
-+  cflags = [
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c8-minmax-avx2.c",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-3x8c8-minmax-avx2.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x8c8-minmax-avx2.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-3x8c8-minmax-avx2.c",
 -    ]
++source_set("f16-vcos_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -41331,7 +43530,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-qs8-vcvt/gen/f16-qs8-vcvt-scalar-imagic-u4.c"
++    "src/src/f16-vcos/gen/f16-vcos-neonfp16arith-rational-3-2-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -41341,7 +43540,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f16-qc8w-gemm_f16c-fma-avx2_standalone") {
+-  source_set("qd8-f32-qb4w-gemm_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
@@ -41356,8 +43555,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c8-minmax-avx2.c",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-3x8c8-minmax-avx2.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x8c8-minmax-avx2.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-3x8c8-minmax-avx2.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -41373,10 +43572,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-qs8-vcvt_arm64_standalone") {
++source_set("f16-vcos_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vcos/gen/f16-vcos-neonfp16arith-rational-3-2-div.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -41385,29 +43589,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 -
 -  source_set(
--      "qd8-f16-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-      "qd8-f32-qb4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
 -      "-mavx512dq",
 -      "-mavx512f",
 -      "-mavx512vl",
+-      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-qs8-vcvt/gen/f16-qs8-vcvt-scalar-imagic-u4.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c8-minmax-avx256skx.c",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-5x8c8-minmax-avx256skx.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-8x16c8-minmax-avx512vnni-prfm.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -41419,41 +43626,36 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+  public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
--      "qd8-f16-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-      "qd8-f32-qb4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
 -      "-mavx512dq",
 -      "-mavx512f",
 -      "-mavx512vl",
+-      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++source_set("f16-vcos_arm64") {
++  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c8-minmax-avx256skx.c",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-5x8c8-minmax-avx256skx.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-8x16c8-minmax-avx512vnni-prfm.c",
 -    ]
-+source_set("f16-qu8-vcvt_arm64") {
-+  cflags = [
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -41465,21 +43667,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-qu8-vcvt/gen/f16-qu8-vcvt-scalar-imagic-u4.c"
++    "src/src/f16-vcos/gen/f16-vcos-scalar-rational-3-2-div.c"
 +  ]
  
+-    public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
 -
 -  source_set(
--      "qd8-f16-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
+-      "qd8-f32-qb4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
@@ -41489,15 +43693,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c8-minmax-avx256vnni.c",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-8x8c8-minmax-avx256vnni.c",
+-      "-mgfni",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -41506,6 +43702,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-14x16c8-minmax-avx512vnnigfni-prfm.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c8-minmax-avx512vnnigfni-prfm.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -41516,18 +43720,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++# This is a target that cannot depend on //base.
++source_set("f16-vcos_arm64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+# This is a target that cannot depend on //base.
-+source_set("f16-qu8-vcvt_arm64_standalone") {
-+  cflags = [
++  ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
--      "qd8-f16-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
+-      "qd8-f32-qb4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni_standalone") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
@@ -41537,17 +43740,104 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
+-      "-mgfni",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vcos/gen/f16-vcos-scalar-rational-3-2-div.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c8-minmax-avx256vnni.c",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-8x8c8-minmax-avx256vnni.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-14x16c8-minmax-avx512vnnigfni-prfm.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c8-minmax-avx512vnnigfni-prfm.c",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  source_set("qd8-f32-qb4w-gemm_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4c8-minmax-sse2-ld128.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4c8-minmax-sse2-ld128.c",
+-    ]
+-
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
 -    ]
++source_set("f16-velu_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-qu8-vcvt/gen/f16-qu8-vcvt-scalar-imagic-u4.c"
++    "src/src/f16-velu/gen/f16-velu-neonfp16arith-rr1-p3-u16.c"
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f32-qb4w-gemm_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4c8-minmax-sse2-ld128.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4c8-minmax-sse2-ld128.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -41560,11 +43850,39 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    public_configs = [ ":xnnpack_config" ]
++# This is a target that cannot depend on //base.
++source_set("f16-velu_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-velu/gen/f16-velu-neonfp16arith-rr1-p3-u16.c"
++  ]
+ 
+-  source_set("qd8-f32-qb4w-gemm_sse4.1-no-sse4.2") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4c8-minmax-sse41-ld128.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-3x4c8-minmax-sse41-ld128.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -41572,38 +43890,164 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f32-qb4w-gemm_sse4.1-no-sse4.2_standalone") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
+-    ]
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4c8-minmax-sse41-ld128.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-3x4c8-minmax-sse41-ld128.c",
+-    ]
+-
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
+-
+-    public_configs = [ ":xnnpack_config" ]
++source_set("f16-vexp_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vexp/gen/f16-vexp-neonfp16arith-poly-3.c"
++  ]
+ 
+-  source_set("qd8-f32-qb4w-gemm_x64") {
+-    cflags = []
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4-minmax-scalar.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4-minmax-scalar.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
 -
--  source_set(
--      "qd8-f16-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8") {
--    cflags = [
--      "-mamx-int8",
--      "-mamx-tile",
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
--      "-mf16c",
--      "-mfma",
--      "-mgfni",
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++# This is a target that cannot depend on //base.
++source_set("f16-vexp_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f32-qb4w-gemm_x64_standalone") {
+-    cflags = []
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vexp/gen/f16-vexp-neonfp16arith-poly-3.c"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-16x64c4-minmax-avx512amx.c",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x64c4-minmax-avx512amx.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4-minmax-scalar.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4-minmax-scalar.c",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
  
+-  source_set("qd8-f32-qc4w-gemm_avx2-avxvnni-f16c-fma") {
+-    cflags = [
+-      "-mavx2",
+-      "-mavxvnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++source_set("f16-vexp_arm64") {
++  cflags = [
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-5x8c8-minmax-avxvnni-prfm.c",
+-    ]
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -41614,42 +44058,24 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("f16-raddstoreexpminusmax_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vexp/gen/f16-vexp-scalar-poly-3.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-raddstoreexpminusmax/gen/f16-raddstoreexpminusmax-neonfp16arith-rr2-p2-u32.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qd8-f16-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8_standalone") {
+-  source_set("qd8-f32-qc4w-gemm_avx2-avxvnni-f16c-fma_standalone") {
 -    cflags = [
--      "-mamx-int8",
--      "-mamx-tile",
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
+-      "-mavx2",
+-      "-mavxvnni",
 -      "-mf16c",
 -      "-mfma",
--      "-mgfni",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-16x64c4-minmax-avx512amx.c",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x64c4-minmax-avx512amx.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -41658,6 +44084,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-5x8c8-minmax-avxvnni-prfm.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -41668,14 +44102,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-raddstoreexpminusmax_arch=armv8.2-a+fp16_standalone") {
++source_set("f16-vexp_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16"
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -41684,13 +44115,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-raddstoreexpminusmax/gen/f16-raddstoreexpminusmax-neonfp16arith-rr2-p2-u32.c"
++    "src/src/f16-vexp/gen/f16-vexp-scalar-poly-3.c"
 +  ]
  
--  source_set("qd8-f16-qc8w-igemm_avx2-avxvnni-f16c-fma") {
+-  source_set("qd8-f32-qc4w-gemm_f16c-fma-avx2") {
 -    cflags = [
 -      "-mavx2",
--      "-mavxvnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
@@ -41700,8 +44130,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c8-minmax-avxvnni-prfm.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-5x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x8c8-minmax-avx2-madd-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x8c8-minmax-avx2-madd-prfm.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -41729,22 +44159,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f16-qc8w-igemm_avx2-avxvnni-f16c-fma_standalone") {
+-  source_set("qd8-f32-qc4w-gemm_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
--      "-mavxvnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
--
++source_set("f16-vgelu_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c8-minmax-avxvnni-prfm.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-5x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x8c8-minmax-avx2-madd-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x8c8-minmax-avx2-madd-prfm.c",
 -    ]
-+source_set("f16-rdminmax_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vgelu/gen/f16-vgelu-neonfp16arith-rational-6-4-div.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -41757,21 +44190,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-rdminmax/gen/f16-rdmax-2p2x-neonfp16arith-c32.c",
-+    "src/src/f16-rdminmax/gen/f16-rdmin-2p2x-neonfp16arith-c32.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -41779,9 +44202,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  source_set("qd8-f16-qc8w-igemm_f16c-fma-avx2") {
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
+-
+-  source_set(
+-      "qd8-f32-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 -    cflags = [
--      "-mavx2",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
 -      "-mf16c",
 -      "-mfma",
 -    ]
@@ -41790,11 +44223,13 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c8-minmax-avx2.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-3x8c8-minmax-avx2.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-avx512skx-madd-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x8c8-minmax-avx256skx-madd-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-8x16c8-minmax-avx512skx-madd-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-8x8c8-minmax-avx256skx-madd-prfm.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-rdminmax_arch=armv8.2-a+fp16_standalone") {
++source_set("f16-vgelu_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
@@ -41811,8 +44246,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-rdminmax/gen/f16-rdmax-2p2x-neonfp16arith-c32.c",
-+    "src/src/f16-rdminmax/gen/f16-rdmin-2p2x-neonfp16arith-c32.c"
++    "src/src/f16-vgelu/gen/f16-vgelu-neonfp16arith-rational-6-4-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -41822,9 +44256,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f16-qc8w-igemm_f16c-fma-avx2_standalone") {
+-  source_set(
+-      "qd8-f32-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
--      "-mavx2",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
 -      "-mf16c",
 -      "-mfma",
 -    ]
@@ -41837,8 +44276,10 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c8-minmax-avx2.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-3x8c8-minmax-avx2.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-avx512skx-madd-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x8c8-minmax-avx256skx-madd-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-8x16c8-minmax-avx512skx-madd-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-8x8c8-minmax-avx256skx-madd-prfm.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
  
@@ -41858,7 +44299,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("f16-rdminmax_arm64") {
++source_set("f16-vgelu_arm64") {
 +  cflags = [
  
 -    if (!(is_android && use_order_profiling)) {
@@ -41867,13 +44308,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 -
 -  source_set(
--      "qd8-f16-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-      "qd8-f32-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
 -      "-mavx512dq",
 -      "-mavx512f",
 -      "-mavx512vl",
+-      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
@@ -41881,13 +44323,12 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c8-minmax-avx256skx.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-5x8c8-minmax-avx256skx.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-8x16c8-minmax-avx512vnni-prfm.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-rdminmax/gen/f16-rdmax-2p2x-scalar-c2.c",
-+    "src/src/f16-rdminmax/gen/f16-rdmin-2p2x-scalar-c2.c"
++    "src/src/f16-vgelu/gen/f16-vgelu-scalar-rational-6-4-div.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -41915,13 +44356,14 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
--      "qd8-f16-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-      "qd8-f32-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
 -      "-mavx512dq",
 -      "-mavx512f",
 -      "-mavx512vl",
+-      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
@@ -41930,11 +44372,11 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c8-minmax-avx256skx.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-5x8c8-minmax-avx256skx.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-8x16c8-minmax-avx512vnni-prfm.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-rdminmax_arm64_standalone") {
++source_set("f16-vgelu_arm64_standalone") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -41952,8 +44394,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-rdminmax/gen/f16-rdmax-2p2x-scalar-c2.c",
-+    "src/src/f16-rdminmax/gen/f16-rdmin-2p2x-scalar-c2.c"
++    "src/src/f16-vgelu/gen/f16-vgelu-scalar-rational-6-4-div.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -41962,7 +44403,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 -
 -  source_set(
--      "qd8-f16-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
+-      "qd8-f32-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
@@ -41972,6 +44413,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
+-      "-mgfni",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -41979,8 +44421,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c8-minmax-avx256vnni.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-8x8c8-minmax-avx256vnni.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-14x16c8-minmax-avx512vnnigfni-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-avx512vnnigfni-prfm.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -42009,7 +44451,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
--      "qd8-f16-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
+-      "qd8-f32-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni_standalone") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
@@ -42019,14 +44461,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
+-      "-mgfni",
 -    ]
--
++source_set("f16-vhswish_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c8-minmax-avx256vnni.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-8x8c8-minmax-avx256vnni.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-14x16c8-minmax-avx512vnnigfni-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-avx512vnnigfni-prfm.c",
 -    ]
--
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vhswish/gen/f16-vhswish-neonfp16arith-u16.c"
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -42037,17 +44488,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("f16-rminmax_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-rminmax/gen/f16-rmax-neonfp16arith-u32-acc4.c",
-+    "src/src/f16-rminmax/gen/f16-rmin-neonfp16arith-u32-acc4.c",
-+    "src/src/f16-rminmax/gen/f16-rminmax-neonfp16arith-u32-acc4.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -42056,7 +44506,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 -
 -  source_set(
--      "qd8-f16-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8") {
+-      "qd8-f32-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8") {
 -    cflags = [
 -      "-mamx-int8",
 -      "-mamx-tile",
@@ -42070,20 +44520,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mfma",
 -      "-mgfni",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-16x64c4-minmax-avx512amx.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x64c4-minmax-avx512amx.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-16x64c4-minmax-avx512amx.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x64c4-minmax-avx512amx.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++# This is a target that cannot depend on //base.
++source_set("f16-vhswish_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -42096,20 +44544,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vhswish/gen/f16-vhswish-neonfp16arith-u16.c"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+# This is a target that cannot depend on //base.
-+source_set("f16-rminmax_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
--      "qd8-f16-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8_standalone") {
+-      "qd8-f32-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8_standalone") {
 -    cflags = [
 -      "-mamx-int8",
 -      "-mamx-tile",
@@ -42123,21 +44571,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mfma",
 -      "-mgfni",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-rminmax/gen/f16-rmax-neonfp16arith-u32-acc4.c",
-+    "src/src/f16-rminmax/gen/f16-rmin-neonfp16arith-u32-acc4.c",
-+    "src/src/f16-rminmax/gen/f16-rminmax-neonfp16arith-u32-acc4.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-16x64c4-minmax-avx512amx.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x64c4-minmax-avx512amx.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-16x64c4-minmax-avx512amx.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x64c4-minmax-avx512amx.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -42149,39 +44595,45 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
++source_set("f16-vlrelu_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vlrelu/gen/f16-vlrelu-neonfp16arith-u16.c"
++  ]
  
--  source_set("qd8-f32-qb4w-gemm_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qd8-f32-qc4w-gemm_sse2-no-sse3") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+source_set("f16-rminmax_arm64") {
-+  cflags = [
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4c8-minmax-avx-ld128.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4c8-minmax-avx-ld128.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4c8-minmax-sse2-ld128.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4c8-minmax-sse2-ld128.c",
 -    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -42194,41 +44646,36 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-rminmax/gen/f16-rmax-scalar-u2-acc2.c",
-+    "src/src/f16-rminmax/gen/f16-rmin-scalar-u2-acc2.c",
-+    "src/src/f16-rminmax/gen/f16-rminmax-scalar-u2-acc2.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++# This is a target that cannot depend on //base.
++source_set("f16-vlrelu_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qb4w-gemm_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("qd8-f32-qc4w-gemm_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vlrelu/gen/f16-vlrelu-neonfp16arith-u16.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4c8-minmax-avx-ld128.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4c8-minmax-avx-ld128.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4c8-minmax-sse2-ld128.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4c8-minmax-sse2-ld128.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -42240,39 +44687,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-rminmax_arm64_standalone") {
-+  cflags = [
- 
--    public_configs = [ ":xnnpack_config" ]
-+  ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-rminmax/gen/f16-rmax-scalar-u2-acc2.c",
-+    "src/src/f16-rminmax/gen/f16-rmin-scalar-u2-acc2.c",
-+    "src/src/f16-rminmax/gen/f16-rminmax-scalar-u2-acc2.c"
-+  ]
- 
--  source_set("qd8-f32-qb4w-gemm_f16c-fma-avx2") {
--    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x8c8-minmax-avx2.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-3x8c8-minmax-avx2.c",
--    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -42280,44 +44694,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
  
--    public_configs = [ ":xnnpack_config" ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
    }
 +}
  
--  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qb4w-gemm_f16c-fma-avx2_standalone") {
+-  source_set("qd8-f32-qc4w-gemm_ssse3-no-sse4.1") {
 -    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse4.1",
+-      "-mssse3",
 -    ]
-+source_set("f16-spmm_arch=armv8.2-a+fp16") {
++source_set("f16-vmulcaddc_arch=armv8.2-a+fp16") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x8c8-minmax-avx2.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-3x8c8-minmax-avx2.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4c8-minmax-ssse3-madd.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4c8-minmax-ssse3-madd.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-spmm/gen/f16-spmm-32x1-minmax-neonfp16arith-pipelined.c"
++    "src/src/f16-vmulcaddc/gen/f16-vmulcaddc-c8-minmax-neonfp16arith-2x.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -42328,13 +44733,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    public_configs = [ ":xnnpack_config" ]
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -42342,33 +44748,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
--
--  source_set(
--      "qd8-f32-qb4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f32-qc4w-gemm_ssse3-no-sse4.1_standalone") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse4.1",
+-      "-mssse3",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c8-minmax-avx512vnni-prfm.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-8x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4c8-minmax-ssse3-madd.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4c8-minmax-ssse3-madd.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-spmm_arch=armv8.2-a+fp16_standalone") {
++source_set("f16-vmulcaddc_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
@@ -42381,32 +44776,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-spmm/gen/f16-spmm-32x1-minmax-neonfp16arith-pipelined.c"
++    "src/src/f16-vmulcaddc/gen/f16-vmulcaddc-c8-minmax-neonfp16arith-2x.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set(
--      "qd8-f32-qb4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
--      "-mf16c",
--      "-mfma",
--    ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -42414,12 +44799,63 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-  source_set("qd8-f32-qc4w-gemm_x64") {
+-    cflags = []
++  public_configs = [ ":xnnpack_config" ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c8-minmax-avx512vnni-prfm.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-8x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4-minmax-scalar.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4-minmax-scalar.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++source_set("f16-vrnd_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vrnd/gen/f16-vrndd-neonfp16arith-u16.c",
++    "src/src/f16-vrnd/gen/f16-vrndne-neonfp16arith-u16.c",
++    "src/src/f16-vrnd/gen/f16-vrndu-neonfp16arith-u16.c",
++    "src/src/f16-vrnd/gen/f16-vrndz-neonfp16arith-u16.c"
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f32-qc4w-gemm_x64_standalone") {
+-    cflags = []
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4-minmax-scalar.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4-minmax-scalar.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -42431,13 +44867,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
++  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("f16-vapproxgelu_arch=armv8.2-a+fp16") {
++# This is a target that cannot depend on //base.
++source_set("f16-vrnd_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
@@ -42446,34 +44881,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set(
--      "qd8-f32-qb4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
--      "-mf16c",
--      "-mfma",
--      "-mgfni",
--    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vapproxgelu/gen/f16-vapproxgelu-neonfp16arith-rational-6-4-div.c"
++    "src/src/f16-vrnd/gen/f16-vrndd-neonfp16arith-u16.c",
++    "src/src/f16-vrnd/gen/f16-vrndne-neonfp16arith-u16.c",
++    "src/src/f16-vrnd/gen/f16-vrndu-neonfp16arith-u16.c",
++    "src/src/f16-vrnd/gen/f16-vrndz-neonfp16arith-u16.c"
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-14x16c8-minmax-avx512vnnigfni-prfm.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c8-minmax-avx512vnnigfni-prfm.c",
+-  source_set("qd8-f32-qc8w-gemm_avx2-avxvnni-f16c-fma") {
+-    cflags = [
+-      "-mavx2",
+-      "-mavxvnni",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-5x8c8-minmax-avxvnni-prfm.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -42484,46 +44922,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+  public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qd8-f32-qb4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni_standalone") {
+-  source_set("qd8-f32-qc8w-gemm_avx2-avxvnni-f16c-fma_standalone") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
+-      "-mavx2",
+-      "-mavxvnni",
 -      "-mf16c",
 -      "-mfma",
--      "-mgfni",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-vapproxgelu_arch=armv8.2-a+fp16_standalone") {
++source_set("f16-vrsqrt_arch=armv8.2-a+fp16") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-14x16c8-minmax-avx512vnnigfni-prfm.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c8-minmax-avx512vnnigfni-prfm.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-5x8c8-minmax-avxvnni-prfm.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vapproxgelu/gen/f16-vapproxgelu-neonfp16arith-rational-6-4-div.c"
++    "src/src/f16-vrsqrt/gen/f16-vrsqrt-neonfp16arith-rsqrt-u16.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -42545,32 +44972,36 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
+-  }
++  public_configs = [ ":xnnpack_config" ]
 +}
  
--  source_set("qd8-f32-qb4w-gemm_sse2-no-sse3") {
+-  source_set("qd8-f32-qc8w-gemm_f16c-fma-avx2") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+source_set("f16-vapproxgelu_arm64") {
++# This is a target that cannot depend on //base.
++source_set("f16-vrsqrt_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4c8-minmax-sse2-ld128.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4c8-minmax-sse2-ld128.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c8-minmax-avx2.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x8c8-minmax-avx2.c",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vrsqrt/gen/f16-vrsqrt-neonfp16arith-rsqrt-u16.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -42583,36 +45014,36 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vapproxgelu/gen/f16-vapproxgelu-scalar-rational-6-4-div.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qb4w-gemm_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f32-qc8w-gemm_f16c-fma-avx2_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4c8-minmax-sse2-ld128.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4c8-minmax-sse2-ld128.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c8-minmax-avx2.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x8c8-minmax-avx2.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
 +}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -42625,26 +45056,33 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-vapproxgelu_arm64_standalone") {
++source_set("f16-vsigmoid_arch=armv8.2-a+fp16") {
 +  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vsigmoid/gen/f16-vsigmoid-neonfp16arith-rr2-p2-nr1fma-u32.c",
++    "src/src/f16-vsigmoid/gen/f16-vsigmoid-neonfp16arith-rr2-p2-nr1recps-u16.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vapproxgelu/gen/f16-vapproxgelu-scalar-rational-6-4-div.c"
-+  ]
- 
--  source_set("qd8-f32-qb4w-gemm_sse4.1-no-sse4.2") {
+-
+-  source_set(
+-      "qd8-f32-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -42652,14 +45090,16 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4c8-minmax-sse41-ld128.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-3x4c8-minmax-sse41-ld128.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c8-minmax-avx512skx-prfm.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c8-minmax-avx256skx.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-8x16c8-minmax-avx512skx-prfm.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-8x8c8-minmax-avx256skx.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -42673,54 +45113,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++# This is a target that cannot depend on //base.
++source_set("f16-vsigmoid_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qb4w-gemm_sse4.1-no-sse4.2_standalone") {
+-  source_set(
+-      "qd8-f32-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+source_set("f16-vbinary_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vsigmoid/gen/f16-vsigmoid-neonfp16arith-rr2-p2-nr1fma-u32.c",
++    "src/src/f16-vsigmoid/gen/f16-vsigmoid-neonfp16arith-rr2-p2-nr1recps-u16.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4c8-minmax-sse41-ld128.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-3x4c8-minmax-sse41-ld128.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c8-minmax-avx512skx-prfm.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c8-minmax-avx256skx.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-8x16c8-minmax-avx512skx-prfm.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-8x8c8-minmax-avx256skx.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vbinary/gen/f16-vadd-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vaddc-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vdiv-aarch64-neonfp16arith-u8.c",
-+    "src/src/f16-vbinary/gen/f16-vdiv-fp16arith-u2.c",
-+    "src/src/f16-vbinary/gen/f16-vdivc-aarch64-neonfp16arith-u8.c",
-+    "src/src/f16-vbinary/gen/f16-vdivc-fp16arith-u2.c",
-+    "src/src/f16-vbinary/gen/f16-vmax-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vmaxc-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vmin-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vminc-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vmul-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vmulc-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vprelu-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vpreluc-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vrdivc-aarch64-neonfp16arith-u8.c",
-+    "src/src/f16-vbinary/gen/f16-vrdivc-fp16arith-u2.c",
-+    "src/src/f16-vbinary/gen/f16-vrpreluc-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vrsubc-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vsqrdiff-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vsqrdiffc-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vsub-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vsubc-neonfp16arith-u16.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -42732,67 +45162,47 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  public_configs = [ ":xnnpack_config" ]
+-
+-  source_set(
+-      "qd8-f32-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
 +}
-+
-+# This is a target that cannot depend on //base.
-+source_set("f16-vbinary_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
- 
--  source_set("qd8-f32-qb4w-gemm_x64") {
--    cflags = []
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vbinary/gen/f16-vadd-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vaddc-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vdiv-aarch64-neonfp16arith-u8.c",
-+    "src/src/f16-vbinary/gen/f16-vdiv-fp16arith-u2.c",
-+    "src/src/f16-vbinary/gen/f16-vdivc-aarch64-neonfp16arith-u8.c",
-+    "src/src/f16-vbinary/gen/f16-vdivc-fp16arith-u2.c",
-+    "src/src/f16-vbinary/gen/f16-vmax-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vmaxc-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vmin-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vminc-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vmul-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vmulc-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vprelu-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vpreluc-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vrdivc-aarch64-neonfp16arith-u8.c",
-+    "src/src/f16-vbinary/gen/f16-vrdivc-fp16arith-u2.c",
-+    "src/src/f16-vbinary/gen/f16-vrpreluc-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vrsubc-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vsqrdiff-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vsqrdiffc-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vsub-neonfp16arith-u16.c",
-+    "src/src/f16-vbinary/gen/f16-vsubc-neonfp16arith-u16.c"
-+  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4-minmax-scalar.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4-minmax-scalar.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-10x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c8-minmax-avx512vnni-prfm.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++source_set("f16-vsin_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -42804,42 +45214,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vsin/gen/f16-vsin-neonfp16arith-rational-3-2-div.c"
 +  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
-+
-+source_set("f16-vclamp_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qb4w-gemm_x64_standalone") {
--    cflags = []
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vclamp/gen/f16-vclamp-neonfp16arith-u16.c"
+-  source_set(
+-      "qd8-f32-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4-minmax-scalar.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4-minmax-scalar.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-10x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c8-minmax-avx512vnni-prfm.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -42851,47 +45263,53 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++# This is a target that cannot depend on //base.
++source_set("f16-vsin_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vsin/gen/f16-vsin-neonfp16arith-rational-3-2-div.c"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+# This is a target that cannot depend on //base.
-+source_set("f16-vclamp_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
- 
--  source_set("qd8-f32-qc4w-gemm_avx2-avxvnni-f16c-fma") {
+-
+-  source_set(
+-      "qd8-f32-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8") {
 -    cflags = [
--      "-mavx2",
--      "-mavxvnni",
+-      "-mamx-int8",
+-      "-mamx-tile",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
+-      "-mgfni",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vclamp/gen/f16-vclamp-neonfp16arith-u16.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x8c8-minmax-avxvnni-prfm.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-5x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-16x64c4-minmax-avx512amx.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x64c4-minmax-avx512amx.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -42903,13 +45321,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
 +  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -42919,27 +45330,29 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc4w-gemm_avx2-avxvnni-f16c-fma_standalone") {
+-  source_set(
+-      "qd8-f32-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8_standalone") {
 -    cflags = [
--      "-mavx2",
--      "-mavxvnni",
+-      "-mamx-int8",
+-      "-mamx-tile",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
+-      "-mgfni",
 -    ]
-+source_set("f16-vcmul_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x8c8-minmax-avxvnni-prfm.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-5x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-16x64c4-minmax-avx512amx.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x64c4-minmax-avx512amx.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vcmul/gen/f16-vcmul-neonfp16arith-u16.c"
-+  ]
++source_set("f16-vsin_arm64") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -42951,11 +45364,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vsin/gen/f16-vsin-scalar-rational-3-2-div.c"
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
+-  source_set("qd8-f32-qc8w-gemm_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -42963,34 +45392,51 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4c8-minmax-sse2-ld64.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4c8-minmax-sse2-ld64.c",
+-    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
--  source_set("qd8-f32-qc4w-gemm_f16c-fma-avx2") {
--    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-vcmul_arch=armv8.2-a+fp16_standalone") {
++source_set("f16-vsin_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16"
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f32-qc8w-gemm_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vsin/gen/f16-vsin-scalar-rational-3-2-div.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x8c8-minmax-avx2-madd-prfm.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x8c8-minmax-avx2-madd-prfm.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4c8-minmax-sse2-ld64.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4c8-minmax-sse2-ld64.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vcmul/gen/f16-vcmul-neonfp16arith-u16.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -43000,47 +45446,45 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
    }
 +}
  
--  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc4w-gemm_f16c-fma-avx2_standalone") {
+-  source_set("qd8-f32-qc8w-gemm_sse4.1-no-sse4.2") {
 -    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
-+source_set("f16-vcos_arch=armv8.2-a+fp16") {
++source_set("f16-vsqrt_arch=armv8.2-a+fp16") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x8c8-minmax-avx2-madd-prfm.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x8c8-minmax-avx2-madd-prfm.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4c8-minmax-sse41-ld64.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4c8-minmax-sse41-ld64.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vcos/gen/f16-vcos-neonfp16arith-rational-3-2-div.c"
++    "src/src/f16-vsqrt/gen/f16-vsqrt-aarch64-neonfp16arith-sqrt-u8.c",
++    "src/src/f16-vsqrt/gen/f16-vsqrt-neonfp16arith-nr1fma1adj-u8.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -43051,13 +45495,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    public_configs = [ ":xnnpack_config" ]
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -43065,34 +45510,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
--
--  source_set(
--      "qd8-f32-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f32-qc8w-gemm_sse4.1-no-sse4.2_standalone") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-avx512skx-madd-prfm.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x8c8-minmax-avx256skx-madd-prfm.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-8x16c8-minmax-avx512skx-madd-prfm.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-8x8c8-minmax-avx256skx-madd-prfm.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4c8-minmax-sse41-ld64.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4c8-minmax-sse41-ld64.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-vcos_arch=armv8.2-a+fp16_standalone") {
++source_set("f16-vsqrt_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
@@ -43105,31 +45538,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vcos/gen/f16-vcos-neonfp16arith-rational-3-2-div.c"
++    "src/src/f16-vsqrt/gen/f16-vsqrt-aarch64-neonfp16arith-sqrt-u8.c",
++    "src/src/f16-vsqrt/gen/f16-vsqrt-neonfp16arith-nr1fma1adj-u8.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set(
--      "qd8-f32-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
--    ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -43137,63 +45562,21 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-avx512skx-madd-prfm.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x8c8-minmax-avx256skx-madd-prfm.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-8x16c8-minmax-avx512skx-madd-prfm.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-8x8c8-minmax-avx256skx-madd-prfm.c",
--    ]
+-  source_set("qd8-f32-qc8w-gemm_x64") {
+-    cflags = []
 +  public_configs = [ ":xnnpack_config" ]
  
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x2-minmax-scalar.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4-minmax-scalar.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4-minmax-scalar.c",
 -    ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
 +  }
 +}
  
--    public_configs = [ ":xnnpack_config" ]
-+source_set("f16-vcos_arm64") {
-+  cflags = [
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
--
--  source_set(
--      "qd8-f32-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
--      "-mf16c",
--      "-mfma",
--    ]
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-avx512vnni-prfm.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-8x16c8-minmax-avx512vnni-prfm.c",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vcos/gen/f16-vcos-scalar-rational-3-2-div.c"
-+  ]
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -43204,43 +45587,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++source_set("f16-vtanh_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vtanh/gen/f16-vtanh-aarch64-neonfp16arith-expm1minus-rr1-p3h2ts-div-u32.c",
++    "src/src/f16-vtanh/gen/f16-vtanh-neonfp16arith-expm1minus-rr1-p3h2ts-nr1fma-u32.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qd8-f32-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
--      "-mf16c",
--      "-mfma",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
+-  source_set("qd8-f32-qc8w-gemm_x64_standalone") {
+-    cflags = []
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-avx512vnni-prfm.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-8x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x2-minmax-scalar.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4-minmax-scalar.c",
+-      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4-minmax-scalar.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-vcos_arm64_standalone") {
-+  cflags = [
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -43252,47 +45624,48 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vcos/gen/f16-vcos-scalar-rational-3-2-div.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set(
--      "qd8-f32-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni") {
++# This is a target that cannot depend on //base.
++source_set("f16-vtanh_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
+ 
+-  source_set("qd8-f32-qc8w-igemm_avx2-avxvnni-f16c-fma") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
+-      "-mavx2",
+-      "-mavxvnni",
 -      "-mf16c",
 -      "-mfma",
--      "-mgfni",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vtanh/gen/f16-vtanh-aarch64-neonfp16arith-expm1minus-rr1-p3h2ts-div-u32.c",
++    "src/src/f16-vtanh/gen/f16-vtanh-neonfp16arith-expm1minus-rr1-p3h2ts-nr1fma-u32.c"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-14x16c8-minmax-avx512vnnigfni-prfm.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-avx512vnnigfni-prfm.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-5x8c8-minmax-avxvnni-prfm.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -43304,41 +45677,38 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  public_configs = [ ":xnnpack_config" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qd8-f32-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni_standalone") {
+-  source_set("qd8-f32-qc8w-igemm_avx2-avxvnni-f16c-fma_standalone") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
+-      "-mavx2",
+-      "-mavxvnni",
 -      "-mf16c",
 -      "-mfma",
--      "-mgfni",
 -    ]
-+source_set("f16-velu_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-14x16c8-minmax-avx512vnnigfni-prfm.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-avx512vnnigfni-prfm.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-5x8c8-minmax-avxvnni-prfm.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-velu/gen/f16-velu-neonfp16arith-rr1-p3-u16.c"
++source_set("f16-vunary_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -43351,11 +45721,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f16-vunary/gen/f16-vabs-neonfp16arith-u16.c",
++    "src/src/f16-vunary/gen/f16-vneg-neonfp16arith-u16.c",
++    "src/src/f16-vunary/gen/f16-vsqr-neonfp16arith-u16.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -43363,36 +45744,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
--
--  source_set(
--      "qd8-f32-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8") {
+-  source_set("qd8-f32-qc8w-igemm_f16c-fma-avx2") {
 -    cflags = [
--      "-mamx-int8",
--      "-mamx-tile",
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
+-      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
--      "-mgfni",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-16x64c4-minmax-avx512amx.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x64c4-minmax-avx512amx.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c8-minmax-avx2.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x8c8-minmax-avx2.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-velu_arch=armv8.2-a+fp16_standalone") {
++source_set("f16-vunary_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
@@ -43409,7 +45776,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-velu/gen/f16-velu-neonfp16arith-rr1-p3-u16.c"
++    "src/src/f16-vunary/gen/f16-vabs-neonfp16arith-u16.c",
++    "src/src/f16-vunary/gen/f16-vneg-neonfp16arith-u16.c",
++    "src/src/f16-vunary/gen/f16-vsqr-neonfp16arith-u16.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -43419,20 +45788,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qd8-f32-qc4w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8_standalone") {
+-  source_set("qd8-f32-qc8w-igemm_f16c-fma-avx2_standalone") {
 -    cflags = [
--      "-mamx-int8",
--      "-mamx-tile",
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
+-      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
--      "-mgfni",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -43443,8 +45803,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-16x64c4-minmax-avx512amx.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x64c4-minmax-avx512amx.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c8-minmax-avx2.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x8c8-minmax-avx2.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
  
@@ -43464,39 +45824,38 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("f16-vexp_arch=armv8.2-a+fp16") {
++source_set("f32-argmaxpool_arm64") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vexp/gen/f16-vexp-neonfp16arith-poly-3.c"
-+  ]
- 
--  source_set("qd8-f32-qc4w-gemm_sse2-no-sse3") {
+-
+-  source_set(
+-      "qd8-f32-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4c8-minmax-sse2-ld128.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4c8-minmax-sse2-ld128.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c8-minmax-avx512skx-prfm.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c8-minmax-avx256skx.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-8x16c8-minmax-avx512skx-prfm.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-8x8c8-minmax-avx256skx.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-neon-c4.c",
++    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -43509,36 +45868,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+# This is a target that cannot depend on //base.
-+source_set("f16-vexp_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc4w-gemm_sse2-no-sse3_standalone") {
+-  source_set(
+-      "qd8-f32-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vexp/gen/f16-vexp-neonfp16arith-poly-3.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4c8-minmax-sse2-ld128.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4c8-minmax-sse2-ld128.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c8-minmax-avx512skx-prfm.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c8-minmax-avx256skx.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-8x16c8-minmax-avx512skx-prfm.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-8x8c8-minmax-avx256skx.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++# This is a target that cannot depend on //base.
++source_set("f32-argmaxpool_arm64_standalone") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -43550,37 +45917,46 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-neon-c4.c",
++    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
- 
--  source_set("qd8-f32-qc4w-gemm_ssse3-no-sse4.1") {
+-  }
+-
+-  source_set(
+-      "qd8-f32-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
 -    cflags = [
--      "-mno-sse4.1",
--      "-mssse3",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+source_set("f16-vexp_arm64") {
-+  cflags = [
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4c8-minmax-ssse3-madd.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4c8-minmax-ssse3-madd.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-10x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c8-minmax-avx512vnni-prfm.c",
 -    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -43593,37 +45969,36 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vexp/gen/f16-vexp-scalar-poly-3.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc4w-gemm_ssse3-no-sse4.1_standalone") {
+-  source_set(
+-      "qd8-f32-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
 -    cflags = [
--      "-mno-sse4.1",
--      "-mssse3",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++source_set("f32-avgpool_arm64") {
++  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4c8-minmax-ssse3-madd.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4c8-minmax-ssse3-madd.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-10x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c8-minmax-avx512vnni-prfm.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -43635,40 +46010,52 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-vexp_arm64_standalone") {
-+  cflags = [
- 
--    public_configs = [ ":xnnpack_config" ]
-+  ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vexp/gen/f16-vexp-scalar-poly-3.c"
++    "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-neon-u4.c",
++    "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-scalar-u1.c"
 +  ]
  
--  source_set("qd8-f32-qc4w-gemm_x64") {
--    cflags = []
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4-minmax-scalar.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4-minmax-scalar.c",
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
+-
+-  source_set(
+-      "qd8-f32-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8") {
+-    cflags = [
+-      "-mamx-int8",
+-      "-mamx-tile",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
+-      "-mgfni",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-16x64c4-minmax-avx512amx.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x64c4-minmax-avx512amx.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -43679,31 +46066,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++# This is a target that cannot depend on //base.
++source_set("f32-avgpool_arm64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc4w-gemm_x64_standalone") {
--    cflags = []
-+source_set("f16-vgelu_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
+-  source_set(
+-      "qd8-f32-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8_standalone") {
+-    cflags = [
+-      "-mamx-int8",
+-      "-mamx-tile",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
+-      "-mgfni",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-neon-u4.c",
++    "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-scalar-u1.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4-minmax-scalar.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4-minmax-scalar.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-16x64c4-minmax-avx512amx.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x64c4-minmax-avx512amx.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vgelu/gen/f16-vgelu-neonfp16arith-rational-6-4-div.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -43715,46 +46115,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
-+  public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
 +}
  
--  source_set("qd8-f32-qc8w-gemm_avx2-avxvnni-f16c-fma") {
+-  source_set("qd8-f32-qc8w-igemm_sse2-no-sse3") {
 -    cflags = [
--      "-mavx2",
--      "-mavxvnni",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-vgelu_arch=armv8.2-a+fp16_standalone") {
++source_set("f32-conv-hwc2chw_arm64") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c8-minmax-avxvnni-prfm.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-5x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4c8-minmax-sse2-ld64.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4c8-minmax-sse2-ld64.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vgelu/gen/f16-vgelu-neonfp16arith-rational-6-4-div.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -43767,37 +46158,38 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-aarch64-neonfma-2x2.c",
++    "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-neon-2x2.c",
++    "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-scalar-1x1.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f32-qc8w-igemm_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc8w-gemm_avx2-avxvnni-f16c-fma_standalone") {
--    cflags = [
--      "-mavx2",
--      "-mavxvnni",
--      "-mf16c",
--      "-mfma",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c8-minmax-avxvnni-prfm.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-5x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4c8-minmax-sse2-ld64.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4c8-minmax-sse2-ld64.c",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
++  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -43810,7 +46202,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("f16-vgelu_arm64") {
++# This is a target that cannot depend on //base.
++source_set("f32-conv-hwc2chw_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -43822,14 +46215,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vgelu/gen/f16-vgelu-scalar-rational-6-4-div.c"
++    "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-aarch64-neonfma-2x2.c",
++    "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-neon-2x2.c",
++    "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-scalar-1x1.c"
 +  ]
  
--  source_set("qd8-f32-qc8w-gemm_f16c-fma-avx2") {
+-  source_set("qd8-f32-qc8w-igemm_sse4.1-no-sse4.2") {
 -    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -43837,14 +46231,14 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c8-minmax-avx2.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x8c8-minmax-avx2.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4c8-minmax-sse41-ld64.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4c8-minmax-sse41-ld64.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -43858,31 +46252,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
-+}
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+# This is a target that cannot depend on //base.
-+source_set("f16-vgelu_arm64_standalone") {
-+  cflags = [
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc8w-gemm_f16c-fma-avx2_standalone") {
+-  source_set("qd8-f32-qc8w-igemm_sse4.1-no-sse4.2_standalone") {
 -    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
-+  ]
++source_set("f32-dwconv2d-chw_arm64") {
++  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c8-minmax-avx2.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x8c8-minmax-avx2.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4c8-minmax-sse41-ld64.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4c8-minmax-sse41-ld64.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vgelu/gen/f16-vgelu-scalar-rational-6-4-div.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -43895,47 +46285,56 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-aarch64-neonfma-3x4.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-neon-2x4.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-2x1-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-4x1.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-aarch64-neonfma-2x4-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-neon-1x4.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-1x1-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-2x1-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-aarch64-neonfma-4x4.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-neon-1x4.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-1x1-acc5.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-2x1-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-aarch64-neonfma-1x4-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-neon-1x4.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-1x1-acc5.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-2x1-acc2.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
--
--  source_set(
--      "qd8-f32-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
--    ]
+-  source_set("qd8-f32-qc8w-igemm_x64") {
+-    cflags = []
 +  public_configs = [ ":xnnpack_config" ]
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c8-minmax-avx512skx-prfm.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c8-minmax-avx256skx.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-8x16c8-minmax-avx512skx-prfm.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-8x8c8-minmax-avx256skx.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x2-minmax-scalar.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4-minmax-scalar.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4-minmax-scalar.c",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++# This is a target that cannot depend on //base.
++source_set("f32-dwconv2d-chw_arm64_standalone") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -43947,48 +46346,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("f16-vhswish_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vhswish/gen/f16-vhswish-neonfp16arith-u16.c"
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-aarch64-neonfma-3x4.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-neon-2x4.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-2x1-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-4x1.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-aarch64-neonfma-2x4-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-neon-1x4.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-1x1-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-2x1-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-aarch64-neonfma-4x4.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-neon-1x4.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-1x1-acc5.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-2x1-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-aarch64-neonfma-1x4-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-neon-1x4.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-1x1-acc5.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-2x1-acc2.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qd8-f32-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
+-  source_set("qd8-f32-qc8w-igemm_x64_standalone") {
+-    cflags = []
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x2-minmax-scalar.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4-minmax-scalar.c",
+-      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4-minmax-scalar.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c8-minmax-avx512skx-prfm.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c8-minmax-avx256skx.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-8x16c8-minmax-avx512skx-prfm.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-8x8c8-minmax-avx256skx.c",
--    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -43999,46 +46394,42 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-vhswish_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
--
--  source_set(
--      "qd8-f32-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  source_set("qs8-dwconv_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
--      "-mf16c",
--      "-mfma",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vhswish/gen/f16-vhswish-neonfp16arith-u16.c"
++source_set("f32-dwconv_arch=armv8.2-a+fp16+dotprod") {
++  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-10x16c8-minmax-avx512vnni-prfm.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p16c-minmax-fp32-avx-mul16-add16.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p16c-minmax-fp32-avx-mul16-add16.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  asmflags = cflags
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -44050,48 +46441,39 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-dwconv/f32-dwconv-9p4c-minmax-asm-aarch64-neonfma-cortex-a55.S"
 +  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qd8-f32-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
+-  source_set("qs8-dwconv_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
--      "-mf16c",
--      "-mfma",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
-+source_set("f16-vlrelu_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-10x16c8-minmax-avx512vnni-prfm.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p16c-minmax-fp32-avx-mul16-add16.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p16c-minmax-fp32-avx-mul16-add16.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vlrelu/gen/f16-vlrelu-neonfp16arith-u16.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -44103,50 +46485,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++# This is a target that cannot depend on //base.
++source_set("f32-dwconv_arch=armv8.2-a+fp16+dotprod_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++  asmflags = cflags
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set(
--      "qd8-f32-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8") {
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-dwconv/f32-dwconv-9p4c-minmax-asm-aarch64-neonfma-cortex-a55.S"
++  ]
+ 
+-  source_set("qs8-dwconv_f16c-fma-avx2") {
 -    cflags = [
--      "-mamx-int8",
--      "-mamx-tile",
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
+-      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
--      "-mgfni",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-16x64c4-minmax-avx512amx.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x64c4-minmax-avx512amx.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p16c-minmax-fp32-avx2-mul32.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p16c-minmax-fp32-avx2-mul32.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-vlrelu_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -44159,46 +46535,30 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vlrelu/gen/f16-vlrelu-neonfp16arith-u16.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--  # This is a target that cannot depend on //base.
--  source_set(
--      "qd8-f32-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8_standalone") {
--    cflags = [
--      "-mamx-int8",
--      "-mamx-tile",
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-dwconv_f16c-fma-avx2_standalone") {
+-    cflags = [
+-      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
--      "-mgfni",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++source_set("f32-dwconv_arm64") {
++  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-16x64c4-minmax-avx512amx.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x64c4-minmax-avx512amx.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p16c-minmax-fp32-avx2-mul32.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p16c-minmax-fp32-avx2-mul32.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -44210,39 +46570,48 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
-+source_set("f16-vmulcaddc_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vmulcaddc/gen/f16-vmulcaddc-c8-minmax-neonfp16arith-2x.c"
++    "src/src/f32-dwconv/gen/f32-dwconv-25p1c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-25p1c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-25p2c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-25p2c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-25p8c-minmax-neon.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-25p8c-minmax-neonfma-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-3p1c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-3p1c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-3p8c-minmax-neon.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-3p8c-minmax-neonfma.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-4p1c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-4p1c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-4p8c-minmax-neon.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-4p8c-minmax-neonfma.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-9p1c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-9p1c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-9p8c-minmax-neon.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-9p8c-minmax-neonfma.c"
 +  ]
  
--  source_set("qd8-f32-qc8w-gemm_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4c8-minmax-sse2-ld64.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4c8-minmax-sse2-ld64.c",
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
+-
+-  source_set(
+-      "qs8-dwconv_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -44251,6 +46620,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p32c-minmax-fp32-avx512skx-mul32.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p32c-minmax-fp32-avx512skx-mul32.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -44261,32 +46638,52 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++# This is a target that cannot depend on //base.
++source_set("f32-dwconv_arm64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+# This is a target that cannot depend on //base.
-+source_set("f16-vmulcaddc_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc8w-gemm_sse2-no-sse3_standalone") {
+-  source_set(
+-      "qs8-dwconv_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vmulcaddc/gen/f16-vmulcaddc-c8-minmax-neonfp16arith-2x.c"
++    "src/src/f32-dwconv/gen/f32-dwconv-25p1c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-25p1c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-25p2c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-25p2c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-25p8c-minmax-neon.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-25p8c-minmax-neonfma-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-3p1c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-3p1c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-3p8c-minmax-neon.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-3p8c-minmax-neonfma.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-4p1c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-4p1c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-4p8c-minmax-neon.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-4p8c-minmax-neonfma.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-9p1c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-9p1c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-9p8c-minmax-neon.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-9p8c-minmax-neonfma.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4c8-minmax-sse2-ld64.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4c8-minmax-sse2-ld64.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p32c-minmax-fp32-avx512skx-mul32.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p32c-minmax-fp32-avx512skx-mul32.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -44320,27 +46717,19 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qd8-f32-qc8w-gemm_sse4.1-no-sse4.2") {
+-  source_set("qs8-dwconv_sse2-no-sse3") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+source_set("f16-vrnd_arch=armv8.2-a+fp16") {
++source_set("f32-f16-vcvt_arm64") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4c8-minmax-sse41-ld64.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4c8-minmax-sse41-ld64.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p8c-minmax-fp32-sse2-mul16-add16.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p8c-minmax-fp32-sse2-mul16-add16.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vrnd/gen/f16-vrndd-neonfp16arith-u16.c",
-+    "src/src/f16-vrnd/gen/f16-vrndne-neonfp16arith-u16.c",
-+    "src/src/f16-vrnd/gen/f16-vrndu-neonfp16arith-u16.c",
-+    "src/src/f16-vrnd/gen/f16-vrndz-neonfp16arith-u16.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -44353,12 +46742,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-neon-u8.c",
++    "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-neonfp16-u16.c",
++    "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-scalar-fabsf-u2.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-dwconv_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -44366,25 +46768,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc8w-gemm_sse4.1-no-sse4.2_standalone") {
--    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4c8-minmax-sse41-ld64.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4c8-minmax-sse41-ld64.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p8c-minmax-fp32-sse2-mul16-add16.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p8c-minmax-fp32-sse2-mul16-add16.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-vrnd_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -44396,33 +46786,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++# This is a target that cannot depend on //base.
++source_set("f32-f16-vcvt_arm64_standalone") {
++  cflags = [
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vrnd/gen/f16-vrndd-neonfp16arith-u16.c",
-+    "src/src/f16-vrnd/gen/f16-vrndne-neonfp16arith-u16.c",
-+    "src/src/f16-vrnd/gen/f16-vrndu-neonfp16arith-u16.c",
-+    "src/src/f16-vrnd/gen/f16-vrndz-neonfp16arith-u16.c"
++    "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-neon-u8.c",
++    "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-neonfp16-u16.c",
++    "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-scalar-fabsf-u2.c"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
+-  source_set("qs8-dwconv_sse4.1-no-sse4.2") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
--
--  source_set("qd8-f32-qc8w-gemm_x64") {
--    cflags = []
--
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x2-minmax-scalar.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4-minmax-scalar.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-2x2-minmax-scalar.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4-minmax-scalar.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p8c-minmax-fp32-sse41-mul16-add16.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p8c-minmax-fp32-sse41-mul16-add16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -44450,21 +46844,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc8w-gemm_x64_standalone") {
--    cflags = []
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x2-minmax-scalar.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4-minmax-scalar.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-2x2-minmax-scalar.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4-minmax-scalar.c",
+-  source_set("qs8-dwconv_sse4.1-no-sse4.2_standalone") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
-+source_set("f16-vrsqrt_arch=armv8.2-a+fp16") {
++source_set("f32-gemm_arch=armv8.2-a+fp16+dotprod") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16"
++    "-march=armv8.2-a+fp16+dotprod"
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p8c-minmax-fp32-sse41-mul16-add16.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p8c-minmax-fp32-sse41-mul16-add16.c",
+-    ]
++  asmflags = cflags
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -44477,7 +46873,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vrsqrt/gen/f16-vrsqrt-neonfp16arith-rsqrt-u16.c"
++    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
++    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
++    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
++    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-cortex-a75.S",
++    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-ld128-acc2.S",
++    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-ld128-acc4.S",
++    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-asm-aarch64-neonfma-ld128.S",
++    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
++    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
++    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a55.S",
++    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
++    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a75.S",
++    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-ld128.S",
++    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
++    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
++    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a55.S",
++    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a73.S",
++    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
++    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-ld128.S"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -44489,6 +46903,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
+-
+-  source_set("qs8-dwconv_x64") {
+-    cflags = []
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -44496,27 +46921,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  source_set("qd8-f32-qc8w-igemm_avx2-avxvnni-f16c-fma") {
--    cflags = [
--      "-mavx2",
--      "-mavxvnni",
--      "-mf16c",
--      "-mfma",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c8-minmax-avxvnni-prfm.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-5x8c8-minmax-avxvnni-prfm.c",
--    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-vrsqrt_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -44527,53 +46931,29 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vrsqrt/gen/f16-vrsqrt-neonfp16arith-rsqrt-u16.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
 +  public_configs = [ ":xnnpack_config" ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
 +}
  
--  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc8w-igemm_avx2-avxvnni-f16c-fma_standalone") {
--    cflags = [
--      "-mavx2",
--      "-mavxvnni",
--      "-mf16c",
--      "-mfma",
--    ]
-+source_set("f16-vsigmoid_arch=armv8.2-a+fp16") {
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++# This is a target that cannot depend on //base.
++source_set("f32-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16"
++    "-march=armv8.2-a+fp16+dotprod"
 +  ]
  
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-dwconv_x64_standalone") {
+-    cflags = []
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c8-minmax-avxvnni-prfm.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-5x8c8-minmax-avxvnni-prfm.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vsigmoid/gen/f16-vsigmoid-neonfp16arith-rr2-p2-nr1fma-u32.c",
-+    "src/src/f16-vsigmoid/gen/f16-vsigmoid-neonfp16arith-rr2-p2-nr1recps-u16.c"
-+  ]
++  asmflags = cflags
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -44585,51 +46965,65 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
++    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
++    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
++    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-cortex-a75.S",
++    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-ld128-acc2.S",
++    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-ld128-acc4.S",
++    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-asm-aarch64-neonfma-ld128.S",
++    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
++    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
++    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a55.S",
++    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
++    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a75.S",
++    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-ld128.S",
++    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
++    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
++    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a55.S",
++    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a73.S",
++    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
++    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-ld128.S"
++  ]
++
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
- 
--    public_configs = [ ":xnnpack_config" ]
++
 +  public_configs = [ ":xnnpack_config" ]
-+}
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
-+# This is a target that cannot depend on //base.
-+source_set("f16-vsigmoid_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
--  source_set("qd8-f32-qc8w-igemm_f16c-fma-avx2") {
+-  source_set("qs8-f16-vcvt_f16c-fma-avx2") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vsigmoid/gen/f16-vsigmoid-neonfp16arith-rr2-p2-nr1fma-u32.c",
-+    "src/src/f16-vsigmoid/gen/f16-vsigmoid-neonfp16arith-rr2-p2-nr1recps-u16.c"
-+  ]
++source_set("f32-gemm_arm64") {
++  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c8-minmax-avx2.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x8c8-minmax-avx2.c",
+-      "src/src/qs8-f16-vcvt/gen/qs8-f16-vcvt-avx2-u16.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -44641,42 +47035,53 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-gemm/gen/f32-gemm-1x4-minmax-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-1x4-relu-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-1x4-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-aarch64-neonfma-lane-ld128.c",
++    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-neon-lane-ld64.c",
++    "src/src/f32-gemm/gen/f32-gemm-1x8s4-minmax-neonfma.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-aarch64-neonfma-lane-ld64.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-neon-lane-ld64.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x2-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x4-minmax-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x4-relu-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x4-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-neon-lane-ld128.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x8s4-minmax-neonfma.c",
++    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-aarch64-neonfma-lane-ld128.c",
++    "src/src/f32-gemm/gen/f32-gemm-6x8s4-minmax-neonfma.c"
 +  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc8w-igemm_f16c-fma-avx2_standalone") {
+-  source_set("qs8-f16-vcvt_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("f16-vsin_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c8-minmax-avx2.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x8c8-minmax-avx2.c",
+-      "src/src/qs8-f16-vcvt/gen/qs8-f16-vcvt-avx2-u16.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vsin/gen/f16-vsin-neonfp16arith-rational-3-2-div.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -44688,82 +47093,52 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++# This is a target that cannot depend on //base.
++source_set("f32-gemm_arm64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set(
--      "qd8-f32-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c8-minmax-avx512skx-prfm.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c8-minmax-avx256skx.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-8x16c8-minmax-avx512skx-prfm.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-8x8c8-minmax-avx256skx.c",
--    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-vsin_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vsin/gen/f16-vsin-neonfp16arith-rational-3-2-div.c"
++    "src/src/f32-gemm/gen/f32-gemm-1x4-minmax-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-1x4-relu-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-1x4-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-aarch64-neonfma-lane-ld128.c",
++    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-neon-lane-ld64.c",
++    "src/src/f32-gemm/gen/f32-gemm-1x8s4-minmax-neonfma.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-aarch64-neonfma-lane-ld64.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-neon-lane-ld64.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x2-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x4-minmax-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x4-relu-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x4-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-neon-lane-ld128.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x8s4-minmax-neonfma.c",
++    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-aarch64-neonfma-lane-ld128.c",
++    "src/src/f32-gemm/gen/f32-gemm-6x8s4-minmax-neonfma.c"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-  source_set("qs8-f32-vcvt_avx-no-avx2-no-f16c-no-fma") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set(
--      "qd8-f32-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-avx-u32.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -44772,15 +47147,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c8-minmax-avx512skx-prfm.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c8-minmax-avx256skx.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-8x16c8-minmax-avx512skx-prfm.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-8x8c8-minmax-avx256skx.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -44789,44 +47155,31 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
  
--    public_configs = [ ":xnnpack_config" ]
-+source_set("f16-vsin_arm64") {
-+  cflags = [
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
--
--  source_set(
--      "qd8-f32-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-f32-vcvt_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
--      "-mf16c",
--      "-mfma",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
-+  ]
++source_set("f32-ibilinear-chw_arm64") {
++  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-10x16c8-minmax-avx512vnni-prfm.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-avx-u32.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vsin/gen/f16-vsin-scalar-rational-3-2-div.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -44837,13 +47190,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-neon-p8.c",
++    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-neonfma-p8.c",
++    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-scalar-p4.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 -  }
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -44852,16 +47215,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set(
--      "qd8-f32-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
+-  source_set("qs8-f32-vcvt_f16c-fma-avx2") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
+-      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
@@ -44870,11 +47226,10 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-10x16c8-minmax-avx512vnni-prfm.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-avx2-u16.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-vsin_arm64_standalone") {
++source_set("f32-ibilinear-chw_arm64_standalone") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -44885,35 +47240,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
+-  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vsin/gen/f16-vsin-scalar-rational-3-2-div.c"
++    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-neon-p8.c",
++    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-neonfma-p8.c",
++    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-scalar-p4.c"
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
--
--  source_set(
--      "qd8-f32-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8") {
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-f32-vcvt_f16c-fma-avx2_standalone") {
 -    cflags = [
--      "-mamx-int8",
--      "-mamx-tile",
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
+-      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
--      "-mgfni",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -44921,8 +47266,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-16x64c4-minmax-avx512amx.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x64c4-minmax-avx512amx.c",
+-      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-avx2-u16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -44939,46 +47283,39 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-   }
++  }
 +}
  
--  # This is a target that cannot depend on //base.
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
+-
 -  source_set(
--      "qd8-f32-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8_standalone") {
+-      "qs8-f32-vcvt_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 -    cflags = [
--      "-mamx-int8",
--      "-mamx-tile",
 -      "-mavx512bw",
 -      "-mavx512cd",
 -      "-mavx512dq",
 -      "-mavx512f",
 -      "-mavx512vl",
--      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
--      "-mgfni",
 -    ]
-+source_set("f16-vsqrt_arch=armv8.2-a+fp16") {
++source_set("f32-ibilinear_arm64") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-16x64c4-minmax-avx512amx.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x64c4-minmax-avx512amx.c",
+-      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-avx512skx-u32.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vsqrt/gen/f16-vsqrt-aarch64-neonfp16arith-sqrt-u8.c",
-+    "src/src/f16-vsqrt/gen/f16-vsqrt-neonfp16arith-nr1fma1adj-u8.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -44989,99 +47326,46 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--  source_set("qd8-f32-qc8w-igemm_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-vsqrt_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4c8-minmax-sse2-ld64.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4c8-minmax-sse2-ld64.c",
+-      "//third_party/pthreadpool",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vsqrt/gen/f16-vsqrt-aarch64-neonfp16arith-sqrt-u8.c",
-+    "src/src/f16-vsqrt/gen/f16-vsqrt-neonfp16arith-nr1fma1adj-u8.c"
++    "src/src/f32-ibilinear/gen/f32-ibilinear-neon-c8.c",
++    "src/src/f32-ibilinear/gen/f32-ibilinear-neonfma-c8.c",
++    "src/src/f32-ibilinear/gen/f32-ibilinear-scalar-c2.c"
 +  ]
  
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
+-  # This is a target that cannot depend on //base.
+-  source_set(
+-      "qs8-f32-vcvt_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
-+
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
- 
--  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc8w-igemm_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
-+source_set("f16-vtanh_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
++    "//third_party/pthreadpool",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4c8-minmax-sse2-ld64.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4c8-minmax-sse2-ld64.c",
+-      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-avx512skx-u32.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vtanh/gen/f16-vtanh-aarch64-neonfp16arith-expm1minus-rr1-p3h2ts-div-u32.c",
-+    "src/src/f16-vtanh/gen/f16-vtanh-neonfp16arith-expm1minus-rr1-p3h2ts-nr1fma-u32.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -45093,45 +47377,42 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++# This is a target that cannot depend on //base.
++source_set("f32-ibilinear_arm64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-ibilinear/gen/f32-ibilinear-neon-c8.c",
++    "src/src/f32-ibilinear/gen/f32-ibilinear-neonfma-c8.c",
++    "src/src/f32-ibilinear/gen/f32-ibilinear-scalar-c2.c"
++  ]
  
--  source_set("qd8-f32-qc8w-igemm_sse4.1-no-sse4.2") {
+-  source_set("qs8-f32-vcvt_sse2-no-sse3") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-vtanh_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4c8-minmax-sse41-ld64.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4c8-minmax-sse41-ld64.c",
+-      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-sse2-u32.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vtanh/gen/f16-vtanh-aarch64-neonfp16arith-expm1minus-rr1-p3h2ts-div-u32.c",
-+    "src/src/f16-vtanh/gen/f16-vtanh-neonfp16arith-expm1minus-rr1-p3h2ts-nr1fma-u32.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -45144,46 +47425,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
-+
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
    }
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc8w-igemm_sse4.1-no-sse4.2_standalone") {
+-  source_set("qs8-f32-vcvt_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+source_set("f16-vunary_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4c8-minmax-sse41-ld64.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4c8-minmax-sse41-ld64.c",
+-      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-sse2-u32.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-vunary/gen/f16-vabs-neonfp16arith-u16.c",
-+    "src/src/f16-vunary/gen/f16-vneg-neonfp16arith-u16.c",
-+    "src/src/f16-vunary/gen/f16-vsqr-neonfp16arith-u16.c"
++source_set("f32-igemm_arch=armv8.2-a+fp16+dotprod") {
++  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -45196,11 +47459,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  asmflags = cflags
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-igemm/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a55.S",
++    "src/src/f32-igemm/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a55.S",
++    "src/src/f32-igemm/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a73.S",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-cortex-a75.S",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-ld64.S",
++    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
++    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
++    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
++    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
++    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a75.S",
++    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-ld128.S",
++    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
++    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
++    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
++    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-asm-aarch64-neonfma-ld128.S"
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
+-  source_set("qs8-f32-vcvt_sse4.1-no-sse4.2") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -45208,20 +47504,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
--
--  source_set("qd8-f32-qc8w-igemm_x64") {
--    cflags = []
--
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x2-minmax-scalar.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4-minmax-scalar.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-2x2-minmax-scalar.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4-minmax-scalar.c",
+-      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-sse41-u16.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -45237,30 +47522,46 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-vunary_arch=armv8.2-a+fp16_standalone") {
++source_set("f32-igemm_arch=armv8.2-a+fp16+dotprod_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16"
++    "-march=armv8.2-a+fp16+dotprod"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
++  asmflags = cflags
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-f32-vcvt_sse4.1-no-sse4.2_standalone") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vunary/gen/f16-vabs-neonfp16arith-u16.c",
-+    "src/src/f16-vunary/gen/f16-vneg-neonfp16arith-u16.c",
-+    "src/src/f16-vunary/gen/f16-vsqr-neonfp16arith-u16.c"
++    "src/src/f32-igemm/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a55.S",
++    "src/src/f32-igemm/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a55.S",
++    "src/src/f32-igemm/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a73.S",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-cortex-a75.S",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-ld64.S",
++    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
++    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
++    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
++    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
++    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a75.S",
++    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-ld128.S",
++    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
++    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
++    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
++    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-asm-aarch64-neonfma-ld128.S"
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc8w-igemm_x64_standalone") {
--    cflags = []
--
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x2-minmax-scalar.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4-minmax-scalar.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-2x2-minmax-scalar.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4-minmax-scalar.c",
+-      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-sse41-u16.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -45294,20 +47595,14 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-dwconv_avx-no-avx2-no-f16c-no-fma") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
-+source_set("f32-argmaxpool_arm64") {
+-  source_set("qs8-f32-vcvt_x64") {
+-    cflags = []
++source_set("f32-igemm_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p16c-minmax-fp32-avx-mul16-add16.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p16c-minmax-fp32-avx-mul16-add16.c",
+-      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u4.c",
 -    ]
 +  ]
  
@@ -45323,8 +47618,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-neon-c4.c",
-+    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c"
++    "src/src/f32-igemm/gen/f32-igemm-1x4-minmax-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x4-relu-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x4-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-aarch64-neonfma-lane-ld128.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-neon-lane-ld64.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x8s4-minmax-neonfma.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-aarch64-neonfma-lane-ld64.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-neon-lane-ld64.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x2-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x4-minmax-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x4-relu-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x4-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-neon-lane-ld128.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x8s4-minmax-neonfma.c",
++    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-aarch64-neonfma-lane-ld128.c",
++    "src/src/f32-igemm/gen/f32-igemm-6x8s4-minmax-neonfma.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -45334,13 +47644,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-dwconv_avx-no-avx2-no-f16c-no-fma_standalone") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
+-  source_set("qs8-f32-vcvt_x64_standalone") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -45350,8 +47655,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p16c-minmax-fp32-avx-mul16-add16.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p16c-minmax-fp32-avx-mul16-add16.c",
+-      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u4.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -45367,7 +47671,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-argmaxpool_arm64_standalone") {
++source_set("f32-igemm_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -45379,13 +47683,29 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-neon-c4.c",
-+    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c"
++    "src/src/f32-igemm/gen/f32-igemm-1x4-minmax-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x4-relu-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x4-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-aarch64-neonfma-lane-ld128.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-neon-lane-ld64.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x8s4-minmax-neonfma.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-aarch64-neonfma-lane-ld64.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-neon-lane-ld64.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x2-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x4-minmax-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x4-relu-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x4-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-neon-lane-ld128.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x8s4-minmax-neonfma.c",
++    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-aarch64-neonfma-lane-ld128.c",
++    "src/src/f32-igemm/gen/f32-igemm-6x8s4-minmax-neonfma.c"
 +  ]
  
--  source_set("qs8-dwconv_f16c-fma-avx2") {
+-  source_set("qs8-packw_avx2-avxvnni-f16c-fma") {
 -    cflags = [
 -      "-mavx2",
+-      "-mavxvnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
@@ -45395,8 +47715,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p16c-minmax-fp32-avx2-mul32.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p16c-minmax-fp32-avx2-mul32.c",
+-      "src/src/qs8-packw/gen/qs8-packw-x8c8-gemm-goi-avxvnni-prfm.c",
+-      "src/src/qs8-packw/gen/qs8-packw-x8c8-gemm-goi-avxvnni.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -45424,19 +47744,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-dwconv_f16c-fma-avx2_standalone") {
+-  source_set("qs8-packw_avx2-avxvnni-f16c-fma_standalone") {
 -    cflags = [
 -      "-mavx2",
+-      "-mavxvnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("f32-avgpool_arm64") {
++source_set("f32-maxpool_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p16c-minmax-fp32-avx2-mul32.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p16c-minmax-fp32-avx2-mul32.c",
+-      "src/src/qs8-packw/gen/qs8-packw-x8c8-gemm-goi-avxvnni-prfm.c",
+-      "src/src/qs8-packw/gen/qs8-packw-x8c8-gemm-goi-avxvnni.c",
 -    ]
 +  ]
  
@@ -45452,8 +47773,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-neon-u4.c",
-+    "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-scalar-u1.c"
++    "src/src/f32-maxpool/gen/f32-maxpool-9p-minmax-neon-u4.c",
++    "src/src/f32-maxpool/gen/f32-maxpool-9p-minmax-scalar-u1.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -45467,13 +47788,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 -
 -  source_set(
--      "qs8-dwconv_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-      "qs8-packw_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
 -      "-mavx512dq",
 -      "-mavx512f",
 -      "-mavx512vl",
+-      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
@@ -45486,8 +47808,9 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p32c-minmax-fp32-avx512skx-mul32.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p32c-minmax-fp32-avx512skx-mul32.c",
+-      "src/src/qs8-packw/gen/qs8-packw-x16c8-gemm-goi-avx256vnni-prfm.c",
+-      "src/src/qs8-packw/gen/qs8-packw-x64c4-gemm-goi-avx256vnni-prfm.c",
+-      "src/src/qs8-packw/gen/qs8-packw-x8c8-gemm-goi-avx256vnni-prfm.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -45503,7 +47826,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-avgpool_arm64_standalone") {
++source_set("f32-maxpool_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -45512,26 +47835,28 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
--      "qs8-dwconv_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-      "qs8-packw_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
 -      "-mavx512dq",
 -      "-mavx512f",
 -      "-mavx512vl",
+-      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-neon-u4.c",
-+    "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-scalar-u1.c"
++    "src/src/f32-maxpool/gen/f32-maxpool-9p-minmax-neon-u4.c",
++    "src/src/f32-maxpool/gen/f32-maxpool-9p-minmax-scalar-u1.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p32c-minmax-fp32-avx512skx-mul32.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p32c-minmax-fp32-avx512skx-mul32.c",
+-      "src/src/qs8-packw/gen/qs8-packw-x16c8-gemm-goi-avx256vnni-prfm.c",
+-      "src/src/qs8-packw/gen/qs8-packw-x64c4-gemm-goi-avx256vnni-prfm.c",
+-      "src/src/qs8-packw/gen/qs8-packw-x8c8-gemm-goi-avx256vnni-prfm.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -45565,18 +47890,14 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-dwconv_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
-+source_set("f32-conv-hwc2chw_arm64") {
+-  source_set("qs8-packw_x64") {
+-    cflags = []
++source_set("f32-qc4w-gemm_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p8c-minmax-fp32-sse2-mul16-add16.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p8c-minmax-fp32-sse2-mul16-add16.c",
+-      "src/src/qs8-packw/gen/qs8-packw-x16c8-gemm-goi-scalar.c",
 -    ]
 +  ]
  
@@ -45592,9 +47913,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-aarch64-neonfma-2x2.c",
-+    "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-neon-2x2.c",
-+    "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-scalar-1x1.c"
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-1x4-minmax-scalar.c",
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-1x8-minmax-aarch64-neonfma-lane-ld128.c",
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-1x8-minmax-neon-lane-ld64.c",
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-4x4-minmax-scalar.c",
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-4x8-minmax-aarch64-neonfma-lane-ld128.c",
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-4x8-minmax-neon-lane-ld64.c",
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-6x8-minmax-aarch64-neonfma-lane-ld128.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -45604,11 +47929,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-dwconv_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("qs8-packw_x64_standalone") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -45618,8 +47940,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p8c-minmax-fp32-sse2-mul16-add16.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p8c-minmax-fp32-sse2-mul16-add16.c",
+-      "src/src/qs8-packw/gen/qs8-packw-x16c8-gemm-goi-scalar.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -45635,7 +47956,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-conv-hwc2chw_arm64_standalone") {
++source_set("f32-qc4w-gemm_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -45647,24 +47968,24 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-aarch64-neonfma-2x2.c",
-+    "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-neon-2x2.c",
-+    "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-scalar-1x1.c"
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-1x4-minmax-scalar.c",
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-1x8-minmax-aarch64-neonfma-lane-ld128.c",
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-1x8-minmax-neon-lane-ld64.c",
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-4x4-minmax-scalar.c",
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-4x8-minmax-aarch64-neonfma-lane-ld128.c",
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-4x8-minmax-neon-lane-ld64.c",
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-6x8-minmax-aarch64-neonfma-lane-ld128.c"
 +  ]
  
--  source_set("qs8-dwconv_sse4.1-no-sse4.2") {
--    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
--    ]
+-  source_set("qs8-qc4w-gemm_x64") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p8c-minmax-fp32-sse41-mul16-add16.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p8c-minmax-fp32-sse41-mul16-add16.c",
+-      "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-1x2-minmax-fp32-scalar-lrintf.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -45692,19 +48013,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-dwconv_sse4.1-no-sse4.2_standalone") {
--    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
--    ]
-+source_set("f32-dwconv2d-chw_arm64") {
-+  cflags = [
- 
+-  source_set("qs8-qc4w-gemm_x64_standalone") {
+-    cflags = []
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p8c-minmax-fp32-sse41-mul16-add16.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p8c-minmax-fp32-sse41-mul16-add16.c",
+-      "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-1x2-minmax-fp32-scalar-lrintf.c",
 -    ]
++source_set("f32-qc8w-gemm_arch=armv8.2-a+fp16+dotprod") {
++  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -45717,47 +48035,30 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  asmflags = cflags
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-aarch64-neonfma-3x4.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-neon-2x4.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-2x1-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-4x1.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-aarch64-neonfma-2x4-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-neon-1x4.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-1x1-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-2x1-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-aarch64-neonfma-4x4.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-neon-1x4.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-1x1-acc5.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-2x1-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-aarch64-neonfma-1x4-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-neon-1x4.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-1x1-acc5.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-2x1-acc2.c"
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x8-minmax-asm-aarch64-neonfma-ld128-acc4.S",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x8-minmax-asm-aarch64-neonfma-ld128.S",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-6x8-minmax-asm-aarch64-neonfma-ld128.S"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set("qs8-dwconv_x64") {
--    cflags = []
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  source_set("qs8-qc8w-dwconv_avx-no-avx2-no-f16c-no-fma") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -45766,6 +48067,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p16c-minmax-fp32-avx-mul16-add16.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-avx-mul16-add16.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-avx-mul16-add16.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -45776,29 +48086,40 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++# This is a target that cannot depend on //base.
++source_set("f32-qc8w-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+# This is a target that cannot depend on //base.
-+source_set("f32-dwconv2d-chw_arm64_standalone") {
-+  cflags = [
++  asmflags = cflags
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-dwconv_x64_standalone") {
--    cflags = []
--
+-  source_set("qs8-qc8w-dwconv_avx-no-avx2-no-f16c-no-fma_standalone") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x8-minmax-asm-aarch64-neonfma-ld128-acc4.S",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x8-minmax-asm-aarch64-neonfma-ld128.S",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-6x8-minmax-asm-aarch64-neonfma-ld128.S"
++  ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p16c-minmax-fp32-avx-mul16-add16.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-avx-mul16-add16.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-avx-mul16-add16.c",
 -    ]
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -45810,30 +48131,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-aarch64-neonfma-3x4.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-neon-2x4.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-2x1-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-4x1.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-aarch64-neonfma-2x4-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-neon-1x4.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-1x1-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-2x1-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-aarch64-neonfma-4x4.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-neon-1x4.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-1x1-acc5.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-2x1-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-aarch64-neonfma-1x4-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-neon-1x4.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-1x1-acc5.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-2x1-acc2.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -45852,22 +48149,22 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-f16-vcvt_f16c-fma-avx2") {
+-  source_set("qs8-qc8w-dwconv_f16c-fma-avx2") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("f32-dwconv_arch=armv8.2-a+fp16+dotprod") {
++source_set("f32-qc8w-gemm_arm64") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-f16-vcvt/gen/qs8-f16-vcvt-avx2-u16.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p16c-minmax-fp32-avx2-mul32.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-avx2-mul32.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-avx2-mul32.c",
 -    ]
-+  asmflags = cflags
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -45881,7 +48178,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-dwconv/f32-dwconv-9p4c-minmax-asm-aarch64-neonfma-cortex-a55.S"
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x4-minmax-scalar.c",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x8-minmax-aarch64-neonfma-lane-ld64.c",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x8-minmax-neon-lane-ld64.c",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-4x4-minmax-scalar.c",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-4x8-minmax-neon-lane-ld64.c",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-6x8-minmax-aarch64-neonfma-lane-ld64.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -45891,7 +48193,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-f16-vcvt_f16c-fma-avx2_standalone") {
+-  source_set("qs8-qc8w-dwconv_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
@@ -45906,7 +48208,9 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-f16-vcvt/gen/qs8-f16-vcvt-avx2-u16.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p16c-minmax-fp32-avx2-mul32.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-avx2-mul32.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-avx2-mul32.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -45922,45 +48226,48 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-dwconv_arch=armv8.2-a+fp16+dotprod_standalone") {
++source_set("f32-qc8w-gemm_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  asmflags = cflags
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
+-
+-  source_set(
+-      "qs8-qc8w-dwconv_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-dwconv/f32-dwconv-9p4c-minmax-asm-aarch64-neonfma-cortex-a55.S"
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x4-minmax-scalar.c",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x8-minmax-aarch64-neonfma-lane-ld64.c",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x8-minmax-neon-lane-ld64.c",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-4x4-minmax-scalar.c",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-4x8-minmax-neon-lane-ld64.c",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-6x8-minmax-aarch64-neonfma-lane-ld64.c"
 +  ]
  
--  source_set("qs8-f32-vcvt_avx-no-avx2-no-f16c-no-fma") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p32c-minmax-fp32-avx512skx-mul32.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p32c-minmax-fp32-avx512skx-mul32.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p32c-minmax-fp32-avx512skx-mul32.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-avx-u32.c",
--    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -45971,30 +48278,42 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  public_configs = [ ":xnnpack_config" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-f32-vcvt_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set(
+-      "qs8-qc8w-dwconv_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+source_set("f32-dwconv_arm64") {
-+  cflags = [
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-avx-u32.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p32c-minmax-fp32-avx512skx-mul32.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p32c-minmax-fp32-avx512skx-mul32.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p32c-minmax-fp32-avx512skx-mul32.c",
 -    ]
-+  ]
++source_set("f32-qs8-vcvt_arm64") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -46006,37 +48325,30 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p1c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p1c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p2c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p2c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p8c-minmax-neon.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p8c-minmax-neonfma-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-3p1c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-3p1c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-3p8c-minmax-neon.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-3p8c-minmax-neonfma.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-4p1c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-4p1c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-4p8c-minmax-neon.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-4p8c-minmax-neonfma.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-9p1c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-9p1c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-9p8c-minmax-neon.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-9p8c-minmax-neonfma.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-neon-u32.c",
++    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-neonv8-u32.c",
++    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u4.c",
++    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-lrintf-u4.c"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  source_set("qs8-qc8w-dwconv_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -46044,22 +48356,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  source_set("qs8-f32-vcvt_f16c-fma-avx2") {
--    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-avx2-u16.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-sse2-mul16.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p8c-minmax-fp32-sse2-mul16.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p8c-minmax-fp32-sse2-mul16.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-dwconv_arm64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -46071,53 +48375,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  ]
++# This is a target that cannot depend on //base.
++source_set("f32-qs8-vcvt_arm64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p1c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p1c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p2c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p2c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p8c-minmax-neon.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p8c-minmax-neonfma-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-3p1c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-3p1c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-3p8c-minmax-neon.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-3p8c-minmax-neonfma.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-4p1c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-4p1c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-4p8c-minmax-neon.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-4p8c-minmax-neonfma.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-9p1c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-9p1c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-9p8c-minmax-neon.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-9p8c-minmax-neonfma.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-f32-vcvt_f16c-fma-avx2_standalone") {
+-  source_set("qs8-qc8w-dwconv_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-neon-u32.c",
++    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-neonv8-u32.c",
++    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u4.c",
++    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-lrintf-u4.c"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-avx2-u16.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-sse2-mul16.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p8c-minmax-fp32-sse2-mul16.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p8c-minmax-fp32-sse2-mul16.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -46129,36 +48417,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
--
--  source_set(
--      "qs8-f32-vcvt_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  source_set("qs8-qc8w-dwconv_sse4.1-no-sse4.2") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
-+source_set("f32-f16-vcvt_arm64") {
++source_set("f32-qu8-vcvt_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-avx512skx-u32.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-sse41-mul16.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p8c-minmax-fp32-sse41-mul16.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p8c-minmax-fp32-sse41-mul16.c",
 -    ]
 +  ]
  
@@ -46174,9 +48463,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-neon-u8.c",
-+    "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-neonfp16-u16.c",
-+    "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-scalar-fabsf-u2.c"
++    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-neon-u32.c",
++    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-neonv8-u32.c",
++    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u4.c",
++    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-lrintf-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -46186,16 +48476,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qs8-f32-vcvt_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-  source_set("qs8-qc8w-dwconv_sse4.1-no-sse4.2_standalone") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -46206,7 +48490,9 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-avx512skx-u32.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-sse41-mul16.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p8c-minmax-fp32-sse41-mul16.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p8c-minmax-fp32-sse41-mul16.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -46222,43 +48508,41 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-f16-vcvt_arm64_standalone") {
++source_set("f32-qu8-vcvt_arm64_standalone") {
 +  cflags = [
++
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-neon-u32.c",
++    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-neonv8-u32.c",
++    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u4.c",
++    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-lrintf-u4.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-neon-u8.c",
-+    "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-neonfp16-u16.c",
-+    "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-scalar-fabsf-u2.c"
-+  ]
- 
--  source_set("qs8-f32-vcvt_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-
+-  source_set("qs8-qc8w-dwconv_x64") {
+-    cflags = []
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p1c-minmax-fp32-scalar-fmagic.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-sse2-u32.c",
--    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -46269,6 +48553,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
 +  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -46278,20 +48569,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-f32-vcvt_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("qs8-qc8w-dwconv_x64_standalone") {
+-    cflags = []
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-sse2-u32.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p1c-minmax-fp32-scalar-fmagic.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
+-      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
 -    ]
-+source_set("f32-gemm_arch=armv8.2-a+fp16+dotprod") {
++source_set("f32-raddstoreexpminusmax_arm64") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -46303,30 +48594,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  asmflags = cflags
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-cortex-a75.S",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-ld128-acc2.S",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-ld128-acc4.S",
-+    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-asm-aarch64-neonfma-ld128.S",
-+    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
-+    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
-+    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a55.S",
-+    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
-+    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a75.S",
-+    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-ld128.S",
-+    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
-+    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
-+    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a55.S",
-+    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a73.S",
-+    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
-+    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-ld128.S"
++    "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-neonfma-rr1-lut64-p2-u16-acc2.c",
++    "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-scalar-rr2-p5-u4-acc2.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -46337,10 +48611,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qs8-f32-vcvt_sse4.1-no-sse4.2") {
+-  source_set("qs8-qc8w-gemm_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -46351,7 +48627,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-sse41-u16.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4c8-minmax-fp32-avx-ld128.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x4c8-minmax-fp32-avx-ld128.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -46367,47 +48644,31 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
++source_set("f32-raddstoreexpminusmax_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  asmflags = cflags
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-f32-vcvt_sse4.1-no-sse4.2_standalone") {
+-  source_set("qs8-qc8w-gemm_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-cortex-a75.S",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-ld128-acc2.S",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-asm-aarch64-neonfma-ld128-acc4.S",
-+    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-asm-aarch64-neonfma-ld128.S",
-+    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
-+    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
-+    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a55.S",
-+    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
-+    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-cortex-a75.S",
-+    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-asm-aarch64-neonfma-ld128.S",
-+    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
-+    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
-+    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a55.S",
-+    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a73.S",
-+    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
-+    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-asm-aarch64-neonfma-ld128.S"
++    "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-neonfma-rr1-lut64-p2-u16-acc2.c",
++    "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-scalar-rr2-p5-u4-acc2.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-sse41-u16.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4c8-minmax-fp32-avx-ld128.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x4c8-minmax-fp32-avx-ld128.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -46441,15 +48702,20 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-f32-vcvt_x64") {
--    cflags = []
-+source_set("f32-gemm_arm64") {
+-  source_set("qs8-qc8w-gemm_avx2-avxvnni-f16c-fma") {
+-    cflags = [
+-      "-mavx2",
+-      "-mavxvnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++source_set("f32-rdminmax_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u1.c",
--      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u4.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-avxvnni-prfm.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-5x8c8-minmax-fp32-avxvnni-prfm.c",
 -    ]
 +  ]
  
@@ -46465,26 +48731,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-gemm/gen/f32-gemm-1x4-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-1x4-relu-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-1x4-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-aarch64-neonfma-lane-ld128.c",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-neon-lane-ld64.c",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8s4-minmax-neonfma.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-relu-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-aarch64-neonfma-lane-ld64.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-neon-lane-ld64.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x2-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x4-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x4-relu-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x4-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-neon-lane-ld128.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x8s4-minmax-neonfma.c",
-+    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-aarch64-neonfma-lane-ld128.c",
-+    "src/src/f32-gemm/gen/f32-gemm-6x8s4-minmax-neonfma.c"
++    "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-neon-c32.c",
++    "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-scalar-c2.c",
++    "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-neon-c32.c",
++    "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-scalar-c2.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -46494,8 +48744,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-f32-vcvt_x64_standalone") {
--    cflags = []
+-  source_set("qs8-qc8w-gemm_avx2-avxvnni-f16c-fma_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mavxvnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -46505,8 +48760,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u1.c",
--      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u4.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-avxvnni-prfm.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-5x8c8-minmax-fp32-avxvnni-prfm.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -46522,7 +48777,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-gemm_arm64_standalone") {
++source_set("f32-rdminmax_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -46534,32 +48789,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-gemm/gen/f32-gemm-1x4-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-1x4-relu-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-1x4-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-aarch64-neonfma-lane-ld128.c",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-neon-lane-ld64.c",
-+    "src/src/f32-gemm/gen/f32-gemm-1x8s4-minmax-neonfma.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-relu-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-aarch64-neonfma-lane-ld64.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-neon-lane-ld64.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x2-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x4-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x4-relu-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x4-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x8-minmax-neon-lane-ld128.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x8s4-minmax-neonfma.c",
-+    "src/src/f32-gemm/gen/f32-gemm-6x8-minmax-aarch64-neonfma-lane-ld128.c",
-+    "src/src/f32-gemm/gen/f32-gemm-6x8s4-minmax-neonfma.c"
++    "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-neon-c32.c",
++    "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-scalar-c2.c",
++    "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-neon-c32.c",
++    "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-scalar-c2.c"
 +  ]
  
--  source_set("qs8-packw_avx2-avxvnni-f16c-fma") {
+-  source_set("qs8-qc8w-gemm_avx2-avxvnniint8-f16c-fma") {
 -    cflags = [
 -      "-mavx2",
--      "-mavxvnni",
+-      "-mavxvnniint8",
 -      "-mf16c",
 -      "-mfma",
 -    ]
@@ -46569,8 +48808,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-packw/gen/qs8-packw-x8c8-gemm-goi-avxvnni-prfm.c",
--      "src/src/qs8-packw/gen/qs8-packw-x8c8-gemm-goi-avxvnni.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-avxvnniint8-prfm.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-5x8c8-minmax-fp32-avxvnniint8-prfm.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -46598,20 +48837,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-packw_avx2-avxvnni-f16c-fma_standalone") {
+-  source_set("qs8-qc8w-gemm_avx2-avxvnniint8-f16c-fma_standalone") {
 -    cflags = [
 -      "-mavx2",
--      "-mavxvnni",
+-      "-mavxvnniint8",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("f32-ibilinear-chw_arm64") {
++source_set("f32-rdsum_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-packw/gen/qs8-packw-x8c8-gemm-goi-avxvnni-prfm.c",
--      "src/src/qs8-packw/gen/qs8-packw-x8c8-gemm-goi-avxvnni.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-avxvnniint8-prfm.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-5x8c8-minmax-fp32-avxvnniint8-prfm.c",
 -    ]
 +  ]
  
@@ -46627,48 +48866,45 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-neon-p8.c",
-+    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-neonfma-p8.c",
-+    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-scalar-p4.c"
++    "src/src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-neon-c16.c",
++    "src/src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-scalar.c"
 +  ]
- 
--    public_configs = [ ":xnnpack_config" ]
++
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
+-    public_configs = [ ":xnnpack_config" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set(
--      "qs8-packw_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-  source_set("qs8-qc8w-gemm_f16c-fma-avx2") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
+-      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++# This is a target that cannot depend on //base.
++source_set("f32-rdsum_arm64_standalone") {
++  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-packw/gen/qs8-packw-x16c8-gemm-goi-avx256vnni-prfm.c",
--      "src/src/qs8-packw/gen/qs8-packw-x64c4-gemm-goi-avx256vnni-prfm.c",
--      "src/src/qs8-packw/gen/qs8-packw-x8c8-gemm-goi-avx256vnni-prfm.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-avx2.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x8c8-minmax-fp32-avx2.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -46680,43 +48916,38 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-ibilinear-chw_arm64_standalone") {
-+  cflags = [
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-neon-c16.c",
++    "src/src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-scalar.c"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qs8-packw_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
+-  source_set("qs8-qc8w-gemm_f16c-fma-avx2_standalone") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
+-      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-neon-p8.c",
-+    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-neonfma-p8.c",
-+    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-scalar-p4.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-packw/gen/qs8-packw-x16c8-gemm-goi-avx256vnni-prfm.c",
--      "src/src/qs8-packw/gen/qs8-packw-x64c4-gemm-goi-avx256vnni-prfm.c",
--      "src/src/qs8-packw/gen/qs8-packw-x8c8-gemm-goi-avx256vnni-prfm.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-avx2.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x8c8-minmax-fp32-avx2.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -46728,33 +48959,48 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
++source_set("f32-rminmax_arm64") {
++  cflags = [
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
- 
--  source_set("qs8-packw_x64") {
--    cflags = []
-+source_set("f32-ibilinear_arm64") {
-+  cflags = [
+-  }
+-
+-  source_set(
+-      "qs8-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-packw/gen/qs8-packw-x16c8-gemm-goi-scalar.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-avx256skx.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x8c8-minmax-fp32-avx256skx.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-7x16c8-minmax-fp32-avx512skx-prfm.c",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-rminmax/gen/f32-rmax-neon-u16-acc4.c",
++    "src/src/f32-rminmax/gen/f32-rmax-scalar-u4-acc4.c",
++    "src/src/f32-rminmax/gen/f32-rmin-neon-u16-acc4.c",
++    "src/src/f32-rminmax/gen/f32-rmin-scalar-u4-acc4.c",
++    "src/src/f32-rminmax/gen/f32-rminmax-neon-u16-acc4.c",
++    "src/src/f32-rminmax/gen/f32-rminmax-scalar-u4-acc4.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -46767,22 +49013,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-ibilinear/gen/f32-ibilinear-neon-c8.c",
-+    "src/src/f32-ibilinear/gen/f32-ibilinear-neonfma-c8.c",
-+    "src/src/f32-ibilinear/gen/f32-ibilinear-scalar-c2.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qs8-packw_x64_standalone") {
--    cflags = []
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -46790,13 +49026,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-packw/gen/qs8-packw-x16c8-gemm-goi-scalar.c",
+-  # This is a target that cannot depend on //base.
+-  source_set(
+-      "qs8-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-avx256skx.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x8c8-minmax-fp32-avx256skx.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-7x16c8-minmax-fp32-avx512skx-prfm.c",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("f32-rminmax_arm64_standalone") {
++  cflags = [
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -46807,33 +49062,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-ibilinear_arm64_standalone") {
-+  cflags = [
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-rminmax/gen/f32-rmax-neon-u16-acc4.c",
++    "src/src/f32-rminmax/gen/f32-rmax-scalar-u4-acc4.c",
++    "src/src/f32-rminmax/gen/f32-rmin-neon-u16-acc4.c",
++    "src/src/f32-rminmax/gen/f32-rmin-scalar-u4-acc4.c",
++    "src/src/f32-rminmax/gen/f32-rminmax-neon-u16-acc4.c",
++    "src/src/f32-rminmax/gen/f32-rminmax-scalar-u4-acc4.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-ibilinear/gen/f32-ibilinear-neon-c8.c",
-+    "src/src/f32-ibilinear/gen/f32-ibilinear-neonfma-c8.c",
-+    "src/src/f32-ibilinear/gen/f32-ibilinear-scalar-c2.c"
-+  ]
- 
--  source_set("qs8-qc4w-gemm_x64") {
--    cflags = []
+-
+-  source_set(
+-      "qs8-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-1x2-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16c8-minmax-fp32-avx512vnni-prfm.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-7x16c8-minmax-fp32-avx512vnni-prfm.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -46861,16 +49127,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc4w-gemm_x64_standalone") {
--    cflags = []
--
+-  source_set(
+-      "qs8-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++source_set("f32-rsum_arm64") {
++  cflags = [
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-1x2-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16c8-minmax-fp32-avx512vnni-prfm.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-7x16c8-minmax-fp32-avx512vnni-prfm.c",
 -    ]
-+source_set("f32-igemm_arch=armv8.2-a+fp16+dotprod") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -46883,45 +49159,36 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  asmflags = cflags
- 
--    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-igemm/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a55.S",
-+    "src/src/f32-igemm/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a55.S",
-+    "src/src/f32-igemm/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a73.S",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-cortex-a75.S",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-ld64.S",
-+    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
-+    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
-+    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
-+    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
-+    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a75.S",
-+    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-ld128.S",
-+    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
-+    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
-+    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
-+    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-asm-aarch64-neonfma-ld128.S"
++    "src/src/f32-rsum/gen/f32-rsum-neon-u16-acc4.c",
++    "src/src/f32-rsum/gen/f32-rsum-scalar-u4-acc4.c"
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qs8-qc8w-dwconv_avx-no-avx2-no-f16c-no-fma") {
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
+-
+-  source_set(
+-      "qs8-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mamx-int8",
+-      "-mamx-tile",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
+-      "-mgfni",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -46932,9 +49199,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p16c-minmax-fp32-avx-mul16-add16.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-avx-mul16-add16.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-avx-mul16-add16.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-16x64c4-minmax-fp32-avx512amx.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x64c4-minmax-fp32-avx512amx.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -46950,50 +49216,39 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-igemm_arch=armv8.2-a+fp16+dotprod_standalone") {
++source_set("f32-rsum_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  asmflags = cflags
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-dwconv_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set(
+-      "qs8-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8_standalone") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mamx-int8",
+-      "-mamx-tile",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
+-      "-mgfni",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-igemm/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a55.S",
-+    "src/src/f32-igemm/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a55.S",
-+    "src/src/f32-igemm/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a73.S",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-cortex-a75.S",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-asm-aarch64-neonfma-ld64.S",
-+    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
-+    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
-+    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
-+    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
-+    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-cortex-a75.S",
-+    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-asm-aarch64-neonfma-ld128.S",
-+    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a53-prfm.S",
-+    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a53.S",
-+    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-asm-aarch64-neonfma-cortex-a75-prfm.S",
-+    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-asm-aarch64-neonfma-ld128.S"
++    "src/src/f32-rsum/gen/f32-rsum-neon-u16-acc4.c",
++    "src/src/f32-rsum/gen/f32-rsum-scalar-u4-acc4.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p16c-minmax-fp32-avx-mul16-add16.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-avx-mul16-add16.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-avx-mul16-add16.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-16x64c4-minmax-fp32-avx512amx.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x64c4-minmax-fp32-avx512amx.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -47027,20 +49282,18 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-qc8w-dwconv_f16c-fma-avx2") {
+-  source_set("qs8-qc8w-gemm_sse2-no-sse3") {
 -    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+source_set("f32-igemm_arm64") {
++source_set("f32-spmm_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p16c-minmax-fp32-avx2-mul32.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-avx2-mul32.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-avx2-mul32.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4c8-minmax-fp32-sse2-ld64.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4c8-minmax-fp32-sse2-ld64.c",
 -    ]
 +  ]
  
@@ -47056,26 +49309,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-igemm/gen/f32-igemm-1x4-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x4-relu-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x4-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-aarch64-neonfma-lane-ld128.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-neon-lane-ld64.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8s4-minmax-neonfma.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-relu-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-aarch64-neonfma-lane-ld64.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-neon-lane-ld64.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x2-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x4-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x4-relu-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x4-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-neon-lane-ld128.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x8s4-minmax-neonfma.c",
-+    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-aarch64-neonfma-lane-ld128.c",
-+    "src/src/f32-igemm/gen/f32-igemm-6x8s4-minmax-neonfma.c"
++    "src/src/f32-spmm/gen/f32-spmm-32x1-minmax-neon.c",
++    "src/src/f32-spmm/gen/f32-spmm-32x1-minmax-neonfma-pipelined.c",
++    "src/src/f32-spmm/gen/f32-spmm-32x2-minmax-aarch64-neonfma.c",
++    "src/src/f32-spmm/gen/f32-spmm-32x4-minmax-aarch64-neonfma.c",
++    "src/src/f32-spmm/gen/f32-spmm-8x1-minmax-scalar.c",
++    "src/src/f32-spmm/gen/f32-spmm-8x2-minmax-scalar.c",
++    "src/src/f32-spmm/gen/f32-spmm-8x4-minmax-scalar.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -47085,11 +49325,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-dwconv_f16c-fma-avx2_standalone") {
+-  source_set("qs8-qc8w-gemm_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -47100,9 +49339,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p16c-minmax-fp32-avx2-mul32.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-avx2-mul32.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-avx2-mul32.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4c8-minmax-fp32-sse2-ld64.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4c8-minmax-fp32-sse2-ld64.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -47118,7 +49356,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-igemm_arm64_standalone") {
++source_set("f32-spmm_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -47128,52 +49366,38 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set(
--      "qs8-qc8w-dwconv_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
--    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-igemm/gen/f32-igemm-1x4-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x4-relu-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x4-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-aarch64-neonfma-lane-ld128.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-neon-lane-ld64.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x8s4-minmax-neonfma.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-relu-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-aarch64-neonfma-lane-ld64.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-neon-lane-ld64.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x2-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x4-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x4-relu-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x4-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x8-minmax-neon-lane-ld128.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x8s4-minmax-neonfma.c",
-+    "src/src/f32-igemm/gen/f32-igemm-6x8-minmax-aarch64-neonfma-lane-ld128.c",
-+    "src/src/f32-igemm/gen/f32-igemm-6x8s4-minmax-neonfma.c"
++    "src/src/f32-spmm/gen/f32-spmm-32x1-minmax-neon.c",
++    "src/src/f32-spmm/gen/f32-spmm-32x1-minmax-neonfma-pipelined.c",
++    "src/src/f32-spmm/gen/f32-spmm-32x2-minmax-aarch64-neonfma.c",
++    "src/src/f32-spmm/gen/f32-spmm-32x4-minmax-aarch64-neonfma.c",
++    "src/src/f32-spmm/gen/f32-spmm-8x1-minmax-scalar.c",
++    "src/src/f32-spmm/gen/f32-spmm-8x2-minmax-scalar.c",
++    "src/src/f32-spmm/gen/f32-spmm-8x4-minmax-scalar.c"
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p32c-minmax-fp32-avx512skx-mul32.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p32c-minmax-fp32-avx512skx-mul32.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p32c-minmax-fp32-avx512skx-mul32.c",
+-  source_set("qs8-qc8w-gemm_sse4.1-no-sse4.2") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4c8-minmax-fp32-sse41-ld64.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4c8-minmax-fp32-sse41-ld64.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -47184,41 +49408,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +  public_configs = [ ":xnnpack_config" ]
  
--  # This is a target that cannot depend on //base.
--  source_set(
--      "qs8-qc8w-dwconv_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
  
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-qc8w-gemm_sse4.1-no-sse4.2_standalone") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
+-    ]
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p32c-minmax-fp32-avx512skx-mul32.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p32c-minmax-fp32-avx512skx-mul32.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p32c-minmax-fp32-avx512skx-mul32.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4c8-minmax-fp32-sse41-ld64.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4c8-minmax-fp32-sse41-ld64.c",
 -    ]
-+source_set("f32-maxpool_arm64") {
++source_set("f32-vapproxgelu_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -47236,8 +49446,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-maxpool/gen/f32-maxpool-9p-minmax-neon-u4.c",
-+    "src/src/f32-maxpool/gen/f32-maxpool-9p-minmax-scalar-u1.c"
++    "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-neon-rational-12-10-div.c",
++    "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-scalar-rational-12-10-div.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -47248,11 +49458,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qs8-qc8w-dwconv_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("qs8-qc8w-gemm_x64") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -47262,9 +49469,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-sse2-mul16.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p8c-minmax-fp32-sse2-mul16.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p8c-minmax-fp32-sse2-mul16.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4-minmax-fp32-scalar-lrintf.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -47280,7 +49486,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-maxpool_arm64_standalone") {
++source_set("f32-vapproxgelu_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -47288,22 +49494,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-dwconv_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("qs8-qc8w-gemm_x64_standalone") {
+-    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-maxpool/gen/f32-maxpool-9p-minmax-neon-u4.c",
-+    "src/src/f32-maxpool/gen/f32-maxpool-9p-minmax-scalar-u1.c"
++    "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-neon-rational-12-10-div.c",
++    "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-scalar-rational-12-10-div.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-sse2-mul16.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p8c-minmax-fp32-sse2-mul16.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p8c-minmax-fp32-sse2-mul16.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4-minmax-fp32-scalar-lrintf.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -47337,19 +49539,20 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-qc8w-dwconv_sse4.1-no-sse4.2") {
+-  source_set("qs8-qc8w-igemm_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
-+source_set("f32-qc4w-gemm_arm64") {
++source_set("f32-vbinary_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-sse41-mul16.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p8c-minmax-fp32-sse41-mul16.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p8c-minmax-fp32-sse41-mul16.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4c8-minmax-fp32-avx-ld128.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x4c8-minmax-fp32-avx-ld128.c",
 -    ]
 +  ]
  
@@ -47365,13 +49568,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-1x4-minmax-scalar.c",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-1x8-minmax-aarch64-neonfma-lane-ld128.c",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-1x8-minmax-neon-lane-ld64.c",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-4x4-minmax-scalar.c",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-4x8-minmax-aarch64-neonfma-lane-ld128.c",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-4x8-minmax-neon-lane-ld64.c",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-6x8-minmax-aarch64-neonfma-lane-ld128.c"
++    "src/src/f32-vbinary/gen/f32-vadd-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vadd-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vaddc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vaddc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vdiv-aarch64-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vdiv-scalar-u2.c",
++    "src/src/f32-vbinary/gen/f32-vdivc-aarch64-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vdivc-scalar-u2.c",
++    "src/src/f32-vbinary/gen/f32-vmax-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmax-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmaxc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmaxc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmin-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmin-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vminc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vminc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmul-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmul-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmulc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmulc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vprelu-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vprelu-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vpreluc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vpreluc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vrdivc-aarch64-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vrdivc-scalar-u2.c",
++    "src/src/f32-vbinary/gen/f32-vrpreluc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vrpreluc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vrsubc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vrsubc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsqrdiff-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsqrdiff-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsqrdiffc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsqrdiffc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsub-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsub-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsubc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsubc-scalar-u8.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -47381,10 +49615,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-dwconv_sse4.1-no-sse4.2_standalone") {
+-  source_set("qs8-qc8w-igemm_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -47395,9 +49631,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-sse41-mul16.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p8c-minmax-fp32-sse41-mul16.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p8c-minmax-fp32-sse41-mul16.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4c8-minmax-fp32-avx-ld128.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x4c8-minmax-fp32-avx-ld128.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -47413,7 +49648,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-qc4w-gemm_arm64_standalone") {
++source_set("f32-vbinary_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -47423,31 +49658,69 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set("qs8-qc8w-dwconv_x64") {
--    cflags = []
--
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vbinary/gen/f32-vadd-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vadd-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vaddc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vaddc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vdiv-aarch64-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vdiv-scalar-u2.c",
++    "src/src/f32-vbinary/gen/f32-vdivc-aarch64-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vdivc-scalar-u2.c",
++    "src/src/f32-vbinary/gen/f32-vmax-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmax-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmaxc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmaxc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmin-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmin-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vminc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vminc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmul-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmul-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmulc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmulc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vprelu-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vprelu-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vpreluc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vpreluc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vrdivc-aarch64-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vrdivc-scalar-u2.c",
++    "src/src/f32-vbinary/gen/f32-vrpreluc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vrpreluc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vrsubc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vrsubc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsqrdiff-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsqrdiff-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsqrdiffc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsqrdiffc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsub-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsub-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsubc-neon-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsubc-scalar-u8.c"
++  ]
+ 
+-  source_set("qs8-qc8w-igemm_avx2-avxvnni-f16c-fma") {
+-    cflags = [
+-      "-mavx2",
+-      "-mavxvnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-lrintf.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-avxvnni-prfm.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-5x8c8-minmax-fp32-avxvnni-prfm.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-1x4-minmax-scalar.c",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-1x8-minmax-aarch64-neonfma-lane-ld128.c",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-1x8-minmax-neon-lane-ld64.c",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-4x4-minmax-scalar.c",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-4x8-minmax-aarch64-neonfma-lane-ld128.c",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-4x8-minmax-neon-lane-ld64.c",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-6x8-minmax-aarch64-neonfma-lane-ld128.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -47460,45 +49733,30 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
-+
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
    }
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-dwconv_x64_standalone") {
--    cflags = []
+-  source_set("qs8-qc8w-igemm_avx2-avxvnni-f16c-fma_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mavxvnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-lrintf.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-avxvnni-prfm.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-5x8c8-minmax-fp32-avxvnni-prfm.c",
 -    ]
-+source_set("f32-qc8w-gemm_arch=armv8.2-a+fp16+dotprod") {
++source_set("f32-vclamp_arm64") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -47510,14 +49768,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  asmflags = cflags
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x8-minmax-asm-aarch64-neonfma-ld128-acc4.S",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x8-minmax-asm-aarch64-neonfma-ld128.S",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-6x8-minmax-asm-aarch64-neonfma-ld128.S"
++    "src/src/f32-vclamp/gen/f32-vclamp-neon.c",
++    "src/src/f32-vclamp/gen/f32-vclamp-scalar.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -47528,12 +49785,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qs8-qc8w-gemm_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qs8-qc8w-igemm_avx2-avxvnniint8-f16c-fma") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mavx2",
+-      "-mavxvnniint8",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -47544,8 +49801,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4c8-minmax-fp32-avx-ld128.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x4c8-minmax-fp32-avx-ld128.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-avxvnniint8-prfm.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-5x8c8-minmax-fp32-avxvnniint8-prfm.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -47561,34 +49818,31 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-qc8w-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
++source_set("f32-vclamp_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  asmflags = cflags
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-gemm_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("qs8-qc8w-igemm_avx2-avxvnniint8-f16c-fma_standalone") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mavx2",
+-      "-mavxvnniint8",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x8-minmax-asm-aarch64-neonfma-ld128-acc4.S",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x8-minmax-asm-aarch64-neonfma-ld128.S",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-6x8-minmax-asm-aarch64-neonfma-ld128.S"
++    "src/src/f32-vclamp/gen/f32-vclamp-neon.c",
++    "src/src/f32-vclamp/gen/f32-vclamp-scalar.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4c8-minmax-fp32-avx-ld128.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x4c8-minmax-fp32-avx-ld128.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-avxvnniint8-prfm.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-5x8c8-minmax-fp32-avxvnniint8-prfm.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -47622,20 +49876,19 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-qc8w-gemm_avx2-avxvnni-f16c-fma") {
+-  source_set("qs8-qc8w-igemm_f16c-fma-avx2") {
 -    cflags = [
 -      "-mavx2",
--      "-mavxvnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("f32-qc8w-gemm_arm64") {
++source_set("f32-vcmul_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-avxvnni-prfm.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-5x8c8-minmax-fp32-avxvnni-prfm.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-avx2.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x8c8-minmax-fp32-avx2.c",
 -    ]
 +  ]
  
@@ -47651,12 +49904,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x4-minmax-scalar.c",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x8-minmax-aarch64-neonfma-lane-ld64.c",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x8-minmax-neon-lane-ld64.c",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-4x4-minmax-scalar.c",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-4x8-minmax-neon-lane-ld64.c",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-6x8-minmax-aarch64-neonfma-lane-ld64.c"
++    "src/src/f32-vcmul/gen/f32-vcmul-neon-u8.c",
++    "src/src/f32-vcmul/gen/f32-vcmul-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -47666,10 +49915,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-gemm_avx2-avxvnni-f16c-fma_standalone") {
+-  source_set("qs8-qc8w-igemm_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
--      "-mavxvnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
@@ -47682,8 +49930,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-avxvnni-prfm.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-5x8c8-minmax-fp32-avxvnni-prfm.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-avx2.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x8c8-minmax-fp32-avx2.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -47699,7 +49947,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-qc8w-gemm_arm64_standalone") {
++source_set("f32-vcmul_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -47709,38 +49957,34 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x4-minmax-scalar.c",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x8-minmax-aarch64-neonfma-lane-ld64.c",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x8-minmax-neon-lane-ld64.c",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-4x4-minmax-scalar.c",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-4x8-minmax-neon-lane-ld64.c",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-6x8-minmax-aarch64-neonfma-lane-ld64.c"
-+  ]
- 
--  source_set("qs8-qc8w-gemm_avx2-avxvnniint8-f16c-fma") {
+-
+-  source_set(
+-      "qs8-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 -    cflags = [
--      "-mavx2",
--      "-mavxvnniint8",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vcmul/gen/f32-vcmul-neon-u8.c",
++    "src/src/f32-vcmul/gen/f32-vcmul-scalar-u4.c"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-avxvnniint8-prfm.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-5x8c8-minmax-fp32-avxvnniint8-prfm.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-avx256skx.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x8c8-minmax-fp32-avx256skx.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-7x16c8-minmax-fp32-avx512skx-prfm.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -47752,6 +49996,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
++
 +  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -47761,20 +50012,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-gemm_avx2-avxvnniint8-f16c-fma_standalone") {
+-  source_set(
+-      "qs8-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
--      "-mavx2",
--      "-mavxvnniint8",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("f32-qs8-vcvt_arm64") {
++source_set("f32-vcopysign_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-avxvnniint8-prfm.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-5x8c8-minmax-fp32-avxvnniint8-prfm.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-avx256skx.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x8c8-minmax-fp32-avx256skx.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-7x16c8-minmax-fp32-avx512skx-prfm.c",
 -    ]
 +  ]
  
@@ -47790,11 +50047,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-neon-u32.c",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-neonv8-u32.c",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u1.c",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u4.c",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-lrintf-u4.c"
++    "src/src/f32-vcopysign/gen/f32-vcopysign-neon.c",
++    "src/src/f32-vcopysign/gen/f32-vcopysign-scalar.c",
++    "src/src/f32-vcopysign/gen/f32-vcopysignc-neon.c",
++    "src/src/f32-vcopysign/gen/f32-vcopysignc-scalar.c",
++    "src/src/f32-vcopysign/gen/f32-vrcopysignc-neon.c",
++    "src/src/f32-vcopysign/gen/f32-vrcopysignc-scalar.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -47806,6 +50064,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
+-
+-  source_set(
+-      "qs8-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -47813,23 +50084,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  source_set("qs8-qc8w-gemm_f16c-fma-avx2") {
--    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-avx2.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x8c8-minmax-fp32-avx2.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-7x16c8-minmax-avx512vnni-prfm.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-qs8-vcvt_arm64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -47841,41 +50102,45 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  ]
++# This is a target that cannot depend on //base.
++source_set("f32-vcopysign_arm64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-neon-u32.c",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-neonv8-u32.c",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u1.c",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u4.c",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-lrintf-u4.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-gemm_f16c-fma-avx2_standalone") {
+-  source_set(
+-      "qs8-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
 -    cflags = [
--      "-mavx2",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vcopysign/gen/f32-vcopysign-neon.c",
++    "src/src/f32-vcopysign/gen/f32-vcopysign-scalar.c",
++    "src/src/f32-vcopysign/gen/f32-vcopysignc-neon.c",
++    "src/src/f32-vcopysign/gen/f32-vcopysignc-scalar.c",
++    "src/src/f32-vcopysign/gen/f32-vrcopysignc-neon.c",
++    "src/src/f32-vcopysign/gen/f32-vrcopysignc-scalar.c"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-avx2.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x8c8-minmax-fp32-avx2.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-7x16c8-minmax-avx512vnni-prfm.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -47887,13 +50152,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
@@ -47901,27 +50168,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 -
 -  source_set(
--      "qs8-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-      "qs8-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8") {
 -    cflags = [
+-      "-mamx-int8",
+-      "-mamx-tile",
 -      "-mavx512bw",
 -      "-mavx512cd",
 -      "-mavx512dq",
 -      "-mavx512f",
 -      "-mavx512vl",
+-      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
+-      "-mgfni",
 -    ]
-+source_set("f32-qu8-vcvt_arm64") {
-+  cflags = [
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-avx256skx.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x8c8-minmax-fp32-avx256skx.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-7x16c8-minmax-fp32-avx512skx-prfm.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-16x64c4-minmax-fp32-avx512amx.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x64c4-minmax-fp32-avx512amx.c",
 -    ]
-+  ]
++source_set("f32-vcos_arm64") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -47933,32 +50205,40 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-neon-u32.c",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-neonv8-u32.c",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u1.c",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u4.c",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-lrintf-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vcos/gen/f32-vcos-neon-rational-5-4-div.c",
++    "src/src/f32-vcos/gen/f32-vcos-scalar-rational-5-4-div.c"
++  ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
--      "qs8-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-      "qs8-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8_standalone") {
 -    cflags = [
+-      "-mamx-int8",
+-      "-mamx-tile",
 -      "-mavx512bw",
 -      "-mavx512cd",
 -      "-mavx512dq",
 -      "-mavx512f",
 -      "-mavx512vl",
+-      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
+-      "-mgfni",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-16x64c4-minmax-fp32-avx512amx.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x64c4-minmax-fp32-avx512amx.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -47967,16 +50247,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-avx256skx.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x8c8-minmax-fp32-avx256skx.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-7x16c8-minmax-fp32-avx512skx-prfm.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -47987,43 +50257,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-qu8-vcvt_arm64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
-+  ]
++# This is a target that cannot depend on //base.
++source_set("f32-vcos_arm64_standalone") {
++  cflags = [
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set(
--      "qs8-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
++  ]
+ 
+-  source_set("qs8-qc8w-igemm_sse2-no-sse3") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-neon-u32.c",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-neonv8-u32.c",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u1.c",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u4.c",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-lrintf-u4.c"
++    "src/src/f32-vcos/gen/f32-vcos-neon-rational-5-4-div.c",
++    "src/src/f32-vcos/gen/f32-vcos-scalar-rational-5-4-div.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16c8-minmax-fp32-avx512vnni-prfm.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-7x16c8-minmax-fp32-avx512vnni-prfm.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4c8-minmax-fp32-sse2-ld64.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4c8-minmax-fp32-sse2-ld64.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -48047,35 +50309,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
+-  }
 +  public_configs = [ ":xnnpack_config" ]
-+
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qs8-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
+-  source_set("qs8-qc8w-igemm_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+source_set("f32-raddstoreexpminusmax_arm64") {
-+  cflags = [
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16c8-minmax-fp32-avx512vnni-prfm.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-7x16c8-minmax-fp32-avx512vnni-prfm.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4c8-minmax-fp32-sse2-ld64.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4c8-minmax-fp32-sse2-ld64.c",
 -    ]
-+  ]
++source_set("f32-velu_arm64") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -48087,36 +50341,29 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-neonfma-rr1-lut64-p2-u16-acc2.c",
-+    "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-scalar-rr2-p5-u4-acc2.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-velu/gen/f32-velu-neon-rr2-lut16-p3-u8.c",
++    "src/src/f32-velu/gen/f32-velu-neonfma-rr1-lut16-p3-u16.c",
++    "src/src/f32-velu/gen/f32-velu-neonfma-rr1-p6-u8.c",
++    "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u4.c"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set(
--      "qs8-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8") {
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  source_set("qs8-qc8w-igemm_sse4.1-no-sse4.2") {
 -    cflags = [
--      "-mamx-int8",
--      "-mamx-tile",
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
--      "-mf16c",
--      "-mfma",
--      "-mgfni",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -48127,8 +50374,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-16x64c4-minmax-fp32-avx512amx.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x64c4-minmax-fp32-avx512amx.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4c8-minmax-fp32-sse41-ld64.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4c8-minmax-fp32-sse41-ld64.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -48144,7 +50391,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-raddstoreexpminusmax_arm64_standalone") {
++source_set("f32-velu_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -48152,31 +50399,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qs8-qc8w-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8_standalone") {
+-  source_set("qs8-qc8w-igemm_sse4.1-no-sse4.2_standalone") {
 -    cflags = [
--      "-mamx-int8",
--      "-mamx-tile",
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
--      "-mf16c",
--      "-mfma",
--      "-mgfni",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-neonfma-rr1-lut64-p2-u16-acc2.c",
-+    "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-scalar-rr2-p5-u4-acc2.c"
++    "src/src/f32-velu/gen/f32-velu-neon-rr2-lut16-p3-u8.c",
++    "src/src/f32-velu/gen/f32-velu-neonfma-rr1-lut16-p3-u16.c",
++    "src/src/f32-velu/gen/f32-velu-neonfma-rr1-p6-u8.c",
++    "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u4.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-16x64c4-minmax-fp32-avx512amx.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x64c4-minmax-fp32-avx512amx.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4c8-minmax-fp32-sse41-ld64.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4c8-minmax-fp32-sse41-ld64.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -48210,18 +50449,15 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-qc8w-gemm_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
-+source_set("f32-rdminmax_arm64") {
+-  source_set("qs8-qc8w-igemm_x64") {
+-    cflags = []
++source_set("f32-vexp_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4c8-minmax-fp32-sse2-ld64.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4c8-minmax-fp32-sse2-ld64.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4-minmax-fp32-scalar-lrintf.c",
 -    ]
 +  ]
  
@@ -48237,10 +50473,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-neon-c32.c",
-+    "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-scalar-c2.c",
-+    "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-neon-c32.c",
-+    "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-scalar-c2.c"
++    "src/src/f32-vexp/gen/f32-vexp-neon-rational-3-2-div.c",
++    "src/src/f32-vexp/gen/f32-vexp-scalar-rational-3-2-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -48250,11 +50484,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-gemm_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("qs8-qc8w-igemm_x64_standalone") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -48264,8 +50495,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4c8-minmax-fp32-sse2-ld64.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4c8-minmax-fp32-sse2-ld64.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4-minmax-fp32-scalar-lrintf.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -48281,7 +50512,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-rdminmax_arm64_standalone") {
++source_set("f32-vexp_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -48293,16 +50524,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-neon-c32.c",
-+    "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-scalar-c2.c",
-+    "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-neon-c32.c",
-+    "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-scalar-c2.c"
++    "src/src/f32-vexp/gen/f32-vexp-neon-rational-3-2-div.c",
++    "src/src/f32-vexp/gen/f32-vexp-scalar-rational-3-2-div.c"
 +  ]
  
--  source_set("qs8-qc8w-gemm_sse4.1-no-sse4.2") {
+-  source_set("qs8-qu8-packw_avx2-avxvnni-f16c-fma") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx2",
+-      "-mavxvnni",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -48310,8 +50541,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4c8-minmax-fp32-sse41-ld64.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4c8-minmax-fp32-sse41-ld64.c",
+-      "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x8c8-gemm-goi-avxvnni.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -48339,18 +50569,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-gemm_sse4.1-no-sse4.2_standalone") {
+-  source_set("qs8-qu8-packw_avx2-avxvnni-f16c-fma_standalone") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx2",
+-      "-mavxvnni",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+source_set("f32-rdsum_arm64") {
++source_set("f32-vgelu_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4c8-minmax-fp32-sse41-ld64.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4c8-minmax-fp32-sse41-ld64.c",
+-      "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x8c8-gemm-goi-avxvnni.c",
 -    ]
 +  ]
  
@@ -48366,8 +50597,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-neon-c16.c",
-+    "src/src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-scalar.c"
++    "src/src/f32-vgelu/gen/f32-vgelu-neon-rational-12-10-div.c",
++    "src/src/f32-vgelu/gen/f32-vgelu-scalar-rational-12-10-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -48380,15 +50611,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    }
 -  }
 -
--  source_set("qs8-qc8w-gemm_x64") {
--    cflags = []
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x2-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4-minmax-fp32-scalar-lrintf.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x2-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4-minmax-fp32-scalar-lrintf.c",
+-  source_set(
+-      "qs8-qu8-packw_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -48397,6 +50630,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-avx256vnni.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -48407,27 +50647,40 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++# This is a target that cannot depend on //base.
++source_set("f32-vgelu_arm64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+# This is a target that cannot depend on //base.
-+source_set("f32-rdsum_arm64_standalone") {
-+  cflags = [
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-gemm_x64_standalone") {
--    cflags = []
--
+-  source_set(
+-      "qs8-qu8-packw_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vgelu/gen/f32-vgelu-neon-rational-12-10-div.c",
++    "src/src/f32-vgelu/gen/f32-vgelu-scalar-rational-12-10-div.c"
++  ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x2-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4-minmax-fp32-scalar-lrintf.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x2-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-avx256vnni.c",
 -    ]
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -48439,16 +50692,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-neon-c16.c",
-+    "src/src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-scalar.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -48467,20 +50710,14 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-qc8w-igemm_avx-no-avx2-no-f16c-no-fma") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
-+source_set("f32-rminmax_arm64") {
+-  source_set("qs8-qu8-packw_x64") {
+-    cflags = []
++source_set("f32-vhswish_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4c8-minmax-fp32-avx-ld128.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x4c8-minmax-fp32-avx-ld128.c",
+-      "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-scalar.c",
 -    ]
 +  ]
  
@@ -48496,12 +50733,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rminmax/gen/f32-rmax-neon-u16-acc4.c",
-+    "src/src/f32-rminmax/gen/f32-rmax-scalar-u4-acc4.c",
-+    "src/src/f32-rminmax/gen/f32-rmin-neon-u16-acc4.c",
-+    "src/src/f32-rminmax/gen/f32-rmin-scalar-u4-acc4.c",
-+    "src/src/f32-rminmax/gen/f32-rminmax-neon-u16-acc4.c",
-+    "src/src/f32-rminmax/gen/f32-rminmax-scalar-u4-acc4.c"
++    "src/src/f32-vhswish/gen/f32-vhswish-neon.c",
++    "src/src/f32-vhswish/gen/f32-vhswish-scalar.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -48511,13 +50744,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-igemm_avx-no-avx2-no-f16c-no-fma_standalone") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
+-  source_set("qs8-qu8-packw_x64_standalone") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -48527,8 +50755,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4c8-minmax-fp32-avx-ld128.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x4c8-minmax-fp32-avx-ld128.c",
+-      "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-scalar.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -48544,7 +50771,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-rminmax_arm64_standalone") {
++source_set("f32-vhswish_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -48556,18 +50783,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rminmax/gen/f32-rmax-neon-u16-acc4.c",
-+    "src/src/f32-rminmax/gen/f32-rmax-scalar-u4-acc4.c",
-+    "src/src/f32-rminmax/gen/f32-rmin-neon-u16-acc4.c",
-+    "src/src/f32-rminmax/gen/f32-rmin-scalar-u4-acc4.c",
-+    "src/src/f32-rminmax/gen/f32-rminmax-neon-u16-acc4.c",
-+    "src/src/f32-rminmax/gen/f32-rminmax-scalar-u4-acc4.c"
++    "src/src/f32-vhswish/gen/f32-vhswish-neon.c",
++    "src/src/f32-vhswish/gen/f32-vhswish-scalar.c"
 +  ]
  
--  source_set("qs8-qc8w-igemm_avx2-avxvnni-f16c-fma") {
+-  source_set("qs8-rdsum_f16c-fma-avx2") {
 -    cflags = [
 -      "-mavx2",
--      "-mavxvnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
@@ -48577,8 +50799,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-avxvnni-prfm.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-5x8c8-minmax-fp32-avxvnni-prfm.c",
+-      "src/src/qs8-rdsum/gen/qs8-rdsum-7p7x-minmax-fp32-avx2-c64.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -48606,21 +50827,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-igemm_avx2-avxvnni-f16c-fma_standalone") {
+-  source_set("qs8-rdsum_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
--      "-mavxvnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
--
++source_set("f32-vlog_arm64") {
++  cflags = [
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-avxvnni-prfm.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-5x8c8-minmax-fp32-avxvnni-prfm.c",
+-      "src/src/qs8-rdsum/gen/qs8-rdsum-7p7x-minmax-fp32-avx2-c64.c",
 -    ]
-+source_set("f32-rsum_arm64") {
-+  cflags = [
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -48632,27 +50852,29 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rsum/gen/f32-rsum-neon-u16-acc4.c",
-+    "src/src/f32-rsum/gen/f32-rsum-scalar-u4-acc4.c"
++    "src/src/f32-vlog/gen/f32-vlog-neon-rational-3-3-div.c",
++    "src/src/f32-vlog/gen/f32-vlog-scalar-rational-3-3-div.c"
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qs8-qc8w-igemm_avx2-avxvnniint8-f16c-fma") {
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
+-
+-  source_set("qs8-rdsum_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 -    cflags = [
--      "-mavx2",
--      "-mavxvnniint8",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
 -      "-mf16c",
 -      "-mfma",
 -    ]
@@ -48665,8 +50887,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-avxvnniint8-prfm.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-5x8c8-minmax-fp32-avxvnniint8-prfm.c",
+-      "src/src/qs8-rdsum/gen/qs8-rdsum-7p7x-minmax-fp32-avx512skx-c64.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -48682,7 +50903,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-rsum_arm64_standalone") {
++source_set("f32-vlog_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -48690,23 +50911,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-igemm_avx2-avxvnniint8-f16c-fma_standalone") {
+-  source_set(
+-      "qs8-rdsum_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
--      "-mavx2",
--      "-mavxvnniint8",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
 -      "-mf16c",
 -      "-mfma",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rsum/gen/f32-rsum-neon-u16-acc4.c",
-+    "src/src/f32-rsum/gen/f32-rsum-scalar-u4-acc4.c"
++    "src/src/f32-vlog/gen/f32-vlog-neon-rational-3-3-div.c",
++    "src/src/f32-vlog/gen/f32-vlog-scalar-rational-3-3-div.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-avxvnniint8-prfm.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-5x8c8-minmax-fp32-avxvnniint8-prfm.c",
+-      "src/src/qs8-rdsum/gen/qs8-rdsum-7p7x-minmax-fp32-avx512skx-c64.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -48740,19 +50964,17 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-qc8w-igemm_f16c-fma-avx2") {
+-  source_set("qs8-rdsum_sse4.1-no-sse4.2") {
 -    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
-+source_set("f32-spmm_arm64") {
++source_set("f32-vlrelu_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-avx2.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x8c8-minmax-fp32-avx2.c",
+-      "src/src/qs8-rdsum/gen/qs8-rdsum-7p7x-minmax-fp32-sse41-c64.c",
 -    ]
 +  ]
  
@@ -48768,13 +50990,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-spmm/gen/f32-spmm-32x1-minmax-neon.c",
-+    "src/src/f32-spmm/gen/f32-spmm-32x1-minmax-neonfma-pipelined.c",
-+    "src/src/f32-spmm/gen/f32-spmm-32x2-minmax-aarch64-neonfma.c",
-+    "src/src/f32-spmm/gen/f32-spmm-32x4-minmax-aarch64-neonfma.c",
-+    "src/src/f32-spmm/gen/f32-spmm-8x1-minmax-scalar.c",
-+    "src/src/f32-spmm/gen/f32-spmm-8x2-minmax-scalar.c",
-+    "src/src/f32-spmm/gen/f32-spmm-8x4-minmax-scalar.c"
++    "src/src/f32-vlrelu/gen/f32-vlrelu-neon-u8.c",
++    "src/src/f32-vlrelu/gen/f32-vlrelu-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -48784,11 +51001,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-igemm_f16c-fma-avx2_standalone") {
+-  source_set("qs8-rdsum_sse4.1-no-sse4.2_standalone") {
 -    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -48799,8 +51015,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-avx2.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x8c8-minmax-fp32-avx2.c",
+-      "src/src/qs8-rdsum/gen/qs8-rdsum-7p7x-minmax-fp32-sse41-c64.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -48816,7 +51031,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-spmm_arm64_standalone") {
++source_set("f32-vlrelu_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -48826,40 +51041,29 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set(
--      "qs8-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
--    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-spmm/gen/f32-spmm-32x1-minmax-neon.c",
-+    "src/src/f32-spmm/gen/f32-spmm-32x1-minmax-neonfma-pipelined.c",
-+    "src/src/f32-spmm/gen/f32-spmm-32x2-minmax-aarch64-neonfma.c",
-+    "src/src/f32-spmm/gen/f32-spmm-32x4-minmax-aarch64-neonfma.c",
-+    "src/src/f32-spmm/gen/f32-spmm-8x1-minmax-scalar.c",
-+    "src/src/f32-spmm/gen/f32-spmm-8x2-minmax-scalar.c",
-+    "src/src/f32-spmm/gen/f32-spmm-8x4-minmax-scalar.c"
++    "src/src/f32-vlrelu/gen/f32-vlrelu-neon-u8.c",
++    "src/src/f32-vlrelu/gen/f32-vlrelu-scalar-u4.c"
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-avx256skx.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x8c8-minmax-fp32-avx256skx.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-7x16c8-minmax-fp32-avx512skx-prfm.c",
--    ]
+-  source_set("qs8-rdsum_x64") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-rdsum/gen/qs8-rdsum-minmax-fp32-scalar-u1-acc1.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -48870,13 +51074,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
 +  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -48886,28 +51083,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qs8-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
--    ]
-+source_set("f32-vapproxgelu_arm64") {
-+  cflags = [
- 
+-  source_set("qs8-rdsum_x64_standalone") {
+-    cflags = []
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-avx256skx.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x8c8-minmax-fp32-avx256skx.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-7x16c8-minmax-fp32-avx512skx-prfm.c",
+-      "src/src/qs8-rdsum/gen/qs8-rdsum-minmax-fp32-scalar-u1-acc1.c",
 -    ]
-+  ]
++source_set("f32-vmulcaddc_arm64") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -48919,31 +51103,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-neon-rational-12-10-div.c",
-+    "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-scalar-rational-12-10-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c1-minmax-scalar-2x.c",
++    "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c4-minmax-neon-2x.c",
++    "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c4-minmax-neonfma-2x.c"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set(
--      "qs8-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  source_set("qs8-rsum_avx2-avxvnni-f16c-fma") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
+-      "-mavx2",
+-      "-mavxvnni",
 -      "-mf16c",
 -      "-mfma",
 -    ]
@@ -48956,8 +51137,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16c8-minmax-avx512vnni-prfm.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-7x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qs8-rsum/gen/qs8-rsum-avxvnni-u128-acc2.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -48973,57 +51153,109 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vapproxgelu_arm64_standalone") {
++source_set("f32-vmulcaddc_arm64_standalone") {
 +  cflags = [
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  ]
-+
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-rsum_avx2-avxvnni-f16c-fma_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mavxvnni",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-neon-rational-12-10-div.c",
-+    "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-scalar-rational-12-10-div.c"
++    "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c1-minmax-scalar-2x.c",
++    "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c4-minmax-neon-2x.c",
++    "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c4-minmax-neonfma-2x.c"
 +  ]
-+
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-rsum/gen/qs8-rsum-avxvnni-u128-acc2.c",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
    }
 +}
  
--  # This is a target that cannot depend on //base.
--  source_set(
--      "qs8-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
+-  source_set("qs8-rsum_f16c-fma-avx2") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
+-      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("f32-vbinary_arm64") {
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-rsum/gen/qs8-rsum-avx2-u64-acc2.c",
+-    ]
+-
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
+-
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++source_set("f32-vrelu_arm64") {
 +  cflags = [
  
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-rsum_f16c-fma-avx2_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++  ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16c8-minmax-avx512vnni-prfm.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-7x16c8-minmax-avx512vnni-prfm.c",
+-      "src/src/qs8-rsum/gen/qs8-rsum-avx2-u64-acc2.c",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vrelu/gen/f32-vrelu-neon.c",
++    "src/src/f32-vrelu/gen/f32-vrelu-scalar.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -49036,87 +51268,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vbinary/gen/f32-vadd-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vadd-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vaddc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vaddc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vdiv-aarch64-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vdiv-scalar-u2.c",
-+    "src/src/f32-vbinary/gen/f32-vdivc-aarch64-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vdivc-scalar-u2.c",
-+    "src/src/f32-vbinary/gen/f32-vmax-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmax-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmaxc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmaxc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmin-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmin-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vminc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vminc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmul-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmul-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmulc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmulc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vprelu-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vprelu-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vpreluc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vpreluc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vrdivc-aarch64-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vrdivc-scalar-u2.c",
-+    "src/src/f32-vbinary/gen/f32-vrpreluc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vrpreluc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vrsubc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vrsubc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsqrdiff-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsqrdiff-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsqrdiffc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsqrdiffc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsub-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsub-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsubc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsubc-scalar-u8.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
+-    public_configs = [ ":xnnpack_config" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
 -
--  source_set(
--      "qs8-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8") {
+-  source_set("qs8-rsum_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 -    cflags = [
--      "-mamx-int8",
--      "-mamx-tile",
 -      "-mavx512bw",
 -      "-mavx512cd",
 -      "-mavx512dq",
 -      "-mavx512f",
 -      "-mavx512vl",
--      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
--      "-mgfni",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-16x64c4-minmax-fp32-avx512amx.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x64c4-minmax-fp32-avx512amx.c",
+-      "src/src/qs8-rsum/gen/qs8-rsum-avx256skx-u64-acc2.c",
+-      "src/src/qs8-rsum/gen/qs8-rsum-avx512skx-u128-acc2.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++# This is a target that cannot depend on //base.
++source_set("f32-vrelu_arm64_standalone") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -49128,80 +51317,43 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-vbinary_arm64_standalone") {
-+  cflags = [
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vrelu/gen/f32-vrelu-neon.c",
++    "src/src/f32-vrelu/gen/f32-vrelu-scalar.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
--      "qs8-qc8w-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni-gfni-amx-tile-amx-int8_standalone") {
+-      "qs8-rsum_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
--      "-mamx-int8",
--      "-mamx-tile",
 -      "-mavx512bw",
 -      "-mavx512cd",
 -      "-mavx512dq",
 -      "-mavx512f",
 -      "-mavx512vl",
--      "-mavx512vnni",
 -      "-mf16c",
 -      "-mfma",
--      "-mgfni",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vbinary/gen/f32-vadd-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vadd-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vaddc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vaddc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vdiv-aarch64-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vdiv-scalar-u2.c",
-+    "src/src/f32-vbinary/gen/f32-vdivc-aarch64-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vdivc-scalar-u2.c",
-+    "src/src/f32-vbinary/gen/f32-vmax-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmax-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmaxc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmaxc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmin-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmin-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vminc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vminc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmul-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmul-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmulc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmulc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vprelu-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vprelu-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vpreluc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vpreluc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vrdivc-aarch64-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vrdivc-scalar-u2.c",
-+    "src/src/f32-vbinary/gen/f32-vrpreluc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vrpreluc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vrsubc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vrsubc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsqrdiff-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsqrdiff-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsqrdiffc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsqrdiffc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsub-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsub-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsubc-neon-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsubc-scalar-u8.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-16x64c4-minmax-fp32-avx512amx.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x64c4-minmax-fp32-avx512amx.c",
+-      "src/src/qs8-rsum/gen/qs8-rsum-avx256skx-u64-acc2.c",
+-      "src/src/qs8-rsum/gen/qs8-rsum-avx512skx-u128-acc2.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -49213,36 +51365,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-   }
++  }
 +}
  
--  source_set("qs8-qc8w-igemm_sse2-no-sse3") {
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
+-
+-  source_set(
+-      "qs8-rsum_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+source_set("f32-vclamp_arm64") {
++source_set("f32-vrnd_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4c8-minmax-fp32-sse2-ld64.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4c8-minmax-fp32-sse2-ld64.c",
+-      "src/src/qs8-rsum/gen/qs8-rsum-avx512vnni-u128-acc2.c",
 -    ]
 +  ]
  
@@ -49258,8 +51411,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vclamp/gen/f32-vclamp-neon-u16.c",
-+    "src/src/f32-vclamp/gen/f32-vclamp-scalar-u4.c"
++    "src/src/f32-vrnd/gen/f32-vrndd-neon-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndd-neonv8-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-neon-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-neonv8-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-neon-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-neonv8-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-neon-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-neonv8-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -49269,10 +51432,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-igemm_sse2-no-sse3_standalone") {
+-  source_set(
+-      "qs8-rsum_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mavx512vnni",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -49283,8 +51453,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4c8-minmax-fp32-sse2-ld64.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4c8-minmax-fp32-sse2-ld64.c",
+-      "src/src/qs8-rsum/gen/qs8-rsum-avx512vnni-u128-acc2.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -49300,7 +51469,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vclamp_arm64_standalone") {
++source_set("f32-vrnd_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -49312,14 +51481,24 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vclamp/gen/f32-vclamp-neon-u16.c",
-+    "src/src/f32-vclamp/gen/f32-vclamp-scalar-u4.c"
++    "src/src/f32-vrnd/gen/f32-vrndd-neon-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndd-neonv8-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-neon-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-neonv8-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-neon-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-neonv8-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-neon-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-neonv8-u8.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c"
 +  ]
  
--  source_set("qs8-qc8w-igemm_sse4.1-no-sse4.2") {
+-  source_set("qs8-rsum_ssse3-no-sse4.1") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mno-sse4.1",
+-      "-mssse3",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -49327,8 +51506,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4c8-minmax-fp32-sse41-ld64.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4c8-minmax-fp32-sse41-ld64.c",
+-      "src/src/qs8-rsum/gen/qs8-rsum-ssse3-u32-acc2.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -49356,20 +51534,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-igemm_sse4.1-no-sse4.2_standalone") {
+-  source_set("qs8-rsum_ssse3-no-sse4.1_standalone") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mno-sse4.1",
+-      "-mssse3",
 -    ]
-+source_set("f32-vcmul_arm64") {
-+  cflags = [
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4c8-minmax-fp32-sse41-ld64.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4c8-minmax-fp32-sse41-ld64.c",
+-      "src/src/qs8-rsum/gen/qs8-rsum-ssse3-u32-acc2.c",
 -    ]
-+  ]
++source_set("f32-vrsqrt_arm64") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -49381,32 +51557,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vcmul/gen/f32-vcmul-neon-u8.c",
-+    "src/src/f32-vcmul/gen/f32-vcmul-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-neon-rsqrt.c",
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-sqrt.c"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set("qs8-qc8w-igemm_x64") {
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  source_set("qs8-rsum_x64") {
 -    cflags = []
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x2-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4-minmax-fp32-scalar-lrintf.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x2-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4-minmax-fp32-scalar-lrintf.c",
--    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -49414,6 +51584,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-rsum/gen/qs8-rsum-scalar-u4.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -49424,27 +51601,31 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++# This is a target that cannot depend on //base.
++source_set("f32-vrsqrt_arm64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+# This is a target that cannot depend on //base.
-+source_set("f32-vcmul_arm64_standalone") {
-+  cflags = [
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qc8w-igemm_x64_standalone") {
+-  source_set("qs8-rsum_x64_standalone") {
 -    cflags = []
--
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-neon-rsqrt.c",
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-sqrt.c"
++  ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x2-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4-minmax-fp32-scalar-lrintf.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x2-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-rsum/gen/qs8-rsum-scalar-u4.c",
 -    ]
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -49456,24 +51637,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vcmul/gen/f32-vcmul-neon-u8.c",
-+    "src/src/f32-vcmul/gen/f32-vcmul-scalar-u4.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -49484,19 +51655,19 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-qu8-packw_avx2-avxvnni-f16c-fma") {
+-  source_set("qs8-vadd_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
--      "-mavx2",
--      "-mavxvnni",
--      "-mf16c",
--      "-mfma",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
-+source_set("f32-vcopysign_arm64") {
++source_set("f32-vsigmoid_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x8c8-gemm-goi-avxvnni.c",
+-      "src/src/qs8-vadd/gen/qs8-vadd-minmax-avx-mul32-ld32-u8.c",
 -    ]
 +  ]
  
@@ -49512,12 +51683,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vcopysign/gen/f32-vcopysign-neon.c",
-+    "src/src/f32-vcopysign/gen/f32-vcopysign-scalar.c",
-+    "src/src/f32-vcopysign/gen/f32-vcopysignc-neon.c",
-+    "src/src/f32-vcopysign/gen/f32-vcopysignc-scalar.c",
-+    "src/src/f32-vcopysign/gen/f32-vrcopysignc-neon.c",
-+    "src/src/f32-vcopysign/gen/f32-vrcopysignc-scalar.c"
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-neon-rr2-lut64-p2-nr2recps-u8.c",
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-neonfma-rr1-lut64-p2-nr2recps-u16.c",
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -49527,12 +51695,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qu8-packw_avx2-avxvnni-f16c-fma_standalone") {
+-  source_set("qs8-vadd_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
--      "-mavx2",
--      "-mavxvnni",
--      "-mf16c",
--      "-mfma",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -49543,7 +51711,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x8c8-gemm-goi-avxvnni.c",
+-      "src/src/qs8-vadd/gen/qs8-vadd-minmax-avx-mul32-ld32-u8.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -49559,7 +51727,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vcopysign_arm64_standalone") {
++source_set("f32-vsigmoid_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -49569,141 +51737,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set(
--      "qs8-qu8-packw_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
--      "-mf16c",
--      "-mfma",
--    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vcopysign/gen/f32-vcopysign-neon.c",
-+    "src/src/f32-vcopysign/gen/f32-vcopysign-scalar.c",
-+    "src/src/f32-vcopysign/gen/f32-vcopysignc-neon.c",
-+    "src/src/f32-vcopysign/gen/f32-vcopysignc-scalar.c",
-+    "src/src/f32-vcopysign/gen/f32-vrcopysignc-neon.c",
-+    "src/src/f32-vcopysign/gen/f32-vrcopysignc-scalar.c"
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-avx256vnni.c",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-neon-rr2-lut64-p2-nr2recps-u8.c",
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-neonfma-rr1-lut64-p2-nr2recps-u16.c",
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c"
 +  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
  
--  # This is a target that cannot depend on //base.
--  source_set(
--      "qs8-qu8-packw_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
+-  source_set("qs8-vadd_f16c-fma-avx2") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
+-      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("f32-vcos_arm64") {
-+  cflags = [
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-avx256vnni.c",
--    ]
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vcos/gen/f32-vcos-neon-rational-5-4-div.c",
-+    "src/src/f32-vcos/gen/f32-vcos-scalar-rational-5-4-div.c"
-+  ]
-+
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
-+  public_configs = [ ":xnnpack_config" ]
-+}
-+
-+# This is a target that cannot depend on //base.
-+source_set("f32-vcos_arm64_standalone") {
-+  cflags = [
- 
--  source_set("qs8-qu8-packw_x64") {
--    cflags = []
-+  ]
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-scalar.c",
+-      "src/src/qs8-vadd/gen/qs8-vadd-minmax-avx2-mul32-ld64-u16.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vcos/gen/f32-vcos-neon-rational-5-4-div.c",
-+    "src/src/f32-vcos/gen/f32-vcos-scalar-rational-5-4-div.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -49730,14 +51784,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-qu8-packw_x64_standalone") {
--    cflags = []
-+source_set("f32-velu_arm64") {
+-  source_set("qs8-vadd_f16c-fma-avx2_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++source_set("f32-vsin_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-scalar.c",
+-      "src/src/qs8-vadd/gen/qs8-vadd-minmax-avx2-mul32-ld64-u16.c",
 -    ]
 +  ]
  
@@ -49753,11 +51811,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-velu/gen/f32-velu-neon-rr2-lut16-p3-u8.c",
-+    "src/src/f32-velu/gen/f32-velu-neonfma-rr1-lut16-p3-u16.c",
-+    "src/src/f32-velu/gen/f32-velu-neonfma-rr1-p6-u8.c",
-+    "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u2.c",
-+    "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u4.c"
++    "src/src/f32-vsin/gen/f32-vsin-neon-rational-5-4-div.c",
++    "src/src/f32-vsin/gen/f32-vsin-scalar-rational-5-4-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -49769,6 +51824,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
+-
+-  source_set("qs8-vadd_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -49776,22 +51842,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  source_set("qs8-rdsum_f16c-fma-avx2") {
--    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rdsum/gen/qs8-rdsum-7p7x-minmax-fp32-avx2-c64.c",
+-      "src/src/qs8-vadd/gen/qs8-vadd-minmax-avx512skx-mul32-ld128-u16.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-velu_arm64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -49803,40 +51859,39 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  ]
++# This is a target that cannot depend on //base.
++source_set("f32-vsin_arm64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-velu/gen/f32-velu-neon-rr2-lut16-p3-u8.c",
-+    "src/src/f32-velu/gen/f32-velu-neonfma-rr1-lut16-p3-u16.c",
-+    "src/src/f32-velu/gen/f32-velu-neonfma-rr1-p6-u8.c",
-+    "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u2.c",
-+    "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u4.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-rdsum_f16c-fma-avx2_standalone") {
+-  source_set(
+-      "qs8-vadd_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
--      "-mavx2",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vsin/gen/f32-vsin-neon-rational-5-4-div.c",
++    "src/src/f32-vsin/gen/f32-vsin-scalar-rational-5-4-div.c"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rdsum/gen/qs8-rdsum-7p7x-minmax-fp32-avx2-c64.c",
+-      "src/src/qs8-vadd/gen/qs8-vadd-minmax-avx512skx-mul32-ld128-u16.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -49848,35 +51903,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
--
--  source_set("qs8-rdsum_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  source_set("qs8-vadd_sse2-no-sse3") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+source_set("f32-vexp_arm64") {
++source_set("f32-vsqrt_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rdsum/gen/qs8-rdsum-7p7x-minmax-fp32-avx512skx-c64.c",
+-      "src/src/qs8-vadd/gen/qs8-vadd-minmax-sse2-mul16-ld64-u8.c",
 -    ]
 +  ]
  
@@ -49892,8 +51947,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vexp/gen/f32-vexp-neon-rational-3-2-div.c",
-+    "src/src/f32-vexp/gen/f32-vexp-scalar-rational-3-2-div.c"
++    "src/src/f32-vsqrt/gen/f32-vsqrt-aarch64-neon-sqrt.c",
++    "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -49903,16 +51958,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qs8-rdsum_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-  source_set("qs8-vadd_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -49923,7 +51972,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rdsum/gen/qs8-rdsum-7p7x-minmax-fp32-avx512skx-c64.c",
+-      "src/src/qs8-vadd/gen/qs8-vadd-minmax-sse2-mul16-ld64-u8.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -49939,7 +51988,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vexp_arm64_standalone") {
++source_set("f32-vsqrt_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -49951,11 +52000,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vexp/gen/f32-vexp-neon-rational-3-2-div.c",
-+    "src/src/f32-vexp/gen/f32-vexp-scalar-rational-3-2-div.c"
++    "src/src/f32-vsqrt/gen/f32-vsqrt-aarch64-neon-sqrt.c",
++    "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt.c"
 +  ]
  
--  source_set("qs8-rdsum_sse4.1-no-sse4.2") {
+-  source_set("qs8-vadd_sse4.1-no-sse4.2") {
 -    cflags = [
 -      "-mno-sse4.2",
 -      "-msse4.1",
@@ -49966,7 +52015,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rdsum/gen/qs8-rdsum-7p7x-minmax-fp32-sse41-c64.c",
+-      "src/src/qs8-vadd/gen/qs8-vadd-minmax-sse41-mul16-ld64-u8.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -49994,7 +52043,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-rdsum_sse4.1-no-sse4.2_standalone") {
+-  source_set("qs8-vadd_sse4.1-no-sse4.2_standalone") {
 -    cflags = [
 -      "-mno-sse4.2",
 -      "-msse4.1",
@@ -50002,9 +52051,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rdsum/gen/qs8-rdsum-7p7x-minmax-fp32-sse41-c64.c",
+-      "src/src/qs8-vadd/gen/qs8-vadd-minmax-sse41-mul16-ld64-u8.c",
 -    ]
-+source_set("f32-vgelu_arm64") {
++source_set("f32-vtanh_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -50022,8 +52071,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vgelu/gen/f32-vgelu-neon-rational-12-10-div.c",
-+    "src/src/f32-vgelu/gen/f32-vgelu-scalar-rational-12-10-div.c"
++    "src/src/f32-vtanh/gen/f32-vtanh-neon-rational-9-8-div.c",
++    "src/src/f32-vtanh/gen/f32-vtanh-scalar-rational-9-8-div.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -50034,7 +52083,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qs8-rdsum_x64") {
+-  source_set("qs8-vadd_x64") {
 -    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -50045,7 +52094,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rdsum/gen/qs8-rdsum-minmax-fp32-scalar-u1-acc1.c",
+-      "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u1.c",
+-      "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u4.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -50061,7 +52111,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vgelu_arm64_standalone") {
++source_set("f32-vtanh_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -50069,17 +52119,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-rdsum_x64_standalone") {
+-  source_set("qs8-vadd_x64_standalone") {
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vgelu/gen/f32-vgelu-neon-rational-12-10-div.c",
-+    "src/src/f32-vgelu/gen/f32-vgelu-scalar-rational-12-10-div.c"
++    "src/src/f32-vtanh/gen/f32-vtanh-neon-rational-9-8-div.c",
++    "src/src/f32-vtanh/gen/f32-vtanh-scalar-rational-9-8-div.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rdsum/gen/qs8-rdsum-minmax-fp32-scalar-u1-acc1.c",
+-      "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u1.c",
+-      "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u4.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -50113,19 +52164,19 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-rsum_avx2-avxvnni-f16c-fma") {
+-  source_set("qs8-vaddc_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
--      "-mavx2",
--      "-mavxvnni",
--      "-mf16c",
--      "-mfma",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
-+source_set("f32-vhswish_arm64") {
++source_set("f32-vunary_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rsum/gen/qs8-rsum-avxvnni-u128-acc2.c",
+-      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-avx-mul32-ld32-u8.c",
 -    ]
 +  ]
  
@@ -50141,8 +52192,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vhswish/gen/f32-vhswish-neon-u16.c",
-+    "src/src/f32-vhswish/gen/f32-vhswish-scalar-u4.c"
++    "src/src/f32-vunary/gen/f32-vabs-neon.c",
++    "src/src/f32-vunary/gen/f32-vabs-scalar.c",
++    "src/src/f32-vunary/gen/f32-vneg-neon.c",
++    "src/src/f32-vunary/gen/f32-vneg-scalar.c",
++    "src/src/f32-vunary/gen/f32-vsqr-neon.c",
++    "src/src/f32-vunary/gen/f32-vsqr-scalar.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -50152,12 +52207,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-rsum_avx2-avxvnni-f16c-fma_standalone") {
+-  source_set("qs8-vaddc_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
--      "-mavx2",
--      "-mavxvnni",
--      "-mf16c",
--      "-mfma",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -50168,7 +52223,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rsum/gen/qs8-rsum-avxvnni-u128-acc2.c",
+-      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-avx-mul32-ld32-u8.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -50184,7 +52239,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vhswish_arm64_standalone") {
++source_set("f32-vunary_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -50196,11 +52251,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vhswish/gen/f32-vhswish-neon-u16.c",
-+    "src/src/f32-vhswish/gen/f32-vhswish-scalar-u4.c"
++    "src/src/f32-vunary/gen/f32-vabs-neon.c",
++    "src/src/f32-vunary/gen/f32-vabs-scalar.c",
++    "src/src/f32-vunary/gen/f32-vneg-neon.c",
++    "src/src/f32-vunary/gen/f32-vneg-scalar.c",
++    "src/src/f32-vunary/gen/f32-vsqr-neon.c",
++    "src/src/f32-vunary/gen/f32-vsqr-scalar.c"
 +  ]
  
--  source_set("qs8-rsum_f16c-fma-avx2") {
+-  source_set("qs8-vaddc_f16c-fma-avx2") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
@@ -50212,7 +52271,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rsum/gen/qs8-rsum-avx2-u64-acc2.c",
+-      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-avx2-mul32-ld64-u16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -50240,18 +52299,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-rsum_f16c-fma-avx2_standalone") {
+-  source_set("qs8-vaddc_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("f32-vlog_arm64") {
++source_set("operators_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rsum/gen/qs8-rsum-avx2-u64-acc2.c",
+-      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-avx2-mul32-ld64-u16.c",
 -    ]
 +  ]
  
@@ -50267,8 +52326,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vlog/gen/f32-vlog-neon-rational-3-3-div.c",
-+    "src/src/f32-vlog/gen/f32-vlog-scalar-rational-3-3-div.c"
++    "src/src/operators/argmax-pooling-nhwc.c",
++    "src/src/operators/average-pooling-nhwc.c",
++    "src/src/operators/batch-matrix-multiply-nc.c",
++    "src/src/operators/binary-elementwise-nd.c",
++    "src/src/operators/constant-pad-nd.c",
++    "src/src/operators/convolution-nchw.c",
++    "src/src/operators/convolution-nhwc.c",
++    "src/src/operators/deconvolution-nhwc.c",
++    "src/src/operators/dynamic-fully-connected-nc.c",
++    "src/src/operators/fully-connected-nc.c",
++    "src/src/operators/max-pooling-nhwc.c",
++    "src/src/operators/pack-lh.c",
++    "src/src/operators/reduce-nd.c",
++    "src/src/operators/resize-bilinear-nchw.c",
++    "src/src/operators/resize-bilinear-nhwc.c",
++    "src/src/operators/rope-nthc.c",
++    "src/src/operators/slice-nd.c",
++    "src/src/operators/softmax-nc.c",
++    "src/src/operators/transpose-nd.c",
++    "src/src/operators/unary-elementwise-nc.c",
++    "src/src/operators/unpooling-nhwc.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -50281,7 +52359,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    }
 -  }
 -
--  source_set("qs8-rsum_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-  source_set("qs8-vaddc_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
@@ -50300,8 +52378,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rsum/gen/qs8-rsum-avx256skx-u64-acc2.c",
--      "src/src/qs8-rsum/gen/qs8-rsum-avx512skx-u128-acc2.c",
+-      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-avx512skx-mul32-ld128-u16.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -50317,7 +52394,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vlog_arm64_standalone") {
++source_set("operators_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -50326,7 +52403,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
--      "qs8-rsum_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-      "qs8-vaddc_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
@@ -50338,14 +52415,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vlog/gen/f32-vlog-neon-rational-3-3-div.c",
-+    "src/src/f32-vlog/gen/f32-vlog-scalar-rational-3-3-div.c"
++    "src/src/operators/argmax-pooling-nhwc.c",
++    "src/src/operators/average-pooling-nhwc.c",
++    "src/src/operators/batch-matrix-multiply-nc.c",
++    "src/src/operators/binary-elementwise-nd.c",
++    "src/src/operators/constant-pad-nd.c",
++    "src/src/operators/convolution-nchw.c",
++    "src/src/operators/convolution-nhwc.c",
++    "src/src/operators/deconvolution-nhwc.c",
++    "src/src/operators/dynamic-fully-connected-nc.c",
++    "src/src/operators/fully-connected-nc.c",
++    "src/src/operators/max-pooling-nhwc.c",
++    "src/src/operators/pack-lh.c",
++    "src/src/operators/reduce-nd.c",
++    "src/src/operators/resize-bilinear-nchw.c",
++    "src/src/operators/resize-bilinear-nhwc.c",
++    "src/src/operators/rope-nthc.c",
++    "src/src/operators/slice-nd.c",
++    "src/src/operators/softmax-nc.c",
++    "src/src/operators/transpose-nd.c",
++    "src/src/operators/unary-elementwise-nc.c",
++    "src/src/operators/unpooling-nhwc.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rsum/gen/qs8-rsum-avx256skx-u64-acc2.c",
--      "src/src/qs8-rsum/gen/qs8-rsum-avx512skx-u128-acc2.c",
+-      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-avx512skx-mul32-ld128-u16.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -50374,32 +52469,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
--
--  source_set(
--      "qs8-rsum_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
--      "-mf16c",
--      "-mfma",
--    ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
  
+-  source_set("qs8-vaddc_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rsum/gen/qs8-rsum-avx512vnni-u128-acc2.c",
+-      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-sse2-mul16-ld64-u8.c",
 -    ]
-+source_set("f32-vlrelu_arm64") {
-+  cflags = [
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -50410,28 +52495,24 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++source_set("pf16-gemm_arch=armv8.2-a+sve+sve2") {
++  cflags = [
++    "-march=armv8.2-a+sve+sve2"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vlrelu/gen/f32-vlrelu-neon-u8.c",
-+    "src/src/f32-vlrelu/gen/f32-vlrelu-scalar-u4.c"
++    "src/src/pf16-gemm/pf16-gemm-1x32c2-minmax-neonsme2.c",
++    "src/src/pf16-gemm/pf16-gemm-32x32c2-minmax-neonsme2.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qs8-rsum_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl-avx512vnni_standalone") {
+-  source_set("qs8-vaddc_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mavx512vnni",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -50439,7 +52520,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rsum/gen/qs8-rsum-avx512vnni-u128-acc2.c",
+-      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-sse2-mul16-ld64-u8.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -50463,34 +52544,41 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vlrelu_arm64_standalone") {
++source_set("pf16-gemm_arch=armv8.2-a+sve+sve2_standalone") {
 +  cflags = [
++    "-march=armv8.2-a+sve+sve2"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  ]
- 
--  source_set("qs8-rsum_ssse3-no-sse4.1") {
--    cflags = [
--      "-mno-sse4.1",
--      "-mssse3",
--    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vlrelu/gen/f32-vlrelu-neon-u8.c",
-+    "src/src/f32-vlrelu/gen/f32-vlrelu-scalar-u4.c"
++    "src/src/pf16-gemm/pf16-gemm-1x32c2-minmax-neonsme2.c",
++    "src/src/pf16-gemm/pf16-gemm-32x32c2-minmax-neonsme2.c"
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-rsum/gen/qs8-rsum-ssse3-u32-acc2.c",
+-  source_set("qs8-vaddc_sse4.1-no-sse4.2") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-sse41-mul16-ld64-u8.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -50501,34 +52589,29 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +  public_configs = [ ":xnnpack_config" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qs8-rsum_ssse3-no-sse4.1_standalone") {
--    cflags = [
--      "-mno-sse4.1",
--      "-mssse3",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
  
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-vaddc_sse4.1-no-sse4.2_standalone") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
+-    ]
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rsum/gen/qs8-rsum-ssse3-u32-acc2.c",
+-      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-sse41-mul16-ld64-u8.c",
 -    ]
-+source_set("f32-vmulcaddc_arm64") {
++source_set("pf32-gemm_arch=armv8.2-a+sve+sve2") {
 +  cflags = [
++    "-march=armv8.2-a+sve+sve2"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -50540,26 +52623,21 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c1-minmax-scalar-2x.c",
-+    "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c4-minmax-neon-2x.c",
-+    "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c4-minmax-neonfma-2x.c"
++    "src/src/pf32-gemm/pf32-gemm-1x32-minmax-neonsme2.c",
++    "src/src/pf32-gemm/pf32-gemm-32x32-minmax-neonsme2.c"
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qs8-rsum_x64") {
--    cflags = []
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -50567,12 +52645,21 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-  source_set("qs8-vaddc_x64") {
+-    cflags = []
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rsum/gen/qs8-rsum-scalar-u4.c",
+-      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u1.c",
+-      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u4.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++# This is a target that cannot depend on //base.
++source_set("pf32-gemm_arch=armv8.2-a+sve+sve2_standalone") {
++  cflags = [
++    "-march=armv8.2-a+sve+sve2"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -50584,31 +52671,34 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-vmulcaddc_arm64_standalone") {
-+  cflags = [
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/pf32-gemm/pf32-gemm-1x32-minmax-neonsme2.c",
++    "src/src/pf32-gemm/pf32-gemm-32x32-minmax-neonsme2.c"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-rsum_x64_standalone") {
+-  source_set("qs8-vaddc_x64_standalone") {
 -    cflags = []
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c1-minmax-scalar-2x.c",
-+    "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c4-minmax-neon-2x.c",
-+    "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c4-minmax-neonfma-2x.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-rsum/gen/qs8-rsum-scalar-u4.c",
+-      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u1.c",
+-      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u4.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -50620,38 +52710,47 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
++source_set("pqs8-qc8w-gemm_arch=armv8.2-a+sve+sve2") {
++  cflags = [
++    "-march=armv8.2-a+sve+sve2"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/pqs8-qc8w-gemm/pqs8-qc8w-gemm-1x32c4-minmax-neonsme2.c",
++    "src/src/pqs8-qc8w-gemm/pqs8-qc8w-gemm-32x32c4-minmax-neonsme2.c"
++  ]
  
--  source_set("qs8-vadd_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qs8-vcvt_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
-+source_set("f32-vrelu_arm64") {
-+  cflags = [
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vadd/gen/qs8-vadd-minmax-avx-mul32-ld32-u8.c",
+-      "src/src/qs8-vcvt/gen/qs8-vcvt-avx-u32.c",
 -    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -50664,38 +52763,38 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vrelu/gen/f32-vrelu-scalar-u8.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++# This is a target that cannot depend on //base.
++source_set("pqs8-qc8w-gemm_arch=armv8.2-a+sve+sve2_standalone") {
++  cflags = [
++    "-march=armv8.2-a+sve+sve2"
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vadd_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("qs8-vcvt_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/pqs8-qc8w-gemm/pqs8-qc8w-gemm-1x32c4-minmax-neonsme2.c",
++    "src/src/pqs8-qc8w-gemm/pqs8-qc8w-gemm-32x32c4-minmax-neonsme2.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vadd/gen/qs8-vadd-minmax-avx-mul32-ld32-u8.c",
+-      "src/src/qs8-vcvt/gen/qs8-vcvt-avx-u32.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -50707,41 +52806,43 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-vrelu_arm64_standalone") {
-+  cflags = [
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vrelu/gen/f32-vrelu-scalar-u8.c"
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
--  source_set("qs8-vadd_f16c-fma-avx2") {
+-  source_set("qs8-vcvt_f16c-fma-avx2") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+dotprod+fp16") {
++  cflags = [
++    "-march=armv8.2-a+dotprod+fp16"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vadd/gen/qs8-vadd-minmax-avx2-mul32-ld64-u16.c",
+-      "src/src/qs8-vcvt/gen/qs8-vcvt-avx2-u32.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16c4-minmax-neondotfp16arith.c",
++    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-4x16c4-minmax-neondotfp16arith.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -50754,28 +52855,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vadd_f16c-fma-avx2_standalone") {
+-  source_set("qs8-vcvt_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("f32-vrnd_arm64") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vadd/gen/qs8-vadd-minmax-avx2-mul32-ld64-u16.c",
+-      "src/src/qs8-vcvt/gen/qs8-vcvt-avx2-u32.c",
 -    ]
++# This is a target that cannot depend on //base.
++source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+dotprod+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+dotprod+fp16"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -50790,22 +52900,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrnd/gen/f32-vrndd-neon-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndd-neonv8-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-neon-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-neonv8-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-neon-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-neonv8-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-neon-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-neonv8-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u4.c"
++    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16c4-minmax-neondotfp16arith.c",
++    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-4x16c4-minmax-neondotfp16arith.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -50817,29 +52913,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set("qs8-vadd_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
--    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-  source_set("qs8-vcvt_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vadd/gen/qs8-vadd-minmax-avx512skx-mul32-ld128-u16.c",
+-      "src/src/qs8-vcvt/gen/qs8-vcvt-sse2-u32.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
 +}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -50852,53 +52946,39 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-vrnd_arm64_standalone") {
++source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+fp16") {
 +  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16-minmax-neonfp16arith-mlal-lane.c",
++    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-6x16-minmax-neonfp16arith-mlal-lane.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qs8-vadd_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-  source_set("qs8-vcvt_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vrnd/gen/f32-vrndd-neon-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndd-neonv8-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-neon-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-neonv8-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-neon-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-neonv8-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-neon-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-neonv8-u8.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u4.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vadd/gen/qs8-vadd-minmax-avx512skx-mul32-ld128-u16.c",
+-      "src/src/qs8-vcvt/gen/qs8-vcvt-sse2-u32.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -50910,6 +52990,39 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    public_configs = [ ":xnnpack_config" ]
++# This is a target that cannot depend on //base.
++source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16-minmax-neonfp16arith-mlal-lane.c",
++    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-6x16-minmax-neonfp16arith-mlal-lane.c"
++  ]
+ 
+-  source_set("qs8-vcvt_sse4.1-no-sse4.2") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vcvt/gen/qs8-vcvt-sse41-u32.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -50917,29 +53030,88 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  public_configs = [ ":xnnpack_config" ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
    }
 +}
  
--  source_set("qs8-vadd_sse2-no-sse3") {
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-vcvt_sse4.1-no-sse4.2_standalone") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
+-    ]
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vcvt/gen/qs8-vcvt-sse41-u32.c",
+-    ]
++source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+i8mm+fp16") {
++  cflags = [
++    "-march=armv8.2-a+i8mm+fp16"
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16c8-minmax-neoni8mm.c",
++    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-4x16c8-minmax-neoni8mm.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-  source_set("qs8-vcvt_ssse3-no-sse4.1") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mno-sse4.1",
+-      "-mssse3",
 -    ]
-+source_set("f32-vrsqrt_arm64") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vadd/gen/qs8-vadd-minmax-sse2-mul16-ld64-u8.c",
+-      "src/src/qs8-vcvt/gen/qs8-vcvt-ssse3-u32.c",
 -    ]
++# This is a target that cannot depend on //base.
++source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -50954,9 +53126,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-neon-rsqrt-u16.c",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u4.c"
++    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16c8-minmax-neoni8mm.c",
++    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-4x16c8-minmax-neoni8mm.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -50966,24 +53137,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vadd_sse2-no-sse3_standalone") {
+-  source_set("qs8-vcvt_ssse3-no-sse4.1_standalone") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mno-sse4.1",
+-      "-mssse3",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vadd/gen/qs8-vadd-minmax-sse2-mul16-ld64-u8.c",
+-      "src/src/qs8-vcvt/gen/qs8-vcvt-ssse3-u32.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
-+}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -50995,11 +53165,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-vrsqrt_arm64_standalone") {
-+  cflags = [
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    public_configs = [ ":xnnpack_config" ]
++source_set("qd8-f16-qc4w-gemm_arch=armv8.2-a+dotprod+fp16") {
++  cflags = [
++    "-march=armv8.2-a+dotprod+fp16"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -51008,29 +53182,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-neon-rsqrt-u16.c",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u4.c"
++    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x16c4-minmax-neondotfp16arith.c",
++    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-4x16c4-minmax-neondotfp16arith.c"
 +  ]
  
--  source_set("qs8-vadd_sse4.1-no-sse4.2") {
--    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
--    ]
+-  source_set("qs8-vcvt_x64") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vadd/gen/qs8-vadd-minmax-sse41-mul16-ld64-u8.c",
+-      "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -51044,26 +53214,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++# This is a target that cannot depend on //base.
++source_set("qd8-f16-qc4w-gemm_arch=armv8.2-a+dotprod+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+dotprod+fp16"
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vadd_sse4.1-no-sse4.2_standalone") {
--    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
--    ]
--
+-  source_set("qs8-vcvt_x64_standalone") {
+-    cflags = []
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x16c4-minmax-neondotfp16arith.c",
++    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-4x16c4-minmax-neondotfp16arith.c"
++  ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vadd/gen/qs8-vadd-minmax-sse41-mul16-ld64-u8.c",
+-      "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c",
 -    ]
-+source_set("f32-vsigmoid_arm64") {
-+  cflags = [
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -51075,40 +53251,45 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-neon-rr2-lut64-p2-nr2recps-u8.c",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-neonfma-rr1-lut64-p2-nr2recps-u16.c",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
--  source_set("qs8-vadd_x64") {
--    cflags = []
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
+-  source_set("qs8-vlrelu_avx-no-avx2-no-f16c-no-fma") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
++source_set("qd8-f16-qc4w-gemm_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u1.c",
--      "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u4.c",
+-      "src/src/qs8-vlrelu/gen/qs8-vlrelu-avx-u32.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x16-minmax-neonfp16arith-mlal-lane.c",
++    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-6x16-minmax-neonfp16arith-mlal-lane.c"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -51120,32 +53301,39 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-vsigmoid_arm64_standalone") {
-+  cflags = [
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vadd_x64_standalone") {
--    cflags = []
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-neon-rr2-lut64-p2-nr2recps-u8.c",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-neonfma-rr1-lut64-p2-nr2recps-u16.c",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c"
-+  ]
+-  source_set("qs8-vlrelu_avx-no-avx2-no-f16c-no-fma_standalone") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u1.c",
--      "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u4.c",
+-      "src/src/qs8-vlrelu/gen/qs8-vlrelu-avx-u32.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++# This is a target that cannot depend on //base.
++source_set("qd8-f16-qc4w-gemm_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -51157,39 +53345,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x16-minmax-neonfp16arith-mlal-lane.c",
++    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-6x16-minmax-neonfp16arith-mlal-lane.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
--  source_set("qs8-vaddc_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qs8-vlrelu_f16c-fma-avx2") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+source_set("f32-vsin_arm64") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-avx-mul32-ld32-u8.c",
+-      "src/src/qs8-vlrelu/gen/qs8-vlrelu-avx2-u32.c",
 -    ]
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -51201,25 +53394,33 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vsin/gen/f32-vsin-neon-rational-5-4-div.c",
-+    "src/src/f32-vsin/gen/f32-vsin-scalar-rational-5-4-div.c"
++source_set("qd8-f16-qc4w-gemm_arch=armv8.2-a+i8mm+fp16") {
++  cflags = [
++    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x16c8-minmax-neoni8mm.c",
++    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-4x16c8-minmax-neoni8mm.c"
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-vlrelu_f16c-fma-avx2_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qs8-vaddc_avx-no-avx2-no-f16c-no-fma_standalone") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vlrelu/gen/qs8-vlrelu-avx2-u32.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -51228,13 +53429,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-avx-mul32-ld32-u8.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -51245,11 +53439,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-vsin_arm64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
++# This is a target that cannot depend on //base.
++source_set("qd8-f16-qc4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -51258,15 +53455,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsin/gen/f32-vsin-neon-rational-5-4-div.c",
-+    "src/src/f32-vsin/gen/f32-vsin-scalar-rational-5-4-div.c"
++    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x16c8-minmax-neoni8mm.c",
++    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-4x16c8-minmax-neoni8mm.c"
 +  ]
  
--  source_set("qs8-vaddc_f16c-fma-avx2") {
+-  source_set("qs8-vlrelu_sse2-no-sse3") {
 -    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -51274,7 +53470,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-avx2-mul32-ld64-u16.c",
+-      "src/src/qs8-vlrelu/gen/qs8-vlrelu-sse2-u32.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -51302,19 +53498,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vaddc_f16c-fma-avx2_standalone") {
+-  source_set("qs8-vlrelu_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+source_set("f32-vsqrt_arm64") {
-+  cflags = [
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-avx2-mul32-ld64-u16.c",
+-      "src/src/qs8-vlrelu/gen/qs8-vlrelu-sse2-u32.c",
 -    ]
++source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+dotprod+fp16") {
++  cflags = [
++    "-march=armv8.2-a+dotprod+fp16"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -51329,8 +53525,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsqrt/gen/f32-vsqrt-aarch64-neon-sqrt-u4.c",
-+    "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt-u1.c"
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x16c4-minmax-neondotfp16arith.c",
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c4-minmax-neondotfp16arith.c",
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-neondotfp16arith.c",
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x8c4-minmax-neondotfp16arith.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -51342,31 +53540,125 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-  source_set("qs8-vlrelu_sse4.1-no-sse4.2") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vlrelu/gen/qs8-vlrelu-sse41-u32.c",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+dotprod+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+dotprod+fp16"
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
 -
--  source_set("qs8-vaddc_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x16c4-minmax-neondotfp16arith.c",
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c4-minmax-neondotfp16arith.c",
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-neondotfp16arith.c",
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x8c4-minmax-neondotfp16arith.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-vlrelu_sse4.1-no-sse4.2_standalone") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-avx512skx-mul32-ld128-u16.c",
+-      "src/src/qs8-vlrelu/gen/qs8-vlrelu-sse41-u32.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
 +}
  
+-    public_configs = [ ":xnnpack_config" ]
++source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c2s4-minmax-neonfp16arith.c",
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-2x8c2s4-minmax-neonfp16arith.c"
++  ]
+ 
+-  source_set("qs8-vlrelu_ssse3-no-sse4.1") {
+-    cflags = [
+-      "-mno-sse4.1",
+-      "-mssse3",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vlrelu/gen/qs8-vlrelu-ssse3-u32.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -51377,35 +53669,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-vsqrt_arm64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
++# This is a target that cannot depend on //base.
++source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qs8-vaddc_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-  source_set("qs8-vlrelu_ssse3-no-sse4.1_standalone") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse4.1",
+-      "-mssse3",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsqrt/gen/f32-vsqrt-aarch64-neon-sqrt-u4.c",
-+    "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt-u1.c"
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c2s4-minmax-neonfp16arith.c",
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-2x8c2s4-minmax-neonfp16arith.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-avx512skx-mul32-ld128-u16.c",
+-      "src/src/qs8-vlrelu/gen/qs8-vlrelu-ssse3-u32.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -51439,19 +53728,18 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-vaddc_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
-+source_set("f32-vtanh_arm64") {
+-  source_set("qs8-vlrelu_x64") {
+-    cflags = []
++source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+fp16+dotprod") {
 +  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-sse2-mul16-ld64-u8.c",
+-      "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c",
 -    ]
-+  ]
++  asmflags = cflags
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -51465,8 +53753,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vtanh/gen/f32-vtanh-neon-rational-9-8-div.c",
-+    "src/src/f32-vtanh/gen/f32-vtanh-scalar-rational-9-8-div.c"
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S",
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondotfp16arith-cortex-a55.S"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -51476,11 +53764,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vaddc_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("qs8-vlrelu_x64_standalone") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -51490,7 +53775,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-sse2-mul16-ld64-u8.c",
+-      "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -51506,11 +53791,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vtanh_arm64_standalone") {
++source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
 +  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  ]
++  asmflags = cflags
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
@@ -51518,14 +53805,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vtanh/gen/f32-vtanh-neon-rational-9-8-div.c",
-+    "src/src/f32-vtanh/gen/f32-vtanh-scalar-rational-9-8-div.c"
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S",
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondotfp16arith-cortex-a55.S"
 +  ]
  
--  source_set("qs8-vaddc_sse4.1-no-sse4.2") {
+-  source_set("qs8-vmul_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -51533,7 +53822,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-sse41-mul16-ld64-u8.c",
+-      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-avx-mul16-ld64-u16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -51561,19 +53850,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vaddc_sse4.1-no-sse4.2_standalone") {
+-  source_set("qs8-vmul_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-sse41-mul16-ld64-u8.c",
+-      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-avx-mul16-ld64-u16.c",
 -    ]
-+source_set("f32-vunary_arm64") {
-+  cflags = [
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -51584,29 +53873,36 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
+-
+-    public_configs = [ ":xnnpack_config" ]
++source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+i8mm+fp16") {
++  cflags = [
++    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vunary/gen/f32-vabs-neon.c",
-+    "src/src/f32-vunary/gen/f32-vabs-scalar.c",
-+    "src/src/f32-vunary/gen/f32-vneg-neon.c",
-+    "src/src/f32-vunary/gen/f32-vneg-scalar.c",
-+    "src/src/f32-vunary/gen/f32-vsqr-neon.c",
-+    "src/src/f32-vunary/gen/f32-vsqr-scalar.c"
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x16c8-minmax-neoni8mm.c",
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c8-minmax-neoni8mm.c"
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-  source_set("qs8-vmul_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qs8-vaddc_x64") {
--    cflags = []
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-sse2-mul16-ld64-u8.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -51614,14 +53910,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u1.c",
--      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u4.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -51632,31 +53920,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-vunary_arm64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
++# This is a target that cannot depend on //base.
++source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vaddc_x64_standalone") {
--    cflags = []
+-  source_set("qs8-vmul_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vunary/gen/f32-vabs-neon.c",
-+    "src/src/f32-vunary/gen/f32-vabs-scalar.c",
-+    "src/src/f32-vunary/gen/f32-vneg-neon.c",
-+    "src/src/f32-vunary/gen/f32-vneg-scalar.c",
-+    "src/src/f32-vunary/gen/f32-vsqr-neon.c",
-+    "src/src/f32-vunary/gen/f32-vsqr-scalar.c"
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x16c8-minmax-neoni8mm.c",
++    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c8-minmax-neoni8mm.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u1.c",
--      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u4.c",
+-      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-sse2-mul16-ld64-u8.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -51690,20 +53979,26 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-vcvt_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qs8-vmul_sse4.1-no-sse4.2") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
-+source_set("operators_arm64") {
++source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+dotprod+fp16") {
 +  cflags = [
++    "-march=armv8.2-a+dotprod+fp16"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-avx-u32.c",
+-      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-sse41-mul16-ld64-u16.c",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x16c4-minmax-neondotfp16arith.c",
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c4-minmax-neondotfp16arith.c",
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-neondotfp16arith.c",
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x8c4-minmax-neondotfp16arith.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -51716,157 +54011,34 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/operators/argmax-pooling-nhwc.c",
-+    "src/src/operators/average-pooling-nhwc.c",
-+    "src/src/operators/batch-matrix-multiply-nc.c",
-+    "src/src/operators/binary-elementwise-nd.c",
-+    "src/src/operators/constant-pad-nd.c",
-+    "src/src/operators/convolution-nchw.c",
-+    "src/src/operators/convolution-nhwc.c",
-+    "src/src/operators/deconvolution-nhwc.c",
-+    "src/src/operators/dynamic-fully-connected-nc.c",
-+    "src/src/operators/fully-connected-nc.c",
-+    "src/src/operators/max-pooling-nhwc.c",
-+    "src/src/operators/pack-lh.c",
-+    "src/src/operators/reduce-nd.c",
-+    "src/src/operators/resize-bilinear-nchw.c",
-+    "src/src/operators/resize-bilinear-nhwc.c",
-+    "src/src/operators/rope-nthc.c",
-+    "src/src/operators/slice-nd.c",
-+    "src/src/operators/softmax-nc.c",
-+    "src/src/operators/transpose-nd.c",
-+    "src/src/operators/unary-elementwise-nc.c",
-+    "src/src/operators/unpooling-nhwc.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qs8-vcvt_avx-no-avx2-no-f16c-no-fma_standalone") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-avx-u32.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+# This is a target that cannot depend on //base.
-+source_set("operators_arm64_standalone") {
-+  cflags = [
- 
 -    public_configs = [ ":xnnpack_config" ]
-+  ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/operators/argmax-pooling-nhwc.c",
-+    "src/src/operators/average-pooling-nhwc.c",
-+    "src/src/operators/batch-matrix-multiply-nc.c",
-+    "src/src/operators/binary-elementwise-nd.c",
-+    "src/src/operators/constant-pad-nd.c",
-+    "src/src/operators/convolution-nchw.c",
-+    "src/src/operators/convolution-nhwc.c",
-+    "src/src/operators/deconvolution-nhwc.c",
-+    "src/src/operators/dynamic-fully-connected-nc.c",
-+    "src/src/operators/fully-connected-nc.c",
-+    "src/src/operators/max-pooling-nhwc.c",
-+    "src/src/operators/pack-lh.c",
-+    "src/src/operators/reduce-nd.c",
-+    "src/src/operators/resize-bilinear-nchw.c",
-+    "src/src/operators/resize-bilinear-nhwc.c",
-+    "src/src/operators/rope-nthc.c",
-+    "src/src/operators/slice-nd.c",
-+    "src/src/operators/softmax-nc.c",
-+    "src/src/operators/transpose-nd.c",
-+    "src/src/operators/unary-elementwise-nc.c",
-+    "src/src/operators/unpooling-nhwc.c"
-+  ]
- 
--  source_set("qs8-vcvt_f16c-fma-avx2") {
--    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-avx2-u32.c",
--    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-vmul_sse4.1-no-sse4.2_standalone") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
 +}
  
--  # This is a target that cannot depend on //base.
--  source_set("qs8-vcvt_f16c-fma-avx2_standalone") {
--    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
--    ]
--
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-avx2-u32.c",
+-      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-sse41-mul16-ld64-u16.c",
 -    ]
-+source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+dotprod+fp16") {
++# This is a target that cannot depend on //base.
++source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+dotprod+fp16_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+dotprod+fp16"
 +  ]
@@ -51883,8 +54055,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16c4-minmax-neondotfp16arith.c",
-+    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-4x16c4-minmax-neondotfp16arith.c"
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x16c4-minmax-neondotfp16arith.c",
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c4-minmax-neondotfp16arith.c",
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-neondotfp16arith.c",
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x8c4-minmax-neondotfp16arith.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -51900,26 +54074,21 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--  source_set("qs8-vcvt_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("qs8-vmul_x64") {
+-    cflags = []
 +  public_configs = [ ":xnnpack_config" ]
-+}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-sse2-u32.c",
+-      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-scalar-u4.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+dotprod+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+dotprod+fp16"
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -51931,37 +54100,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16c4-minmax-neondotfp16arith.c",
-+    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-4x16c4-minmax-neondotfp16arith.c"
++source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+fp16") {
++  cflags = [
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c2s4-minmax-neonfp16arith-mlal.c",
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-2x8c2s4-minmax-neonfp16arith-mlal.c"
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-vmul_x64_standalone") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qs8-vcvt_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-scalar-u4.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-sse2-u32.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -51972,13 +54141,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
++  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+fp16") {
++# This is a target that cannot depend on //base.
++source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16"
 +  ]
@@ -51989,14 +54157,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16-minmax-neonfp16arith-mlal-lane.c",
-+    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-6x16-minmax-neonfp16arith-mlal-lane.c"
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c2s4-minmax-neonfp16arith-mlal.c",
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-2x8c2s4-minmax-neonfp16arith-mlal.c"
 +  ]
  
--  source_set("qs8-vcvt_sse4.1-no-sse4.2") {
+-  source_set("qs8-vmulc_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -52004,13 +54174,13 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-sse41-u32.c",
+-      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-avx-mul16-ld64-u16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -52024,35 +54194,30 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
-+}
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vcvt_sse4.1-no-sse4.2_standalone") {
+-  source_set("qs8-vmulc_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16-minmax-neonfp16arith-mlal-lane.c",
-+    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-6x16-minmax-neonfp16arith-mlal-lane.c"
-+  ]
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-sse41-u32.c",
+-      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-avx-mul16-ld64-u16.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+fp16+dotprod") {
++  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -52064,43 +54229,41 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  asmflags = cflags
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qs8-vcvt_ssse3-no-sse4.1") {
+-  source_set("qs8-vmulc_sse2-no-sse3") {
 -    cflags = [
--      "-mno-sse4.1",
--      "-mssse3",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+i8mm+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-ssse3-u32.c",
+-      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-sse2-mul16-ld64-u8.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16c8-minmax-neoni8mm.c",
-+    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-4x16c8-minmax-neoni8mm.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -52112,37 +54275,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++# This is a target that cannot depend on //base.
++source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+fp16+dotprod_standalone") {
++  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++  asmflags = cflags
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vcvt_ssse3-no-sse4.1_standalone") {
+-  source_set("qs8-vmulc_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mno-sse4.1",
--      "-mssse3",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-ssse3-u32.c",
+-      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-sse2-mul16-ld64-u8.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -52154,21 +54315,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16c8-minmax-neoni8mm.c",
-+    "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-4x16c8-minmax-neoni8mm.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -52176,20 +54322,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--  source_set("qs8-vcvt_x64") {
--    cflags = []
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u1.c",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c",
--    ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
  
+-  source_set("qs8-vmulc_sse4.1-no-sse4.2") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
+-    ]
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-sse41-mul16-ld64-u16.c",
+-    ]
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -52200,30 +54354,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("qd8-f16-qc4w-gemm_arch=armv8.2-a+dotprod+fp16") {
++source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+i8mm+fp16") {
 +  cflags = [
-+    "-march=armv8.2-a+dotprod+fp16"
++    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x16c4-minmax-neondotfp16arith.c",
-+    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-4x16c4-minmax-neondotfp16arith.c"
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x16c8-minmax-neoni8mm.c",
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c8-minmax-neoni8mm.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vcvt_x64_standalone") {
--    cflags = []
+-  source_set("qs8-vmulc_sse4.1-no-sse4.2_standalone") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u1.c",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c",
+-      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-sse41-mul16-ld64-u16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -52247,9 +54403,9 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("qd8-f16-qc4w-gemm_arch=armv8.2-a+dotprod+fp16_standalone") {
++source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+i8mm+fp16_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+dotprod+fp16"
++    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -52258,24 +54414,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x16c4-minmax-neondotfp16arith.c",
-+    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-4x16c4-minmax-neondotfp16arith.c"
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x16c8-minmax-neoni8mm.c",
++    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c8-minmax-neoni8mm.c"
 +  ]
  
--  source_set("qs8-vlrelu_avx-no-avx2-no-f16c-no-fma") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
+-  source_set("qs8-vmulc_x64") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-avx-u32.c",
+-      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-scalar-u4.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -52303,21 +54454,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vlrelu_avx-no-avx2-no-f16c-no-fma_standalone") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
+-  source_set("qs8-vmulc_x64_standalone") {
+-    cflags = []
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-avx-u32.c",
+-      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-scalar-u4.c",
 -    ]
-+source_set("qd8-f16-qc4w-gemm_arch=armv8.2-a+fp16") {
++source_set("qd8-f32-qb4w-gemm_arch=armv8.2-a+dotprod") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16"
++    "-march=armv8.2-a+dotprod"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -52332,8 +54478,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x16-minmax-neonfp16arith-mlal-lane.c",
-+    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-6x16-minmax-neonfp16arith-mlal-lane.c"
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c4-minmax-neondot.c",
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x16c4-minmax-neondot.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -52352,7 +54498,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  source_set("qs8-vlrelu_f16c-fma-avx2") {
+-  source_set("qs8-vprelu_f16c-fma-avx2") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
@@ -52363,12 +54509,12 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-avx2-u32.c",
+-      "src/src/qs8-vprelu/gen/qs8-vprelu-avx2-u16.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qd8-f16-qc4w-gemm_arch=armv8.2-a+fp16_standalone") {
++source_set("qd8-f32-qb4w-gemm_arch=armv8.2-a+dotprod_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16"
++    "-march=armv8.2-a+dotprod"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -52383,8 +54529,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x16-minmax-neonfp16arith-mlal-lane.c",
-+    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-6x16-minmax-neonfp16arith-mlal-lane.c"
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c4-minmax-neondot.c",
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x16c4-minmax-neondot.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -52394,7 +54540,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vlrelu_f16c-fma-avx2_standalone") {
+-  source_set("qs8-vprelu_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
@@ -52409,7 +54555,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-avx2-u32.c",
+-      "src/src/qs8-vprelu/gen/qs8-vprelu-avx2-u16.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
  
@@ -52429,7 +54575,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("qd8-f16-qc4w-gemm_arch=armv8.2-a+i8mm+fp16") {
++source_set("qd8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16") {
 +  cflags = [
 +    "-march=armv8.2-a+i8mm+fp16"
 +  ]
@@ -52440,22 +54586,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x16c8-minmax-neoni8mm.c",
-+    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-4x16c8-minmax-neoni8mm.c"
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c8-minmax-neoni8mm.c",
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x16c8-minmax-neoni8mm.c"
 +  ]
  
--  source_set("qs8-vlrelu_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("qs8-vprelu_x64") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-sse2-u32.c",
+-      "src/src/qs8-vprelu/gen/qs8-vprelu-scalar-u8.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -52480,26 +54623,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("qd8-f16-qc4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
++source_set("qd8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vlrelu_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("qs8-vprelu_x64_standalone") {
+-    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-1x16c8-minmax-neoni8mm.c",
-+    "src/src/qd8-f16-qc4w-gemm/gen/qd8-f16-qc4w-gemm-4x16c8-minmax-neoni8mm.c"
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c8-minmax-neoni8mm.c",
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x16c8-minmax-neoni8mm.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-sse2-u32.c",
+-      "src/src/qs8-vprelu/gen/qs8-vprelu-scalar-u8.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -52533,26 +54673,19 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-vlrelu_sse4.1-no-sse4.2") {
+-  source_set("qs8-vpreluc_f16c-fma-avx2") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+dotprod+fp16") {
++source_set("qd8-f32-qb4w-gemm_arm64") {
 +  cflags = [
-+    "-march=armv8.2-a+dotprod+fp16"
-+  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-sse41-u32.c",
+-      "src/src/qs8-vpreluc/gen/qs8-vpreluc-avx2-u16.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x16c4-minmax-neondotfp16arith.c",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c4-minmax-neondotfp16arith.c",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-neondotfp16arith.c",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x8c4-minmax-neondotfp16arith.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -52565,12 +54698,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16-minmax-neon-mlal-lane.c",
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4-minmax-scalar.c",
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4-minmax-scalar.c",
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-6x16-minmax-neon-mlal-lane.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-vpreluc_f16c-fma-avx2_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -52578,24 +54726,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qs8-vlrelu_sse4.1-no-sse4.2_standalone") {
--    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-sse41-u32.c",
+-      "src/src/qs8-vpreluc/gen/qs8-vpreluc-avx2-u16.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+dotprod+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+dotprod+fp16"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -52607,23 +54743,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++# This is a target that cannot depend on //base.
++source_set("qd8-f32-qb4w-gemm_arm64_standalone") {
++  cflags = [
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x16c4-minmax-neondotfp16arith.c",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c4-minmax-neondotfp16arith.c",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-neondotfp16arith.c",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x8c4-minmax-neondotfp16arith.c"
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16-minmax-neon-mlal-lane.c",
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4-minmax-scalar.c",
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4-minmax-scalar.c",
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-6x16-minmax-neon-mlal-lane.c"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
+-  source_set("qs8-vpreluc_x64") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vpreluc/gen/qs8-vpreluc-scalar-u8.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -52631,22 +54779,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
--  source_set("qs8-vlrelu_ssse3-no-sse4.1") {
--    cflags = [
--      "-mno-sse4.1",
--      "-mssse3",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-ssse3-u32.c",
--    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -52657,38 +54789,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c2s4-minmax-neonfp16arith.c",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-2x8c2s4-minmax-neonfp16arith.c"
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vlrelu_ssse3-no-sse4.1_standalone") {
--    cflags = [
--      "-mno-sse4.1",
--      "-mssse3",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
+-  source_set("qs8-vpreluc_x64_standalone") {
+-    cflags = []
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-ssse3-u32.c",
+-      "src/src/qs8-vpreluc/gen/qs8-vpreluc-scalar-u8.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++source_set("qd8-f32-qc4w-gemm_arch=armv8.2-a+dotprod") {
++  cflags = [
++    "-march=armv8.2-a+dotprod"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -52701,42 +54820,45 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c4-minmax-neondot.c",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x16c4-minmax-neondot.c"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c2s4-minmax-neonfp16arith.c",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-2x8c2s4-minmax-neonfp16arith.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
--  source_set("qs8-vlrelu_x64") {
--    cflags = []
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-  source_set("qs8-vrpreluc_f16c-fma-avx2") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-select-u4.c",
+-      "src/src/qs8-vrpreluc/gen/qs8-vrpreluc-avx2-u16.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++# This is a target that cannot depend on //base.
++source_set("qd8-f32-qc4w-gemm_arch=armv8.2-a+dotprod_standalone") {
++  cflags = [
++    "-march=armv8.2-a+dotprod"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -52749,27 +54871,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c4-minmax-neondot.c",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x16c4-minmax-neondot.c"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vlrelu_x64_standalone") {
--    cflags = []
--
+-  source_set("qs8-vrpreluc_f16c-fma-avx2_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-select-u4.c",
+-      "src/src/qs8-vrpreluc/gen/qs8-vrpreluc-avx2-u16.c",
 -    ]
-+source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+fp16+dotprod") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -52781,29 +54913,36 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  asmflags = cflags
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondotfp16arith-cortex-a55.S"
++source_set("qd8-f32-qc4w-gemm_arch=armv8.2-a+i8mm+fp16") {
++  cflags = [
++    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-neoni8mm.c",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x16c8-minmax-neoni8mm.c"
++  ]
+ 
+-  source_set("qs8-vrpreluc_x64") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qs8-vmul_avx-no-avx2-no-f16c-no-fma") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vrpreluc/gen/qs8-vrpreluc-scalar-u8.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -52812,13 +54951,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-avx-mul16-ld64-u16.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -52829,33 +54961,29 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  asmflags = cflags
++# This is a target that cannot depend on //base.
++source_set("qd8-f32-qc4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+i8mm+fp16"
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vmul_avx-no-avx2-no-f16c-no-fma_standalone") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
+-  source_set("qs8-vrpreluc_x64_standalone") {
+-    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondotfp16arith-cortex-a55.S"
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-neoni8mm.c",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x16c8-minmax-neoni8mm.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-avx-mul16-ld64-u16.c",
+-      "src/src/qs8-vrpreluc/gen/qs8-vrpreluc-scalar-u8.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -52889,24 +55017,21 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-vmul_sse2-no-sse3") {
+-  source_set("qu8-dwconv_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
-+source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+i8mm+fp16") {
++source_set("qd8-f32-qc4w-gemm_arm64") {
 +  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
-+  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-sse2-mul16-ld64-u8.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-25p16c-minmax-fp32-avx-mul16.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-fp32-avx-mul16.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x16c8-minmax-neoni8mm.c",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c8-minmax-neoni8mm.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -52919,12 +55044,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16-minmax-neon-mlal-lane.c",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4-minmax-scalar.c",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4-minmax-scalar.c",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-6x16-minmax-neon-mlal-lane.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("qu8-dwconv_avx-no-avx2-no-f16c-no-fma_standalone") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -52932,24 +55073,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qs8-vmul_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-sse2-mul16-ld64-u8.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-25p16c-minmax-fp32-avx-mul16.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-fp32-avx-mul16.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -52961,43 +55091,46 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x16c8-minmax-neoni8mm.c",
-+    "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c8-minmax-neoni8mm.c"
-+  ]
++# This is a target that cannot depend on //base.
++source_set("qd8-f32-qc4w-gemm_arm64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16-minmax-neon-mlal-lane.c",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4-minmax-scalar.c",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4-minmax-scalar.c",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-6x16-minmax-neon-mlal-lane.c"
 +  ]
  
--  source_set("qs8-vmul_sse4.1-no-sse4.2") {
+-  source_set("qu8-dwconv_f16c-fma-avx2") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-sse41-mul16-ld64-u16.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-25p16c-minmax-fp32-avx2-mul32.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-fp32-avx2-mul32.c",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -53009,42 +55142,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+dotprod+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+dotprod+fp16"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x16c4-minmax-neondotfp16arith.c",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c4-minmax-neondotfp16arith.c",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-neondotfp16arith.c",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x8c4-minmax-neondotfp16arith.c"
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vmul_sse4.1-no-sse4.2_standalone") {
+-  source_set("qu8-dwconv_f16c-fma-avx2_standalone") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-sse41-mul16-ld64-u16.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-25p16c-minmax-fp32-avx2-mul32.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-fp32-avx2-mul32.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -53055,43 +55174,50 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++source_set("qd8-f32-qc8w-gemm_arch=armv8.2-a+dotprod") {
++  cflags = [
++    "-march=armv8.2-a+dotprod"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+dotprod+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+dotprod+fp16"
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c4-minmax-neondot.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c4-minmax-neondot.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-neondot.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x8c4-minmax-neondot.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x16c4-minmax-neondotfp16arith.c",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c4-minmax-neondotfp16arith.c",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-neondotfp16arith.c",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x8c4-minmax-neondotfp16arith.c"
-+  ]
- 
--  source_set("qs8-vmul_x64") {
--    cflags = []
+-
+-  source_set(
+-      "qu8-dwconv_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-scalar-u4.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-25p32c-minmax-fp32-avx512skx-mul32.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-9p32c-minmax-fp32-avx512skx-mul32.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -53105,30 +55231,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++# This is a target that cannot depend on //base.
++source_set("qd8-f32-qc8w-gemm_arch=armv8.2-a+dotprod_standalone") {
++  cflags = [
++    "-march=armv8.2-a+dotprod"
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vmul_x64_standalone") {
--    cflags = []
-+source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16"
+-  source_set(
+-      "qu8-dwconv_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c4-minmax-neondot.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c4-minmax-neondot.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-neondot.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x8c4-minmax-neondot.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-scalar-u4.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-25p32c-minmax-fp32-avx512skx-mul32.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-9p32c-minmax-fp32-avx512skx-mul32.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c2s4-minmax-neonfp16arith-mlal.c",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-2x8c2s4-minmax-neonfp16arith-mlal.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -53140,47 +55280,40 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
-+  public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
 +}
  
--  source_set("qs8-vmulc_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qu8-dwconv_sse2-no-sse3") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+fp16_standalone") {
++source_set("qd8-f32-qc8w-gemm_arch=armv8.2-a+fp16+dotprod") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16"
++    "-march=armv8.2-a+fp16+dotprod"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-avx-mul16-ld64-u16.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-fp32-sse2-mul16.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-9p8c-minmax-fp32-sse2-mul16.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c2s4-minmax-neonfp16arith-mlal.c",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-2x8c2s4-minmax-neonfp16arith-mlal.c"
-+  ]
++  asmflags = cflags
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -53192,36 +55325,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("qu8-dwconv_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qs8-vmulc_avx-no-avx2-no-f16c-no-fma_standalone") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-avx-mul16-ld64-u16.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-fp32-sse2-mul16.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-9p8c-minmax-fp32-sse2-mul16.c",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
++  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -53234,7 +55368,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+fp16+dotprod") {
++# This is a target that cannot depend on //base.
++source_set("qd8-f32-qc8w-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16+dotprod"
 +  ]
@@ -53248,14 +55383,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S"
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S"
 +  ]
  
--  source_set("qs8-vmulc_sse2-no-sse3") {
+-  source_set("qu8-dwconv_sse4.1-no-sse4.2") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -53263,13 +55398,14 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-sse2-mul16-ld64-u8.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-fp32-sse41-mul16.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-9p8c-minmax-fp32-sse41-mul16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -53283,84 +55419,33 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+fp16+dotprod_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("qs8-vmulc_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
-+  asmflags = cflags
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-sse2-mul16-ld64-u8.c",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
    }
 +}
  
--  source_set("qs8-vmulc_sse4.1-no-sse4.2") {
+-  # This is a target that cannot depend on //base.
+-  source_set("qu8-dwconv_sse4.1-no-sse4.2_standalone") {
 -    cflags = [
 -      "-mno-sse4.2",
 -      "-msse4.1",
 -    ]
-+source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+i8mm+fp16") {
++source_set("qd8-f32-qc8w-gemm_arch=armv8.2-a+i8mm+fp16") {
 +  cflags = [
 +    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-sse41-mul16-ld64-u16.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-fp32-sse41-mul16.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-9p8c-minmax-fp32-sse41-mul16.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x16c8-minmax-neoni8mm.c",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c8-minmax-neoni8mm.c"
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c8-minmax-neoni8mm.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c8-minmax-neoni8mm.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -53371,14 +55456,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -53386,68 +55470,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qs8-vmulc_sse4.1-no-sse4.2_standalone") {
--    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-sse41-mul16-ld64-u16.c",
--    ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+i8mm+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x16c8-minmax-neoni8mm.c",
-+    "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c8-minmax-neoni8mm.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--  source_set("qs8-vmulc_x64") {
+-
+-  source_set("qu8-dwconv_x64") {
 -    cflags = []
-+  public_configs = [ ":xnnpack_config" ]
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-scalar-u4.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
++  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -53460,36 +55498,34 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("qd8-f32-qb4w-gemm_arch=armv8.2-a+dotprod") {
++# This is a target that cannot depend on //base.
++source_set("qd8-f32-qc8w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+dotprod"
++    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c4-minmax-neondot.c",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x16c4-minmax-neondot.c"
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c8-minmax-neoni8mm.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c8-minmax-neoni8mm.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vmulc_x64_standalone") {
+-  source_set("qu8-dwconv_x64_standalone") {
 -    cflags = []
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-scalar-u4.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
+-      "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -53501,47 +55537,38 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qb4w-gemm_arch=armv8.2-a+dotprod_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+dotprod"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c4-minmax-neondot.c",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x16c4-minmax-neondot.c"
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
--  source_set("qu8-dwconv_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qu8-f32-vcvt_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++source_set("qd8-f32-qc8w-gemm_arm64") {
++  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p16c-minmax-fp32-avx-mul16.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-fp32-avx-mul16.c",
+-      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx-u32.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -53554,37 +55581,42 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x2-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c2s4-minmax-neon-mlal.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-2x8c2s4-minmax-neon-mlal.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4-minmax-scalar.c"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-dwconv_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("qu8-f32-vcvt_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
-+source_set("qd8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p16c-minmax-fp32-avx-mul16.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-fp32-avx-mul16.c",
+-      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx-u32.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c8-minmax-neoni8mm.c",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x16c8-minmax-neoni8mm.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -53596,46 +55628,45 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++# This is a target that cannot depend on //base.
++source_set("qd8-f32-qc8w-gemm_arm64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x2-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c2s4-minmax-neon-mlal.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-2x8c2s4-minmax-neon-mlal.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4-minmax-scalar.c"
++  ]
  
--  source_set("qu8-dwconv_f16c-fma-avx2") {
+-  source_set("qu8-f32-vcvt_f16c-fma-avx2") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p16c-minmax-fp32-avx2-mul32.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-fp32-avx2-mul32.c",
+-      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx2-u16.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c8-minmax-neoni8mm.c",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x16c8-minmax-neoni8mm.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -53648,38 +55679,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-dwconv_f16c-fma-avx2_standalone") {
+-  source_set("qu8-f32-vcvt_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p16c-minmax-fp32-avx2-mul32.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-fp32-avx2-mul32.c",
+-      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx2-u16.c",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -53690,10 +55710,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("qd8-f32-qb4w-gemm_arm64") {
++source_set("qd8-f32-qc8w-igemm_arch=armv8.2-a+dotprod") {
 +  cflags = [
++    "-march=armv8.2-a+dotprod"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c4-minmax-neondot.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c4-minmax-neondot.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-neondot.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x8c4-minmax-neondot.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -53702,7 +55730,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 -
 -  source_set(
--      "qu8-dwconv_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-      "qu8-f32-vcvt_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
@@ -53712,22 +55740,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16-minmax-neon-mlal-lane.c",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-6x16-minmax-neon-mlal-lane.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p32c-minmax-fp32-avx512skx-mul32.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p32c-minmax-fp32-avx512skx-mul32.c",
+-      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx512skx-u32.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -53739,21 +55765,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  public_configs = [ ":xnnpack_config" ]
-+}
++# This is a target that cannot depend on //base.
++source_set("qd8-f32-qc8w-igemm_arch=armv8.2-a+dotprod_standalone") {
++  cflags = [
++    "-march=armv8.2-a+dotprod"
++  ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
--      "qu8-dwconv_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-      "qu8-f32-vcvt_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
@@ -53763,16 +55788,21 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qb4w-gemm_arm64_standalone") {
-+  cflags = [
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c4-minmax-neondot.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c4-minmax-neondot.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-neondot.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x8c4-minmax-neondot.c"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p32c-minmax-fp32-avx512skx-mul32.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p32c-minmax-fp32-avx512skx-mul32.c",
+-      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx512skx-u32.c",
 -    ]
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -53784,46 +55814,39 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16-minmax-neon-mlal-lane.c",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-6x16-minmax-neon-mlal-lane.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
--  source_set("qu8-dwconv_sse2-no-sse3") {
+-  source_set("qu8-f32-vcvt_sse2-no-sse3") {
 -    cflags = [
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++source_set("qd8-f32-qc8w-igemm_arch=armv8.2-a+fp16+dotprod") {
++  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-fp32-sse2-mul16.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p8c-minmax-fp32-sse2-mul16.c",
+-      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-sse2-u32.c",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++  asmflags = cflags
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -53835,34 +55858,24 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("qd8-f32-qc4w-gemm_arch=armv8.2-a+dotprod") {
-+  cflags = [
-+    "-march=armv8.2-a+dotprod"
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c4-minmax-neondot.c",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x16c4-minmax-neondot.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-dwconv_sse2-no-sse3_standalone") {
+-  source_set("qu8-f32-vcvt_sse2-no-sse3_standalone") {
 -    cflags = [
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-fp32-sse2-mul16.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p8c-minmax-fp32-sse2-mul16.c",
--    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -53870,6 +55883,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-sse2-u32.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -53880,27 +55900,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qc4w-gemm_arch=armv8.2-a+dotprod_standalone") {
++source_set("qd8-f32-qc8w-igemm_arch=armv8.2-a+fp16+dotprod_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+dotprod"
++    "-march=armv8.2-a+fp16+dotprod"
 +  ]
  
+-    public_configs = [ ":xnnpack_config" ]
++  asmflags = cflags
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c4-minmax-neondot.c",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x16c4-minmax-neondot.c"
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S"
 +  ]
  
--  source_set("qu8-dwconv_sse4.1-no-sse4.2") {
+-  source_set("qu8-f32-vcvt_sse4.1-no-sse4.2") {
 -    cflags = [
 -      "-mno-sse4.2",
 -      "-msse4.1",
@@ -53911,8 +55930,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-fp32-sse41-mul16.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p8c-minmax-fp32-sse41-mul16.c",
+-      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-sse41-u16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -53940,27 +55958,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-dwconv_sse4.1-no-sse4.2_standalone") {
+-  source_set("qu8-f32-vcvt_sse4.1-no-sse4.2_standalone") {
 -    cflags = [
 -      "-mno-sse4.2",
 -      "-msse4.1",
 -    ]
-+source_set("qd8-f32-qc4w-gemm_arch=armv8.2-a+i8mm+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
-+  ]
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-fp32-sse41-mul16.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p8c-minmax-fp32-sse41-mul16.c",
+-      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-sse41-u16.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-neoni8mm.c",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x16c8-minmax-neoni8mm.c"
-+  ]
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -53971,37 +55979,39 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
+-
 -    public_configs = [ ":xnnpack_config" ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++source_set("qd8-f32-qc8w-igemm_arch=armv8.2-a+i8mm+fp16") {
++  cflags = [
++    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set("qu8-dwconv_x64") {
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c8-minmax-neoni8mm.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c8-minmax-neoni8mm.c"
++  ]
+ 
+-  source_set("qu8-f32-vcvt_x64") {
 -    cflags = []
--
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -54013,32 +56023,29 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qc4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
++source_set("qd8-f32-qc8w-igemm_arch=armv8.2-a+i8mm+fp16_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("qu8-f32-vcvt_x64_standalone") {
+-    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-neoni8mm.c",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x16c8-minmax-neoni8mm.c"
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c8-minmax-neoni8mm.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c8-minmax-neoni8mm.c"
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qu8-dwconv_x64_standalone") {
--    cflags = []
--
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -54072,19 +56079,20 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-f32-vcvt_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qu8-gemm_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
-+source_set("qd8-f32-qc4w-gemm_arm64") {
++source_set("qd8-f32-qc8w-igemm_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx-u32.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-avx-ld128.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-2x4c8-minmax-fp32-avx-ld128.c",
 -    ]
 +  ]
  
@@ -54100,10 +56108,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16-minmax-neon-mlal-lane.c",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-6x16-minmax-neon-mlal-lane.c"
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x2-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8-minmax-neon-mlal-lane.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c2s4-minmax-neon-mlal.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-2x8c2s4-minmax-neon-mlal.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x8-minmax-neon-mlal-lane.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -54113,7 +56124,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-f32-vcvt_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("qu8-gemm_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
@@ -54129,7 +56140,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx-u32.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-avx-ld128.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-2x4c8-minmax-fp32-avx-ld128.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -54145,7 +56157,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qc4w-gemm_arm64_standalone") {
++source_set("qd8-f32-qc8w-igemm_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -54157,13 +56169,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16-minmax-neon-mlal-lane.c",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-6x16-minmax-neon-mlal-lane.c"
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x2-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8-minmax-neon-mlal-lane.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c2s4-minmax-neon-mlal.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-2x8c2s4-minmax-neon-mlal.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x8-minmax-neon-mlal-lane.c"
 +  ]
  
--  source_set("qu8-f32-vcvt_f16c-fma-avx2") {
+-  source_set("qu8-gemm_f16c-fma-avx2") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
@@ -54175,7 +56190,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx2-u16.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-1x8c8-minmax-fp32-avx2.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-3x8c8-minmax-fp32-avx2.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -54203,18 +56219,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-f32-vcvt_f16c-fma-avx2_standalone") {
+-  source_set("qu8-gemm_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
--
++source_set("qp8-f32-qb4w-gemm_arch=armv8.2-a+dotprod") {
++  cflags = [
++    "-march=armv8.2-a+dotprod"
++  ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx2-u16.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-1x8c8-minmax-fp32-avx2.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-3x8c8-minmax-fp32-avx2.c",
 -    ]
--
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-1x4c16s2-aarch64-neondot.c",
++    "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-1x8c16s2-aarch64-neondot.c"
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -54225,18 +56251,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("qd8-f32-qc8w-gemm_arch=armv8.2-a+dotprod") {
-+  cflags = [
-+    "-march=armv8.2-a+dotprod"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c4-minmax-neondot.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c4-minmax-neondot.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-neondot.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x8c4-minmax-neondot.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -54244,8 +56268,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    }
 -  }
 -
--  source_set(
--      "qu8-f32-vcvt_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-  source_set("qu8-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
@@ -54255,19 +56278,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx512skx-u32.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-7x16c8-minmax-fp32-avx512skx-prfm.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++# This is a target that cannot depend on //base.
++source_set("qp8-f32-qb4w-gemm_arch=armv8.2-a+dotprod_standalone") {
++  cflags = [
++    "-march=armv8.2-a+dotprod"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -54280,20 +56302,21 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-1x4c16s2-aarch64-neondot.c",
++    "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-1x8c16s2-aarch64-neondot.c"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qc8w-gemm_arch=armv8.2-a+dotprod_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+dotprod"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
--      "qu8-f32-vcvt_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-      "qu8-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
@@ -54303,21 +56326,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c4-minmax-neondot.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c4-minmax-neondot.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-neondot.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x8c4-minmax-neondot.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-avx512skx-u32.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-7x16c8-minmax-fp32-avx512skx-prfm.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -54329,39 +56350,46 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
++source_set("qp8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16") {
++  cflags = [
++    "-march=armv8.2-a+i8mm+fp16"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-16x4c16s2-mstep4-neoni8mm.c"
++  ]
  
--  source_set("qu8-f32-vcvt_sse2-no-sse3") {
+-  source_set("qu8-gemm_sse2-no-sse3") {
 -    cflags = [
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+source_set("qd8-f32-qc8w-gemm_arch=armv8.2-a+fp16+dotprod") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-sse2-u32.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-sse2-ld64.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-3x4c8-minmax-fp32-sse2-ld64.c",
 -    ]
-+  asmflags = cflags
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -54373,37 +56401,36 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++# This is a target that cannot depend on //base.
++source_set("qp8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+i8mm+fp16"
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-f32-vcvt_sse2-no-sse3_standalone") {
+-  source_set("qu8-gemm_sse2-no-sse3_standalone") {
 -    cflags = [
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-16x4c16s2-mstep4-neoni8mm.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-sse2-u32.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-sse2-ld64.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-3x4c8-minmax-fp32-sse2-ld64.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -54415,43 +56442,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qc8w-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  asmflags = cflags
++  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S"
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
--  source_set("qu8-f32-vcvt_sse4.1-no-sse4.2") {
+-  source_set("qu8-gemm_sse4.1-no-sse4.2") {
 -    cflags = [
 -      "-mno-sse4.2",
 -      "-msse4.1",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+dotprod") {
++  cflags = [
++    "-march=armv8.2-a+dotprod"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-sse41-u16.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-sse41-ld64.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-3x4c8-minmax-fp32-sse41-ld64.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-16x4c8s2-mstep4-aarch64-neondot.c",
++    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x4c8s2-aarch64-neondot.c",
++    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x8c16s2-aarch64-neondot.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -54464,26 +56492,39 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-f32-vcvt_sse4.1-no-sse4.2_standalone") {
+-  source_set("qu8-gemm_sse4.1-no-sse4.2_standalone") {
 -    cflags = [
 -      "-mno-sse4.2",
 -      "-msse4.1",
 -    ]
--
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-sse41-u16.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-sse41-ld64.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-3x4c8-minmax-fp32-sse41-ld64.c",
 -    ]
--
++# This is a target that cannot depend on //base.
++source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+dotprod_standalone") {
++  cflags = [
++    "-march=armv8.2-a+dotprod"
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -54494,41 +56535,43 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
--
--    public_configs = [ ":xnnpack_config" ]
-+source_set("qd8-f32-qc8w-gemm_arch=armv8.2-a+i8mm+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
-+  ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c8-minmax-neoni8mm.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c8-minmax-neoni8mm.c"
++    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-16x4c8s2-mstep4-aarch64-neondot.c",
++    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x4c8s2-aarch64-neondot.c",
++    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x8c16s2-aarch64-neondot.c"
 +  ]
  
--  source_set("qu8-f32-vcvt_x64") {
--    cflags = []
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u1.c",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c",
--    ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-  source_set("qu8-gemm_x64") {
+-    cflags = []
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c",
+-    ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -54539,35 +56582,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qc8w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
++source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+i8mm+fp16") {
 +  cflags = [
 +    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qu8-f32-vcvt_x64_standalone") {
--    cflags = []
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c8-minmax-neoni8mm.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c8-minmax-neoni8mm.c"
++    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-8x8c16s2-mstep2-neoni8mm.c"
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u1.c",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c",
--    ]
+-  # This is a target that cannot depend on //base.
+-  source_set("qu8-gemm_x64_standalone") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -54578,39 +56623,46 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
++# This is a target that cannot depend on //base.
++source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+i8mm+fp16"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-8x8c16s2-mstep2-neoni8mm.c"
++  ]
  
--  source_set("qu8-gemm_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qu8-igemm_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
-+source_set("qd8-f32-qc8w-gemm_arm64") {
-+  cflags = [
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-avx-ld128.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-2x4c8-minmax-fp32-avx-ld128.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-avx-ld128.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-2x4c8-minmax-fp32-avx-ld128.c",
 -    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -54623,44 +56675,37 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x2-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c2s4-minmax-neon-mlal.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-2x2-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-2x8c2s4-minmax-neon-mlal.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4-minmax-scalar.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-gemm_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("qu8-igemm_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+sve+sve2") {
++  cflags = [
++    "-march=armv8.2-a+sve+sve2"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-avx-ld128.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-2x4c8-minmax-fp32-avx-ld128.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-avx-ld128.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-2x4c8-minmax-fp32-avx-ld128.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x128c4-neonsme2.c",
++    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-32x128c4-neonsme2.c"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -54672,47 +56717,46 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qc8w-gemm_arm64_standalone") {
-+  cflags = [
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    public_configs = [ ":xnnpack_config" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x2-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c2s4-minmax-neon-mlal.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-2x2-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-2x8c2s4-minmax-neon-mlal.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4-minmax-scalar.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
--  source_set("qu8-gemm_f16c-fma-avx2") {
+-  source_set("qu8-igemm_f16c-fma-avx2") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++# This is a target that cannot depend on //base.
++source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+sve+sve2_standalone") {
++  cflags = [
++    "-march=armv8.2-a+sve+sve2"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-gemm/gen/qu8-gemm-1x8c8-minmax-fp32-avx2.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-3x8c8-minmax-fp32-avx2.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-1x8c8-minmax-fp32-avx2.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-3x8c8-minmax-fp32-avx2.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x128c4-neonsme2.c",
++    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-32x128c4-neonsme2.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -54725,28 +56769,38 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-gemm_f16c-fma-avx2_standalone") {
+-  source_set("qu8-igemm_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
--
++  public_configs = [ ":xnnpack_config" ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-gemm/gen/qu8-gemm-1x8c8-minmax-fp32-avx2.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-3x8c8-minmax-fp32-avx2.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-1x8c8-minmax-fp32-avx2.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-3x8c8-minmax-fp32-avx2.c",
 -    ]
--
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -54757,7 +56811,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("qd8-f32-qc8w-igemm_arch=armv8.2-a+dotprod") {
++source_set("qp8-f32-qc8w-gemm_arch=armv8.2-a+dotprod") {
 +  cflags = [
 +    "-march=armv8.2-a+dotprod"
 +  ]
@@ -54765,10 +56819,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c4-minmax-neondot.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c4-minmax-neondot.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-neondot.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x8c4-minmax-neondot.c"
++    "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-16x4c4-mstep4-aarch64-neondot.c",
++    "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-1x4c4-aarch64-neondot.c",
++    "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-1x4c8-aarch64-neondot.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -54776,7 +56829,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    }
 -  }
 -
--  source_set("qu8-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-  source_set("qu8-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
@@ -54792,8 +56845,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-gemm/gen/qu8-gemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-7x16c8-minmax-fp32-avx512skx-prfm.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-7x16c8-minmax-fp32-avx512skx-prfm.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -54818,14 +56871,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qc8w-igemm_arch=armv8.2-a+dotprod_standalone") {
++source_set("qp8-f32-qc8w-gemm_arch=armv8.2-a+dotprod_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+dotprod"
 +  ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
--      "qu8-gemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-      "qu8-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
@@ -54837,16 +56890,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c4-minmax-neondot.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c4-minmax-neondot.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-neondot.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x8c4-minmax-neondot.c"
++    "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-16x4c4-mstep4-aarch64-neondot.c",
++    "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-1x4c4-aarch64-neondot.c",
++    "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-1x4c8-aarch64-neondot.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-gemm/gen/qu8-gemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-7x16c8-minmax-fp32-avx512skx-prfm.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-7x16c8-minmax-fp32-avx512skx-prfm.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -54880,22 +56932,25 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-gemm_sse2-no-sse3") {
+-  source_set("qu8-igemm_sse2-no-sse3") {
 -    cflags = [
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+source_set("qd8-f32-qc8w-igemm_arch=armv8.2-a+fp16+dotprod") {
++source_set("qp8-f32-qc8w-gemm_arch=armv8.2-a+i8mm+fp16") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
++    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-sse2-ld64.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-3x4c8-minmax-fp32-sse2-ld64.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-sse2-ld64.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-3x4c8-minmax-fp32-sse2-ld64.c",
 -    ]
-+  asmflags = cflags
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-16x4c8-mstep4-neoni8mm.c"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -54907,24 +56962,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qu8-gemm_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -54932,104 +56975,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-sse2-ld64.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-3x4c8-minmax-fp32-sse2-ld64.c",
+-  # This is a target that cannot depend on //base.
+-  source_set("qu8-igemm_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qc8w-igemm_arch=armv8.2-a+fp16+dotprod_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  asmflags = cflags
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S"
-+  ]
- 
--  source_set("qu8-gemm_sse4.1-no-sse4.2") {
--    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
--    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-sse41-ld64.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-3x4c8-minmax-fp32-sse41-ld64.c",
--    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
- 
--  # This is a target that cannot depend on //base.
--  source_set("qu8-gemm_sse4.1-no-sse4.2_standalone") {
--    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-sse2-ld64.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-3x4c8-minmax-fp32-sse2-ld64.c",
 -    ]
-+source_set("qd8-f32-qc8w-igemm_arch=armv8.2-a+i8mm+fp16") {
++# This is a target that cannot depend on //base.
++source_set("qp8-f32-qc8w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-gemm/gen/qu8-gemm-1x4c8-minmax-fp32-sse41-ld64.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-3x4c8-minmax-fp32-sse41-ld64.c",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c8-minmax-neoni8mm.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c8-minmax-neoni8mm.c"
-+  ]
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -55040,85 +57005,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
--
--  source_set("qu8-gemm_x64") {
--    cflags = []
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-gemm/gen/qu8-gemm-1x2-minmax-fp32-scalar-imagic.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-2x2-minmax-fp32-scalar-imagic.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qc8w-igemm_arch=armv8.2-a+i8mm+fp16_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c8-minmax-neoni8mm.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c8-minmax-neoni8mm.c"
++    "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-16x4c8-mstep4-neoni8mm.c"
 +  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("qu8-gemm_x64_standalone") {
--    cflags = []
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-gemm/gen/qu8-gemm-1x2-minmax-fp32-scalar-imagic.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-2x2-minmax-fp32-scalar-imagic.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c",
--    ]
++
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
++
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -55137,20 +57032,18 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-igemm_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qu8-igemm_sse4.1-no-sse4.2") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
-+source_set("qd8-f32-qc8w-igemm_arm64") {
++source_set("qs8-dwconv_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-avx-ld128.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-2x4c8-minmax-fp32-avx-ld128.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-sse41-ld64.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-3x4c8-minmax-fp32-sse41-ld64.c",
 -    ]
 +  ]
  
@@ -55166,14 +57059,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x2-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8-minmax-neon-mlal-lane.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c2s4-minmax-neon-mlal.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-2x2-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-2x8c2s4-minmax-neon-mlal.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x8-minmax-neon-mlal-lane.c"
++    "src/src/qs8-dwconv/gen/qs8-dwconv-25p16c-minmax-rndnu-neon-mla8-ld64.c",
++    "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qs8-dwconv/gen/qs8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
++    "src/src/qs8-dwconv/gen/qs8-dwconv-25p8c-minmax-rndnu-neon-mla8-ld64.c",
++    "src/src/qs8-dwconv/gen/qs8-dwconv-9p16c-minmax-rndnu-neon-mla8-ld64.c",
++    "src/src/qs8-dwconv/gen/qs8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -55183,12 +57075,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-igemm_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("qu8-igemm_sse4.1-no-sse4.2_standalone") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -55199,8 +57089,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-avx-ld128.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-2x4c8-minmax-fp32-avx-ld128.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-sse41-ld64.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-3x4c8-minmax-fp32-sse41-ld64.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -55216,7 +57106,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qc8w-igemm_arm64_standalone") {
++source_set("qs8-dwconv_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -55228,30 +57118,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x2-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8-minmax-neon-mlal-lane.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c2s4-minmax-neon-mlal.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-2x2-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-2x8c2s4-minmax-neon-mlal.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x8-minmax-neon-mlal-lane.c"
++    "src/src/qs8-dwconv/gen/qs8-dwconv-25p16c-minmax-rndnu-neon-mla8-ld64.c",
++    "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qs8-dwconv/gen/qs8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
++    "src/src/qs8-dwconv/gen/qs8-dwconv-25p8c-minmax-rndnu-neon-mla8-ld64.c",
++    "src/src/qs8-dwconv/gen/qs8-dwconv-9p16c-minmax-rndnu-neon-mla8-ld64.c",
++    "src/src/qs8-dwconv/gen/qs8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
  
--  source_set("qu8-igemm_f16c-fma-avx2") {
--    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
--    ]
+-  source_set("qu8-igemm_x64") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-igemm/gen/qu8-igemm-1x8c8-minmax-fp32-avx2.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-3x8c8-minmax-fp32-avx2.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -55279,26 +57164,21 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-igemm_f16c-fma-avx2_standalone") {
--    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
--    ]
-+source_set("qp8-f32-qb4w-gemm_arch=armv8.2-a+dotprod") {
+-  source_set("qu8-igemm_x64_standalone") {
+-    cflags = []
++source_set("qs8-f16-vcvt_arch=armv8.2-a+fp16") {
 +  cflags = [
-+    "-march=armv8.2-a+dotprod"
++    "-march=armv8.2-a+fp16"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-igemm/gen/qu8-igemm-1x8c8-minmax-fp32-avx2.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-3x8c8-minmax-fp32-avx2.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-1x4c16s2-aarch64-neondot.c",
-+    "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-1x8c16s2-aarch64-neondot.c"
++    "src/src/qs8-f16-vcvt/gen/qs8-f16-vcvt-neonfp16arith-u32.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -55327,29 +57207,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set("qu8-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
--    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-igemm/gen/qu8-igemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-7x16c8-minmax-fp32-avx512skx-prfm.c",
+-  source_set("qu8-rdsum_ssse3-no-sse4.1") {
+-    cflags = [
+-      "-mno-sse4.1",
+-      "-mssse3",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qp8-f32-qb4w-gemm_arch=armv8.2-a+dotprod_standalone") {
++source_set("qs8-f16-vcvt_arch=armv8.2-a+fp16_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+dotprod"
++    "-march=armv8.2-a+fp16"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-rdsum/gen/qu8-rdsum-7p7x-ssse3-c64.c",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-f16-vcvt/gen/qs8-f16-vcvt-neonfp16arith-u32.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -55362,30 +57240,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-1x4c16s2-aarch64-neondot.c",
-+    "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-1x8c16s2-aarch64-neondot.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set(
--      "qu8-igemm_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
--    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -55393,12 +57253,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-  # This is a target that cannot depend on //base.
+-  source_set("qu8-rdsum_ssse3-no-sse4.1_standalone") {
+-    cflags = [
+-      "-mno-sse4.1",
+-      "-mssse3",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-igemm/gen/qu8-igemm-1x16c8-minmax-fp32-avx512skx-prfm.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-7x16c8-minmax-fp32-avx512skx-prfm.c",
+-      "src/src/qu8-rdsum/gen/qu8-rdsum-7p7x-ssse3-c64.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -55410,15 +57280,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++source_set("qs8-f32-vcvt_arm64") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("qp8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -55427,22 +57292,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-16x4c16s2-mstep4-neoni8mm.c"
++    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-neon-u32.c",
++    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u4.c"
 +  ]
  
--  source_set("qu8-igemm_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("qu8-rdsum_x64") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-sse2-ld64.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-3x4c8-minmax-fp32-sse2-ld64.c",
+-      "src/src/qu8-rdsum/gen/qu8-rdsum-scalar.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -55467,30 +57329,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("qp8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
++source_set("qs8-f32-vcvt_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
-+  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-igemm_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-16x4c16s2-mstep4-neoni8mm.c"
+-  source_set("qu8-rdsum_x64_standalone") {
+-    cflags = []
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-sse2-ld64.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-3x4c8-minmax-fp32-sse2-ld64.c",
+-      "src/src/qu8-rdsum/gen/qu8-rdsum-scalar.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-neon-u32.c",
++    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u4.c"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -55502,6 +57357,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -55520,26 +57379,19 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-igemm_sse4.1-no-sse4.2") {
+-  source_set("qu8-rsum_f16c-fma-avx2") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+dotprod") {
++source_set("qs8-packw_arm64") {
 +  cflags = [
-+    "-march=armv8.2-a+dotprod"
-+  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-sse41-ld64.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-3x4c8-minmax-fp32-sse41-ld64.c",
+-      "src/src/qu8-rsum/gen/qu8-rsum-avx2-u64-acc2.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-16x4c8s2-mstep4-aarch64-neondot.c",
-+    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x4c8s2-aarch64-neondot.c",
-+    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x8c16s2-aarch64-neondot.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -55552,12 +57404,24 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-packw/gen/qs8-packw-x16c8-gemm-goi-scalar.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("qu8-rsum_f16c-fma-avx2_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -55565,25 +57429,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qu8-igemm_sse4.1-no-sse4.2_standalone") {
--    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-igemm/gen/qu8-igemm-1x4c8-minmax-fp32-sse41-ld64.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-3x4c8-minmax-fp32-sse41-ld64.c",
+-      "src/src/qu8-rsum/gen/qu8-rsum-avx2-u64-acc2.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+dotprod_standalone") {
-+  cflags = [
-+    "-march=armv8.2-a+dotprod"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -55595,32 +57446,34 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++# This is a target that cannot depend on //base.
++source_set("qs8-packw_arm64_standalone") {
++  cflags = [
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-16x4c8s2-mstep4-aarch64-neondot.c",
-+    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x4c8s2-aarch64-neondot.c",
-+    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x8c16s2-aarch64-neondot.c"
++    "src/src/qs8-packw/gen/qs8-packw-x16c8-gemm-goi-scalar.c"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
+-  source_set("qu8-rsum_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
--
--  source_set("qu8-igemm_x64") {
--    cflags = []
--
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-igemm/gen/qu8-igemm-1x2-minmax-fp32-scalar-imagic.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-2x2-minmax-fp32-scalar-imagic.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qu8-rsum/gen/qu8-rsum-sse2-u32-acc2.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -55648,17 +57501,21 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-igemm_x64_standalone") {
--    cflags = []
+-  source_set("qu8-rsum_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-igemm/gen/qu8-igemm-1x2-minmax-fp32-scalar-imagic.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-2x2-minmax-fp32-scalar-imagic.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qu8-rsum/gen/qu8-rsum-sse2-u32-acc2.c",
 -    ]
--
++source_set("qs8-qc4w-gemm_arch=armv8.2-a+fp16+dotprod") {
++  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -55669,35 +57526,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
--
++  asmflags = cflags
+ 
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+i8mm+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-1x16-minmax-fp32-asm-aarch64-neondot-ld128.S",
++    "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-5x16-minmax-fp32-asm-aarch64-neondot-ld128.S"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-8x8c16s2-mstep2-neoni8mm.c"
-+  ]
- 
--  source_set("qu8-rdsum_ssse3-no-sse4.1") {
--    cflags = [
--      "-mno-sse4.1",
--      "-mssse3",
--    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-rdsum/gen/qu8-rdsum-7p7x-ssse3-c64.c",
--    ]
+-  source_set("qu8-rsum_x64") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -55705,6 +57552,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-rsum/gen/qu8-rsum-scalar-u4.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -55715,31 +57569,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +# This is a target that cannot depend on //base.
-+source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
++source_set("qs8-qc4w-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
++    "-march=armv8.2-a+fp16+dotprod"
 +  ]
  
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  asmflags = cflags
+ 
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-rdsum_ssse3-no-sse4.1_standalone") {
--    cflags = [
--      "-mno-sse4.1",
--      "-mssse3",
--    ]
+-  source_set("qu8-rsum_x64_standalone") {
+-    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-8x8c16s2-mstep2-neoni8mm.c"
++    "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-1x16-minmax-fp32-asm-aarch64-neondot-ld128.S",
++    "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-5x16-minmax-fp32-asm-aarch64-neondot-ld128.S"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-rdsum/gen/qu8-rdsum-7p7x-ssse3-c64.c",
+-      "src/src/qu8-rsum/gen/qu8-rsum-scalar-u4.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -55773,22 +57624,20 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-rdsum_x64") {
--    cflags = []
-+source_set("qp8-f32-qc8w-gemm_arch=armv8.2-a+dotprod") {
+-  source_set("qu8-vadd_avx-no-avx2-no-f16c-no-fma") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
++source_set("qs8-qc4w-gemm_arm64") {
 +  cflags = [
-+    "-march=armv8.2-a+dotprod"
-+  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-rdsum/gen/qu8-rdsum-scalar.c",
+-      "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx-mul32-ld32-u8.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-16x4c4-mstep4-aarch64-neondot.c",
-+    "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-1x4c4-aarch64-neondot.c",
-+    "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-1x4c8-aarch64-neondot.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -55801,12 +57650,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-1x2-minmax-fp32-scalar-lrintf.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("qu8-vadd_avx-no-avx2-no-f16c-no-fma_standalone") {
+-    cflags = [
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -55814,20 +57676,92 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qu8-rdsum_x64_standalone") {
--    cflags = []
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx-mul32-ld32-u8.c",
+-    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("qs8-qc4w-gemm_arm64_standalone") {
++  cflags = [
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-1x2-minmax-fp32-scalar-lrintf.c"
++  ]
+ 
+-  source_set("qu8-vadd_f16c-fma-avx2") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-rdsum/gen/qu8-rdsum-scalar.c",
+-      "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx2-mul32-ld64-u16.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qp8-f32-qc8w-gemm_arch=armv8.2-a+dotprod_standalone") {
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qu8-vadd_f16c-fma-avx2_standalone") {
+-    cflags = [
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
+-    ]
++source_set("qs8-qc8w-dwconv_arm64") {
 +  cflags = [
-+    "-march=armv8.2-a+dotprod"
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx2-mul32-ld64-u16.c",
+-    ]
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -55842,9 +57776,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-16x4c4-mstep4-aarch64-neondot.c",
-+    "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-1x4c4-aarch64-neondot.c",
-+    "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-1x4c8-aarch64-neondot.c"
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p16c-minmax-fp32-neonv8-mla8-ld64.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-neon-mla8-ld64.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-neonv8-mla8-ld64.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-neon-mla8-ld128.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-neonv8-mla8-ld128.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-neonv8-mla8-ld64.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-lrintf.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-neon-mla8-ld64.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-neonv8-mla8-ld64.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -55856,30 +57801,126 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
+-
+-  source_set("qu8-vadd_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-    cflags = [
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
+-      "-mf16c",
+-      "-mfma",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
--  source_set("qu8-rsum_f16c-fma-avx2") {
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx512skx-mul32-ld128-u16.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("qs8-qc8w-dwconv_arm64_standalone") {
++  cflags = [
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set(
+-      "qu8-vadd_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
--      "-mavx2",
+-      "-mavx512bw",
+-      "-mavx512cd",
+-      "-mavx512dq",
+-      "-mavx512f",
+-      "-mavx512vl",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p16c-minmax-fp32-neonv8-mla8-ld64.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-neon-mla8-ld64.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-neonv8-mla8-ld64.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-neon-mla8-ld128.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-neonv8-mla8-ld128.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-neonv8-mla8-ld64.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-lrintf.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-neon-mla8-ld64.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-neonv8-mla8-ld64.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-rsum/gen/qu8-rsum-avx2-u64-acc2.c",
+-      "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx512skx-mul32-ld128-u16.c",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
-+  }
+   }
 +}
  
+-  source_set("qu8-vadd_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-vadd/gen/qu8-vadd-minmax-sse2-mul16-ld64-u8.c",
+-    ]
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -55890,24 +57931,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("qp8-f32-qc8w-gemm_arch=armv8.2-a+i8mm+fp16") {
++source_set("qs8-qc8w-gemm_arch=armv8.2-a+dotprod") {
 +  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
++    "-march=armv8.2-a+dotprod"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-16x4c8-mstep4-neoni8mm.c"
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16c4-minmax-fp32-neondot.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c4-minmax-fp32-neondot.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16c4-minmax-fp32-neondot.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x8c4-minmax-fp32-neondot.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-rsum_f16c-fma-avx2_standalone") {
+-  source_set("qu8-vadd_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -55915,7 +57958,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-rsum/gen/qu8-rsum-avx2-u64-acc2.c",
+-      "src/src/qu8-vadd/gen/qu8-vadd-minmax-sse2-mul16-ld64-u8.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -55939,9 +57982,9 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("qp8-f32-qc8w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
++source_set("qs8-qc8w-gemm_arch=armv8.2-a+dotprod_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
++    "-march=armv8.2-a+dotprod"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -55950,13 +57993,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-16x4c8-mstep4-neoni8mm.c"
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16c4-minmax-fp32-neondot.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c4-minmax-fp32-neondot.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16c4-minmax-fp32-neondot.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x8c4-minmax-fp32-neondot.c"
 +  ]
  
--  source_set("qu8-rsum_sse2-no-sse3") {
+-  source_set("qu8-vadd_sse4.1-no-sse4.2") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -55964,7 +58010,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-rsum/gen/qu8-rsum-sse2-u32-acc2.c",
+-      "src/src/qu8-vadd/gen/qu8-vadd-minmax-sse41-mul16-ld64-u8.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -55992,18 +58038,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-rsum_sse2-no-sse3_standalone") {
+-  source_set("qu8-vadd_sse4.1-no-sse4.2_standalone") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-rsum/gen/qu8-rsum-sse2-u32-acc2.c",
+-      "src/src/qu8-vadd/gen/qu8-vadd-minmax-sse41-mul16-ld64-u8.c",
 -    ]
-+source_set("qs8-dwconv_arm64") {
++source_set("qs8-qc8w-gemm_arch=armv8.2-a+fp16+dotprod") {
 +  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -56015,20 +58063,21 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
++  asmflags = cflags
  
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-25p16c-minmax-rndnu-neon-mla8-ld64.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-25p8c-minmax-rndnu-neon-mla8-ld64.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-9p16c-minmax-rndnu-neon-mla8-ld64.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal-cortex-a53-prfm.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal-prfm.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal-cortex-a53-prfm.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal-prfm.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16-minmax-fp32-asm-aarch64-neon-mlal-lane-cortex-a53.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16-minmax-fp32-asm-aarch64-neon-mlal-lane-ld64.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16c4-minmax-fp32-asm-aarch64-neondot-cortex-a55.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16c4-minmax-fp32-asm-aarch64-neondot-ld128.S"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -56039,7 +58088,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qu8-rsum_x64") {
+-  source_set("qu8-vadd_x64") {
 -    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -56050,7 +58099,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-rsum/gen/qu8-rsum-scalar-u4.c",
+-      "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u1.c",
+-      "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u4.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -56066,32 +58116,36 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-dwconv_arm64_standalone") {
++source_set("qs8-qc8w-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
 +  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  ]
++  asmflags = cflags
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-rsum_x64_standalone") {
+-  source_set("qu8-vadd_x64_standalone") {
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-25p16c-minmax-rndnu-neon-mla8-ld64.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-25p8c-minmax-rndnu-neon-mla8-ld64.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-9p16c-minmax-rndnu-neon-mla8-ld64.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal-cortex-a53-prfm.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal-prfm.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal-cortex-a53-prfm.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal-prfm.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16-minmax-fp32-asm-aarch64-neon-mlal-lane-cortex-a53.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16-minmax-fp32-asm-aarch64-neon-mlal-lane-ld64.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16c4-minmax-fp32-asm-aarch64-neondot-cortex-a55.S",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16c4-minmax-fp32-asm-aarch64-neondot-ld128.S"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-rsum/gen/qu8-rsum-scalar-u4.c",
+-      "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u1.c",
+-      "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u4.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -56125,7 +58179,7 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-vadd_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qu8-vaddc_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
@@ -56135,7 +58189,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx-mul32-ld32-u8.c",
+-      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx-mul32-ld32-u8.c",
 -    ]
 -
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -56148,20 +58202,21 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("qs8-f16-vcvt_arch=armv8.2-a+fp16") {
++source_set("qs8-qc8w-gemm_arch=armv8.2-a+i8mm+fp16") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16"
++    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-f16-vcvt/gen/qs8-f16-vcvt-neonfp16arith-u32.c"
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16c8-minmax-fp32-neoni8mm.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16c8-minmax-fp32-neoni8mm.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vadd_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("qu8-vaddc_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
@@ -56174,7 +58229,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx-mul32-ld32-u8.c",
+-      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx-mul32-ld32-u8.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -56198,9 +58253,9 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-f16-vcvt_arch=armv8.2-a+fp16_standalone") {
++source_set("qs8-qc8w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16"
++    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -56209,10 +58264,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-f16-vcvt/gen/qs8-f16-vcvt-neonfp16arith-u32.c"
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16c8-minmax-fp32-neoni8mm.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16c8-minmax-fp32-neoni8mm.c"
 +  ]
  
--  source_set("qu8-vadd_f16c-fma-avx2") {
+-  source_set("qu8-vaddc_f16c-fma-avx2") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
@@ -56224,7 +58280,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx2-mul32-ld64-u16.c",
+-      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx2-mul32-ld64-u16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -56252,18 +58308,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vadd_f16c-fma-avx2_standalone") {
+-  source_set("qu8-vaddc_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("qs8-f32-vcvt_arm64") {
++source_set("qs8-qc8w-gemm_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx2-mul32-ld64-u16.c",
+-      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx2-mul32-ld64-u16.c",
 -    ]
 +  ]
  
@@ -56279,9 +58335,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-neon-u32.c",
-+    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u1.c",
-+    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u4.c"
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16-minmax-fp32-neonv8-mlal-lane.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c2s4-minmax-fp32-neon-mlal.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c2s4-minmax-fp32-neonv8-mlal.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c2s4-minmax-fp32-neon-mlal.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c2s4-minmax-fp32-neonv8-mlal.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -56294,7 +58354,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    }
 -  }
 -
--  source_set("qu8-vadd_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
+-  source_set("qu8-vaddc_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
@@ -56313,7 +58373,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx512skx-mul32-ld128-u16.c",
+-      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx512skx-mul32-ld128-u16.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -56329,7 +58389,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-f32-vcvt_arm64_standalone") {
++source_set("qs8-qc8w-gemm_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -56338,7 +58398,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
--      "qu8-vadd_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-      "qu8-vaddc_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
 -    cflags = [
 -      "-mavx512bw",
 -      "-mavx512cd",
@@ -56350,14 +58410,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-neon-u32.c",
-+    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u1.c",
-+    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u4.c"
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16-minmax-fp32-neonv8-mlal-lane.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c2s4-minmax-fp32-neon-mlal.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c2s4-minmax-fp32-neonv8-mlal.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c2s4-minmax-fp32-neon-mlal.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c2s4-minmax-fp32-neonv8-mlal.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vadd/gen/qu8-vadd-minmax-avx512skx-mul32-ld128-u16.c",
+-      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx512skx-mul32-ld128-u16.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -56391,20 +58455,17 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-vadd_sse2-no-sse3") {
+-  source_set("qu8-vaddc_sse2-no-sse3") {
 -    cflags = [
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+source_set("qs8-packw_arm64") {
-+  cflags = [
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vadd/gen/qu8-vadd-minmax-sse2-mul16-ld64-u8.c",
+-      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-sse2-mul16-ld64-u8.c",
 -    ]
-+  ]
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -56415,23 +58476,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-packw/gen/qs8-packw-x16c8-gemm-goi-scalar.c"
++source_set("qs8-qc8w-igemm_arch=armv8.2-a+dotprod") {
++  cflags = [
++    "-march=armv8.2-a+dotprod"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16c4-minmax-fp32-neondot.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c4-minmax-fp32-neondot.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16c4-minmax-fp32-neondot.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x8c4-minmax-fp32-neondot.c"
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vadd_sse2-no-sse3_standalone") {
+-  source_set("qu8-vaddc_sse2-no-sse3_standalone") {
 -    cflags = [
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-sse2-mul16-ld64-u8.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -56439,13 +58512,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-vadd/gen/qu8-vadd-minmax-sse2-mul16-ld64-u8.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -56456,11 +58522,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qs8-packw_arm64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
++# This is a target that cannot depend on //base.
++source_set("qs8-qc8w-igemm_arch=armv8.2-a+dotprod_standalone") {
++  cflags = [
++    "-march=armv8.2-a+dotprod"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -56469,10 +58538,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-packw/gen/qs8-packw-x16c8-gemm-goi-scalar.c"
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16c4-minmax-fp32-neondot.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c4-minmax-fp32-neondot.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16c4-minmax-fp32-neondot.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x8c4-minmax-fp32-neondot.c"
 +  ]
  
--  source_set("qu8-vadd_sse4.1-no-sse4.2") {
+-  source_set("qu8-vaddc_sse4.1-no-sse4.2") {
 -    cflags = [
 -      "-mno-sse4.2",
 -      "-msse4.1",
@@ -56483,7 +58555,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vadd/gen/qu8-vadd-minmax-sse41-mul16-ld64-u8.c",
+-      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-sse41-mul16-ld64-u8.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -56511,7 +58583,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vadd_sse4.1-no-sse4.2_standalone") {
+-  source_set("qu8-vaddc_sse4.1-no-sse4.2_standalone") {
 -    cflags = [
 -      "-mno-sse4.2",
 -      "-msse4.1",
@@ -56519,9 +58591,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vadd/gen/qu8-vadd-minmax-sse41-mul16-ld64-u8.c",
+-      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-sse41-mul16-ld64-u8.c",
 -    ]
-+source_set("qs8-qc4w-gemm_arch=armv8.2-a+fp16+dotprod") {
++source_set("qs8-qc8w-igemm_arch=armv8.2-a+fp16+dotprod") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16+dotprod"
 +  ]
@@ -56541,8 +58613,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-1x16-minmax-fp32-asm-aarch64-neondot-ld128.S",
-+    "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-5x16-minmax-fp32-asm-aarch64-neondot-ld128.S"
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal-cortex-a53-prfm.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal-prfm.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal-cortex-a53-prfm.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal-prfm.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16-minmax-fp32-asm-aarch64-neon-mlal-lane-cortex-a53.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16-minmax-fp32-asm-aarch64-neon-mlal-lane-ld64.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16c4-minmax-fp32-asm-aarch64-neondot-cortex-a55.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16c4-minmax-fp32-asm-aarch64-neondot-ld128.S"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -56553,7 +58633,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qu8-vadd_x64") {
+-  source_set("qu8-vaddc_x64") {
 -    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -56564,8 +58644,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u1.c",
--      "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u4.c",
+-      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
+-      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -56581,7 +58661,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-qc4w-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
++source_set("qs8-qc8w-igemm_arch=armv8.2-a+fp16+dotprod_standalone") {
 +  cflags = [
 +    "-march=armv8.2-a+fp16+dotprod"
 +  ]
@@ -56591,18 +58671,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  asmflags = cflags
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vadd_x64_standalone") {
+-  source_set("qu8-vaddc_x64_standalone") {
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-1x16-minmax-fp32-asm-aarch64-neondot-ld128.S",
-+    "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-5x16-minmax-fp32-asm-aarch64-neondot-ld128.S"
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal-cortex-a53-prfm.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal-prfm.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal-cortex-a53-prfm.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal-prfm.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16-minmax-fp32-asm-aarch64-neon-mlal-lane-cortex-a53.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16-minmax-fp32-asm-aarch64-neon-mlal-lane-ld64.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16c4-minmax-fp32-asm-aarch64-neondot-cortex-a55.S",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16c4-minmax-fp32-asm-aarch64-neondot-ld128.S"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u1.c",
--      "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u4.c",
+-      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
+-      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -56636,22 +58724,19 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-vaddc_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qu8-vcvt_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
-+source_set("qs8-qc4w-gemm_arm64") {
-+  cflags = [
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx-mul32-ld32-u8.c",
+-      "src/src/qu8-vcvt/gen/qu8-vcvt-avx-u32.c",
 -    ]
-+  ]
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -56662,25 +58747,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-1x2-minmax-fp32-scalar-lrintf.c"
++source_set("qs8-qc8w-igemm_arch=armv8.2-a+i8mm+fp16") {
++  cflags = [
++    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16c8-minmax-fp32-neoni8mm.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16c8-minmax-fp32-neoni8mm.c"
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vaddc_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("qu8-vcvt_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-vcvt/gen/qu8-vcvt-avx-u32.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -56688,13 +58783,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx-mul32-ld32-u8.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -56705,11 +58793,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qs8-qc4w-gemm_arm64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
++# This is a target that cannot depend on //base.
++source_set("qs8-qc8w-igemm_arch=armv8.2-a+i8mm+fp16_standalone") {
++  cflags = [
++    "-march=armv8.2-a+i8mm+fp16"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -56718,10 +58809,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-1x2-minmax-fp32-scalar-lrintf.c"
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16c8-minmax-fp32-neoni8mm.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16c8-minmax-fp32-neoni8mm.c"
 +  ]
  
--  source_set("qu8-vaddc_f16c-fma-avx2") {
+-  source_set("qu8-vcvt_f16c-fma-avx2") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
@@ -56733,7 +58825,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx2-mul32-ld64-u16.c",
+-      "src/src/qu8-vcvt/gen/qu8-vcvt-avx2-u32.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -56761,20 +58853,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vaddc_f16c-fma-avx2_standalone") {
+-  source_set("qu8-vcvt_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("qs8-qc8w-dwconv_arm64") {
-+  cflags = [
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx2-mul32-ld64-u16.c",
+-      "src/src/qu8-vcvt/gen/qu8-vcvt-avx2-u32.c",
 -    ]
-+  ]
++source_set("qs8-qc8w-igemm_arm64") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -56786,46 +58877,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p16c-minmax-fp32-neonv8-mla8-ld64.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-neon-mla8-ld64.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-neonv8-mla8-ld64.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-neon-mla8-ld128.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-neonv8-mla8-ld128.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-neonv8-mla8-ld64.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-neon-mla8-ld64.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-neonv8-mla8-ld64.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16-minmax-fp32-neonv8-mlal-lane.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c2s4-minmax-fp32-neon-mlal.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c2s4-minmax-fp32-neonv8-mlal.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c2s4-minmax-fp32-neon-mlal.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c2s4-minmax-fp32-neonv8-mlal.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4-minmax-fp32-scalar-lrintf.c"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set("qu8-vaddc_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl") {
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  source_set("qu8-vcvt_sse2-no-sse3") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -56836,7 +58913,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx512skx-mul32-ld128-u16.c",
+-      "src/src/qu8-vcvt/gen/qu8-vcvt-sse2-u32.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -56852,7 +58929,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-qc8w-dwconv_arm64_standalone") {
++source_set("qs8-qc8w-igemm_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -56860,41 +58937,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set(
--      "qu8-vaddc_f16c-fma-avx512f-avx512cd-avx512bw-avx512dq-avx512vl_standalone") {
+-  source_set("qu8-vcvt_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mavx512bw",
--      "-mavx512cd",
--      "-mavx512dq",
--      "-mavx512f",
--      "-mavx512vl",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p16c-minmax-fp32-neonv8-mla8-ld64.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-neon-mla8-ld64.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-neonv8-mla8-ld64.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-neon-mla8-ld128.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-neonv8-mla8-ld128.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-neonv8-mla8-ld64.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-neon-mla8-ld64.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-neonv8-mla8-ld64.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16-minmax-fp32-neonv8-mlal-lane.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c2s4-minmax-fp32-neon-mlal.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c2s4-minmax-fp32-neonv8-mlal.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c2s4-minmax-fp32-neon-mlal.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c2s4-minmax-fp32-neonv8-mlal.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-avx512skx-mul32-ld128-u16.c",
+-      "src/src/qu8-vcvt/gen/qu8-vcvt-sse2-u32.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -56928,17 +58989,20 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-vaddc_sse2-no-sse3") {
+-  source_set("qu8-vcvt_sse4.1-no-sse4.2") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
--
++source_set("qs8-qu8-packw_arm64") {
++  cflags = [
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-sse2-mul16-ld64-u8.c",
+-      "src/src/qu8-vcvt/gen/qu8-vcvt-sse41-u32.c",
 -    ]
--
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -56949,34 +59013,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("qs8-qc8w-gemm_arch=armv8.2-a+dotprod") {
-+  cflags = [
-+    "-march=armv8.2-a+dotprod"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16c4-minmax-fp32-neondot.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c4-minmax-fp32-neondot.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16c4-minmax-fp32-neondot.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x8c4-minmax-fp32-neondot.c"
++    "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-scalar.c"
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qu8-vaddc_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-sse2-mul16-ld64-u8.c",
+-  # This is a target that cannot depend on //base.
+-  source_set("qu8-vcvt_sse4.1-no-sse4.2_standalone") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -56985,6 +59037,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-vcvt/gen/qu8-vcvt-sse41-u32.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -56995,14 +59054,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-qc8w-gemm_arch=armv8.2-a+dotprod_standalone") {
++source_set("qs8-qu8-packw_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+dotprod"
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -57011,16 +59067,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16c4-minmax-fp32-neondot.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c4-minmax-fp32-neondot.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16c4-minmax-fp32-neondot.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x8c4-minmax-fp32-neondot.c"
++    "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-scalar.c"
 +  ]
  
--  source_set("qu8-vaddc_sse4.1-no-sse4.2") {
+-  source_set("qu8-vcvt_ssse3-no-sse4.1") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mno-sse4.1",
+-      "-mssse3",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -57028,7 +59081,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-sse41-mul16-ld64-u8.c",
+-      "src/src/qu8-vcvt/gen/qu8-vcvt-ssse3-u32.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -57056,20 +59109,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vaddc_sse4.1-no-sse4.2_standalone") {
+-  source_set("qu8-vcvt_ssse3-no-sse4.1_standalone") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mno-sse4.1",
+-      "-mssse3",
 -    ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-sse41-mul16-ld64-u8.c",
+-      "src/src/qu8-vcvt/gen/qu8-vcvt-ssse3-u32.c",
 -    ]
-+source_set("qs8-qc8w-gemm_arch=armv8.2-a+fp16+dotprod") {
++source_set("qs8-rdsum_arm64") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -57081,21 +59132,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  asmflags = cflags
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal-cortex-a53-prfm.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal-prfm.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal-cortex-a53-prfm.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal-prfm.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16-minmax-fp32-asm-aarch64-neon-mlal-lane-cortex-a53.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16-minmax-fp32-asm-aarch64-neon-mlal-lane-ld64.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16c4-minmax-fp32-asm-aarch64-neondot-cortex-a55.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16c4-minmax-fp32-asm-aarch64-neondot-ld128.S"
++    "src/src/qs8-rdsum/gen/qs8-rdsum-7p7x-minmax-fp32-neon-c32.c",
++    "src/src/qs8-rdsum/gen/qs8-rdsum-minmax-fp32-scalar-u1-acc1.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -57106,7 +59149,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qu8-vaddc_x64") {
+-  source_set("qu8-vcvt_x64") {
 -    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -57117,8 +59160,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
--      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c",
+-      "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -57134,36 +59176,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-qc8w-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
++source_set("qs8-rdsum_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  asmflags = cflags
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vaddc_x64_standalone") {
+-  source_set("qu8-vcvt_x64_standalone") {
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal-cortex-a53-prfm.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal-prfm.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal-cortex-a53-prfm.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal-prfm.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16-minmax-fp32-asm-aarch64-neon-mlal-lane-cortex-a53.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16-minmax-fp32-asm-aarch64-neon-mlal-lane-ld64.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16c4-minmax-fp32-asm-aarch64-neondot-cortex-a55.S",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16c4-minmax-fp32-asm-aarch64-neondot-ld128.S"
++    "src/src/qs8-rdsum/gen/qs8-rdsum-7p7x-minmax-fp32-neon-c32.c",
++    "src/src/qs8-rdsum/gen/qs8-rdsum-minmax-fp32-scalar-u1-acc1.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
--      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c",
+-      "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -57197,7 +59228,7 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-vcvt_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qu8-vlrelu_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
@@ -57207,7 +59238,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-avx-u32.c",
+-      "src/src/qu8-vlrelu/gen/qu8-vlrelu-avx-u32.c",
 -    ]
 -
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -57220,21 +59251,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("qs8-qc8w-gemm_arch=armv8.2-a+i8mm+fp16") {
++source_set("qs8-rsum_arch=armv8.2-a+dotprod") {
 +  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
++    "-march=armv8.2-a+dotprod"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16c8-minmax-fp32-neoni8mm.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16c8-minmax-fp32-neoni8mm.c"
++    "src/src/qs8-rsum/gen/qs8-rsum-neondot-u32-acc2.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vcvt_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("qu8-vlrelu_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
@@ -57247,7 +59277,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-avx-u32.c",
+-      "src/src/qu8-vlrelu/gen/qu8-vlrelu-avx-u32.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -57271,9 +59301,9 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-qc8w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
++source_set("qs8-rsum_arch=armv8.2-a+dotprod_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
++    "-march=armv8.2-a+dotprod"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -57282,11 +59312,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16c8-minmax-fp32-neoni8mm.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-4x16c8-minmax-fp32-neoni8mm.c"
++    "src/src/qs8-rsum/gen/qs8-rsum-neondot-u32-acc2.c"
 +  ]
  
--  source_set("qu8-vcvt_f16c-fma-avx2") {
+-  source_set("qu8-vlrelu_f16c-fma-avx2") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
@@ -57298,7 +59327,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-avx2-u32.c",
+-      "src/src/qu8-vlrelu/gen/qu8-vlrelu-avx2-u32.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -57326,7 +59355,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vcvt_f16c-fma-avx2_standalone") {
+-  source_set("qu8-vlrelu_f16c-fma-avx2_standalone") {
 -    cflags = [
 -      "-mavx2",
 -      "-mf16c",
@@ -57335,9 +59364,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-avx2-u32.c",
+-      "src/src/qu8-vlrelu/gen/qu8-vlrelu-avx2-u32.c",
 -    ]
-+source_set("qs8-qc8w-gemm_arm64") {
++source_set("qs8-rsum_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -57355,15 +59384,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16-minmax-fp32-neonv8-mlal-lane.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c2s4-minmax-fp32-neon-mlal.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c2s4-minmax-fp32-neonv8-mlal.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c2s4-minmax-fp32-neon-mlal.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c2s4-minmax-fp32-neonv8-mlal.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4-minmax-fp32-scalar-lrintf.c"
++    "src/src/qs8-rsum/gen/qs8-rsum-neon-u32-acc2.c",
++    "src/src/qs8-rsum/gen/qs8-rsum-scalar-u4.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -57374,7 +59396,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qu8-vcvt_sse2-no-sse3") {
+-  source_set("qu8-vlrelu_sse2-no-sse3") {
 -    cflags = [
 -      "-mno-sse3",
 -      "-msse2",
@@ -57388,7 +59410,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-sse2-u32.c",
+-      "src/src/qu8-vlrelu/gen/qu8-vlrelu-sse2-u32.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -57404,7 +59426,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-qc8w-gemm_arm64_standalone") {
++source_set("qs8-rsum_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -57412,27 +59434,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vcvt_sse2-no-sse3_standalone") {
+-  source_set("qu8-vlrelu_sse2-no-sse3_standalone") {
 -    cflags = [
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16-minmax-fp32-neonv8-mlal-lane.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c2s4-minmax-fp32-neon-mlal.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c2s4-minmax-fp32-neonv8-mlal.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c2s4-minmax-fp32-neon-mlal.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c2s4-minmax-fp32-neonv8-mlal.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4-minmax-fp32-scalar-lrintf.c"
++    "src/src/qs8-rsum/gen/qs8-rsum-neon-u32-acc2.c",
++    "src/src/qs8-rsum/gen/qs8-rsum-scalar-u4.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-sse2-u32.c",
+-      "src/src/qu8-vlrelu/gen/qu8-vlrelu-sse2-u32.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -57466,17 +59481,20 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-vcvt_sse4.1-no-sse4.2") {
+-  source_set("qu8-vlrelu_sse4.1-no-sse4.2") {
 -    cflags = [
 -      "-mno-sse4.2",
 -      "-msse4.1",
 -    ]
--
++source_set("qs8-vadd_arm64") {
++  cflags = [
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-sse41-u32.c",
+-      "src/src/qu8-vlrelu/gen/qu8-vlrelu-sse41-u32.c",
 -    ]
--
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -57487,35 +59505,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("qs8-qc8w-igemm_arch=armv8.2-a+dotprod") {
-+  cflags = [
-+    "-march=armv8.2-a+dotprod"
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-vadd/gen/qs8-vadd-minmax-neon-ld64-u16.c",
++    "src/src/qs8-vadd/gen/qs8-vadd-minmax-neon-ld64-u32.c",
++    "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u1.c",
++    "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16c4-minmax-fp32-neondot.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c4-minmax-fp32-neondot.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16c4-minmax-fp32-neondot.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x8c4-minmax-fp32-neondot.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vcvt_sse4.1-no-sse4.2_standalone") {
+-  source_set("qu8-vlrelu_sse4.1-no-sse4.2_standalone") {
 -    cflags = [
 -      "-mno-sse4.2",
 -      "-msse4.1",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-sse41-u32.c",
--    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -57523,6 +59532,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-vlrelu/gen/qu8-vlrelu-sse41-u32.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -57533,14 +59549,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-qc8w-igemm_arch=armv8.2-a+dotprod_standalone") {
++source_set("qs8-vadd_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+dotprod"
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -57549,13 +59562,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16c4-minmax-fp32-neondot.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c4-minmax-fp32-neondot.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16c4-minmax-fp32-neondot.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x8c4-minmax-fp32-neondot.c"
++    "src/src/qs8-vadd/gen/qs8-vadd-minmax-neon-ld64-u16.c",
++    "src/src/qs8-vadd/gen/qs8-vadd-minmax-neon-ld64-u32.c",
++    "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u1.c",
++    "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u4.c"
 +  ]
  
--  source_set("qu8-vcvt_ssse3-no-sse4.1") {
+-  source_set("qu8-vlrelu_ssse3-no-sse4.1") {
 -    cflags = [
 -      "-mno-sse4.1",
 -      "-mssse3",
@@ -57566,7 +59579,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-ssse3-u32.c",
+-      "src/src/qu8-vlrelu/gen/qu8-vlrelu-ssse3-u32.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -57594,7 +59607,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vcvt_ssse3-no-sse4.1_standalone") {
+-  source_set("qu8-vlrelu_ssse3-no-sse4.1_standalone") {
 -    cflags = [
 -      "-mno-sse4.1",
 -      "-mssse3",
@@ -57602,12 +59615,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-ssse3-u32.c",
+-      "src/src/qu8-vlrelu/gen/qu8-vlrelu-ssse3-u32.c",
 -    ]
-+source_set("qs8-qc8w-igemm_arch=armv8.2-a+fp16+dotprod") {
++source_set("qs8-vaddc_arm64") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -57619,21 +59630,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  asmflags = cflags
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal-cortex-a53-prfm.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal-prfm.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal-cortex-a53-prfm.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal-prfm.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16-minmax-fp32-asm-aarch64-neon-mlal-lane-cortex-a53.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16-minmax-fp32-asm-aarch64-neon-mlal-lane-ld64.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16c4-minmax-fp32-asm-aarch64-neondot-cortex-a55.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16c4-minmax-fp32-asm-aarch64-neondot-ld128.S"
++    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-neon-ld64-u16.c",
++    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-neon-ld64-u32.c",
++    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u1.c",
++    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u4.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -57644,7 +59649,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qu8-vcvt_x64") {
+-  source_set("qu8-vlrelu_x64") {
 -    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -57655,8 +59660,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u1.c",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c",
+-      "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -57672,36 +59676,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-qc8w-igemm_arch=armv8.2-a+fp16+dotprod_standalone") {
++source_set("qs8-vaddc_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  asmflags = cflags
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vcvt_x64_standalone") {
+-  source_set("qu8-vlrelu_x64_standalone") {
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal-cortex-a53-prfm.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal-prfm.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c8-minmax-fp32-asm-aarch64-neon-mlal.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal-cortex-a53-prfm.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal-prfm.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c8-minmax-fp32-asm-aarch64-neon-mlal.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16-minmax-fp32-asm-aarch64-neon-mlal-lane-cortex-a53.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16-minmax-fp32-asm-aarch64-neon-mlal-lane-ld64.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16c4-minmax-fp32-asm-aarch64-neondot-cortex-a55.S",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16c4-minmax-fp32-asm-aarch64-neondot-ld128.S"
++    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-neon-ld64-u16.c",
++    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-neon-ld64-u32.c",
++    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u1.c",
++    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u4.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u1.c",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c",
+-      "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -57735,19 +59730,22 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-vlrelu_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qu8-vmul_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
--
++source_set("qs8-vcvt_arm64") {
++  cflags = [
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-avx-u32.c",
+-      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-avx-mul16-ld64-u16.c",
 -    ]
--
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -57758,35 +59756,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("qs8-qc8w-igemm_arch=armv8.2-a+i8mm+fp16") {
-+  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-vcvt/gen/qs8-vcvt-neon-u32.c",
++    "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16c8-minmax-fp32-neoni8mm.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16c8-minmax-fp32-neoni8mm.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vlrelu_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("qu8-vmul_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
 -      "-mavx",
 -      "-mno-avx2",
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-avx-u32.c",
--    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -57794,6 +59783,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-avx-mul16-ld64-u16.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -57804,14 +59800,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-qc8w-igemm_arch=armv8.2-a+i8mm+fp16_standalone") {
++source_set("qs8-vcvt_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+i8mm+fp16"
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -57820,15 +59813,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16c8-minmax-fp32-neoni8mm.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-4x16c8-minmax-fp32-neoni8mm.c"
++    "src/src/qs8-vcvt/gen/qs8-vcvt-neon-u32.c",
++    "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c"
 +  ]
  
--  source_set("qu8-vlrelu_f16c-fma-avx2") {
+-  source_set("qu8-vmul_sse2-no-sse3") {
 -    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -57836,7 +59828,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-avx2-u32.c",
+-      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-sse2-mul16-ld64-u8.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -57864,18 +59856,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vlrelu_f16c-fma-avx2_standalone") {
+-  source_set("qu8-vmul_sse2-no-sse3_standalone") {
 -    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
+-      "-mno-sse3",
+-      "-msse2",
 -    ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-avx2-u32.c",
+-      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-sse2-mul16-ld64-u8.c",
 -    ]
-+source_set("qs8-qc8w-igemm_arm64") {
++source_set("qs8-vlrelu_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -57893,15 +59884,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16-minmax-fp32-neonv8-mlal-lane.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c2s4-minmax-fp32-neon-mlal.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c2s4-minmax-fp32-neonv8-mlal.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c2s4-minmax-fp32-neon-mlal.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c2s4-minmax-fp32-neonv8-mlal.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4-minmax-fp32-scalar-lrintf.c"
++    "src/src/qs8-vlrelu/gen/qs8-vlrelu-neon-u32.c",
++    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -57912,10 +59896,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qu8-vlrelu_sse2-no-sse3") {
+-  source_set("qu8-vmul_sse4.1-no-sse4.2") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -57926,7 +59910,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-sse2-u32.c",
+-      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-sse41-mul16-ld64-u16.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -57942,7 +59926,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-qc8w-igemm_arm64_standalone") {
++source_set("qs8-vlrelu_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -57950,27 +59934,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vlrelu_sse2-no-sse3_standalone") {
+-  source_set("qu8-vmul_sse4.1-no-sse4.2_standalone") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16-minmax-fp32-neonv8-mlal-lane.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c2s4-minmax-fp32-neon-mlal.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c2s4-minmax-fp32-neonv8-mlal.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c2s4-minmax-fp32-neon-mlal.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c2s4-minmax-fp32-neonv8-mlal.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4-minmax-fp32-scalar-lrintf.c"
++    "src/src/qs8-vlrelu/gen/qs8-vlrelu-neon-u32.c",
++    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-sse2-u32.c",
+-      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-sse41-mul16-ld64-u16.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -58004,17 +59981,14 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-vlrelu_sse4.1-no-sse4.2") {
--    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
--    ]
-+source_set("qs8-qu8-packw_arm64") {
+-  source_set("qu8-vmul_x64") {
+-    cflags = []
++source_set("qs8-vmul_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-sse41-u32.c",
+-      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c",
 -    ]
 +  ]
  
@@ -58030,7 +60004,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-scalar.c"
++    "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-scalar-u4.c",
++    "src/src/qs8-vmul/gen/qs8-vmul-minmax-rndnu-neon-ld64-u16.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -58040,11 +60015,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vlrelu_sse4.1-no-sse4.2_standalone") {
--    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
--    ]
+-  source_set("qu8-vmul_x64_standalone") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -58054,7 +60026,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-sse41-u32.c",
+-      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -58070,7 +60042,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-qu8-packw_arm64_standalone") {
++source_set("qs8-vmul_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -58082,13 +60054,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-scalar.c"
++    "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-scalar-u4.c",
++    "src/src/qs8-vmul/gen/qs8-vmul-minmax-rndnu-neon-ld64-u16.c"
 +  ]
  
--  source_set("qu8-vlrelu_ssse3-no-sse4.1") {
+-  source_set("qu8-vmulc_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
--      "-mno-sse4.1",
--      "-mssse3",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -58096,7 +60071,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-ssse3-u32.c",
+-      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-avx-mul16-ld64-u16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -58124,17 +60099,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vlrelu_ssse3-no-sse4.1_standalone") {
+-  source_set("qu8-vmulc_avx-no-avx2-no-f16c-no-fma_standalone") {
 -    cflags = [
--      "-mno-sse4.1",
--      "-mssse3",
+-      "-mavx",
+-      "-mno-avx2",
+-      "-mno-f16c",
+-      "-mno-fma",
 -    ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-ssse3-u32.c",
+-      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-avx-mul16-ld64-u16.c",
 -    ]
-+source_set("qs8-rdsum_arm64") {
++source_set("qs8-vmulc_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -58152,8 +60129,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-rdsum/gen/qs8-rdsum-7p7x-minmax-fp32-neon-c32.c",
-+    "src/src/qs8-rdsum/gen/qs8-rdsum-minmax-fp32-scalar-u1-acc1.c"
++    "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-scalar-u4.c",
++    "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-rndnu-neon-ld64-u16.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -58164,8 +60141,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qu8-vlrelu_x64") {
--    cflags = []
+-  source_set("qu8-vmulc_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -58175,8 +60155,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-select-u4.c",
+-      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-sse2-mul16-ld64-u8.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -58192,7 +60171,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-rdsum_arm64_standalone") {
++source_set("qs8-vmulc_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -58200,18 +60179,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vlrelu_x64_standalone") {
--    cflags = []
+-  source_set("qu8-vmulc_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-rdsum/gen/qs8-rdsum-7p7x-minmax-fp32-neon-c32.c",
-+    "src/src/qs8-rdsum/gen/qs8-rdsum-minmax-fp32-scalar-u1-acc1.c"
++    "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-scalar-u4.c",
++    "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-rndnu-neon-ld64-u16.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-select-u4.c",
+-      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-sse2-mul16-ld64-u8.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -58245,19 +60226,20 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-vmul_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qu8-vmulc_sse4.1-no-sse4.2") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
--
++source_set("qs8-vprelu_arm64") {
++  cflags = [
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-avx-mul16-ld64-u16.c",
+-      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-sse41-mul16-ld64-u16.c",
 -    ]
--
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -58268,33 +60250,22 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("qs8-rsum_arch=armv8.2-a+dotprod") {
-+  cflags = [
-+    "-march=armv8.2-a+dotprod"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-rsum/gen/qs8-rsum-neondot-u32-acc2.c"
++    "src/src/qs8-vprelu/gen/qs8-vprelu-scalar-u8.c"
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qu8-vmul_avx-no-avx2-no-f16c-no-fma_standalone") {
--    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-avx-mul16-ld64-u16.c",
+-  # This is a target that cannot depend on //base.
+-  source_set("qu8-vmulc_sse4.1-no-sse4.2_standalone") {
+-    cflags = [
+-      "-mno-sse4.2",
+-      "-msse4.1",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -58303,6 +60274,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-sse41-mul16-ld64-u16.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -58313,14 +60291,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-rsum_arch=armv8.2-a+dotprod_standalone") {
++source_set("qs8-vprelu_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+dotprod"
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -58329,21 +60304,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-rsum/gen/qs8-rsum-neondot-u32-acc2.c"
++    "src/src/qs8-vprelu/gen/qs8-vprelu-scalar-u8.c"
 +  ]
  
--  source_set("qu8-vmul_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("qu8-vmulc_x64") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-sse2-mul16-ld64-u8.c",
+-      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -58371,17 +60343,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vmul_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("qu8-vmulc_x64_standalone") {
+-    cflags = []
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-sse2-mul16-ld64-u8.c",
+-      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c",
 -    ]
-+source_set("qs8-rsum_arm64") {
++source_set("qs8-vpreluc_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -58399,8 +60368,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-rsum/gen/qs8-rsum-neon-u32-acc2.c",
-+    "src/src/qs8-rsum/gen/qs8-rsum-scalar-u4.c"
++    "src/src/qs8-vpreluc/gen/qs8-vpreluc-scalar-u8.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -58411,10 +60379,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qu8-vmul_sse4.1-no-sse4.2") {
+-  source_set("qu8-vprelu_f16c-fma-avx2") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -58425,7 +60394,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-sse41-mul16-ld64-u16.c",
+-      "src/src/qu8-vprelu/gen/qu8-vprelu-avx2-u16.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -58441,7 +60410,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-rsum_arm64_standalone") {
++source_set("qs8-vpreluc_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -58449,20 +60418,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vmul_sse4.1-no-sse4.2_standalone") {
+-  source_set("qu8-vprelu_f16c-fma-avx2_standalone") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-rsum/gen/qs8-rsum-neon-u32-acc2.c",
-+    "src/src/qs8-rsum/gen/qs8-rsum-scalar-u4.c"
++    "src/src/qs8-vpreluc/gen/qs8-vpreluc-scalar-u8.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-sse41-mul16-ld64-u16.c",
+-      "src/src/qu8-vprelu/gen/qu8-vprelu-avx2-u16.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -58496,14 +60465,14 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-vmul_x64") {
+-  source_set("qu8-vprelu_x64") {
 -    cflags = []
-+source_set("qs8-vadd_arm64") {
++source_set("qs8-vrpreluc_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c",
+-      "src/src/qu8-vprelu/gen/qu8-vprelu-scalar-u8.c",
 -    ]
 +  ]
  
@@ -58519,10 +60488,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vadd/gen/qs8-vadd-minmax-neon-ld64-u16.c",
-+    "src/src/qs8-vadd/gen/qs8-vadd-minmax-neon-ld64-u32.c",
-+    "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u1.c",
-+    "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u4.c"
++    "src/src/qs8-vrpreluc/gen/qs8-vrpreluc-scalar-u8.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -58532,7 +60498,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vmul_x64_standalone") {
+-  source_set("qu8-vprelu_x64_standalone") {
 -    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -58543,7 +60509,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c",
+-      "src/src/qu8-vprelu/gen/qu8-vprelu-scalar-u8.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -58559,7 +60525,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-vadd_arm64_standalone") {
++source_set("qs8-vrpreluc_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -58571,18 +60537,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vadd/gen/qs8-vadd-minmax-neon-ld64-u16.c",
-+    "src/src/qs8-vadd/gen/qs8-vadd-minmax-neon-ld64-u32.c",
-+    "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u1.c",
-+    "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u4.c"
++    "src/src/qs8-vrpreluc/gen/qs8-vrpreluc-scalar-u8.c"
 +  ]
  
--  source_set("qu8-vmulc_avx-no-avx2-no-f16c-no-fma") {
+-  source_set("qu8-vpreluc_f16c-fma-avx2") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -58590,7 +60552,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-avx-mul16-ld64-u16.c",
+-      "src/src/qu8-vpreluc/gen/qu8-vpreluc-avx2-u16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -58618,19 +60580,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vmulc_avx-no-avx2-no-f16c-no-fma_standalone") {
+-  source_set("qu8-vpreluc_f16c-fma-avx2_standalone") {
 -    cflags = [
--      "-mavx",
--      "-mno-avx2",
--      "-mno-f16c",
--      "-mno-fma",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-avx-mul16-ld64-u16.c",
+-      "src/src/qu8-vpreluc/gen/qu8-vpreluc-avx2-u16.c",
 -    ]
-+source_set("qs8-vaddc_arm64") {
++source_set("qu8-dwconv_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -58648,10 +60609,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-neon-ld64-u16.c",
-+    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-neon-ld64-u32.c",
-+    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u1.c",
-+    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u4.c"
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-rndnu-neon-mul8.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-rndnu-neon-mul8.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -58662,11 +60625,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qu8-vmulc_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("qu8-vpreluc_x64") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -58676,7 +60636,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-sse2-mul16-ld64-u8.c",
+-      "src/src/qu8-vpreluc/gen/qu8-vpreluc-scalar-u8.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -58692,7 +60652,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-vaddc_arm64_standalone") {
++source_set("qu8-dwconv_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -58700,22 +60660,21 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vmulc_sse2-no-sse3_standalone") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("qu8-vpreluc_x64_standalone") {
+-    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-neon-ld64-u16.c",
-+    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-neon-ld64-u32.c",
-+    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u1.c",
-+    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u4.c"
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-rndnu-neon-mul8.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-rndnu-neon-mul8.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-sse2-mul16-ld64-u8.c",
+-      "src/src/qu8-vpreluc/gen/qu8-vpreluc-scalar-u8.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -58749,17 +60708,18 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-vmulc_sse4.1-no-sse4.2") {
+-  source_set("qu8-vrpreluc_f16c-fma-avx2") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
-+source_set("qs8-vcvt_arm64") {
++source_set("qu8-f32-vcvt_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-sse41-mul16-ld64-u16.c",
+-      "src/src/qu8-vrpreluc/gen/qu8-vrpreluc-avx2-u16.c",
 -    ]
 +  ]
  
@@ -58775,9 +60735,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vcvt/gen/qs8-vcvt-neon-u32.c",
-+    "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u1.c",
-+    "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c"
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-neon-u32.c",
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -58787,10 +60746,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vmulc_sse4.1-no-sse4.2_standalone") {
+-  source_set("qu8-vrpreluc_f16c-fma-avx2_standalone") {
 -    cflags = [
--      "-mno-sse4.2",
--      "-msse4.1",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -58801,7 +60761,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-sse41-mul16-ld64-u16.c",
+-      "src/src/qu8-vrpreluc/gen/qu8-vrpreluc-avx2-u16.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -58817,7 +60777,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-vcvt_arm64_standalone") {
++source_set("qu8-f32-vcvt_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -58829,12 +60789,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vcvt/gen/qs8-vcvt-neon-u32.c",
-+    "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u1.c",
-+    "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c"
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-neon-u32.c",
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c"
 +  ]
  
--  source_set("qu8-vmulc_x64") {
+-  source_set("qu8-vrpreluc_x64") {
 -    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -58842,7 +60801,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c",
+-      "src/src/qu8-vrpreluc/gen/qu8-vrpreluc-scalar-u8.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -58870,16 +60829,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vmulc_x64_standalone") {
+-  source_set("qu8-vrpreluc_x64_standalone") {
 -    cflags = []
-+source_set("qs8-vlrelu_arm64") {
++source_set("qu8-gemm_arch=armv8.2-a+fp16+dotprod") {
 +  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c",
+-      "src/src/qu8-vrpreluc/gen/qu8-vrpreluc-scalar-u8.c",
 -    ]
-+  ]
++  asmflags = cflags
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -58893,9 +60854,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vlrelu/gen/qs8-vlrelu-neon-u32.c",
-+    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c",
-+    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-select-u4.c"
++    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S",
++    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75-prfm.S",
++    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75.S",
++    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu16-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -58926,8 +60888,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/reference/unary-elementwise.cc",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-vlrelu_arm64_standalone") {
++source_set("qu8-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
 +  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -58939,15 +60903,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  ]
++  asmflags = cflags
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vlrelu/gen/qs8-vlrelu-neon-u32.c",
-+    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c",
-+    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-select-u4.c"
++    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S",
++    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75-prfm.S",
++    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75.S",
++    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu16-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -58997,7 +60962,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+source_set("qs8-vmul_arm64") {
++source_set("qu8-gemm_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -59018,8 +60983,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-scalar-u4.c",
-+    "src/src/qs8-vmul/gen/qs8-vmul-minmax-rndnu-neon-ld64-u16.c"
++    "src/src/qu8-gemm/gen/qu8-gemm-1x16-minmax-rndnu-neon-mlal-lane.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-1x16-minmax-rndnu16-neon-mlal-lane.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-1x8-minmax-rndnu-neon-mlal-lane.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-3x8-minmax-rndnu-neon-mlal-lane.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-neon-mlal-lane.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -59059,7 +61029,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-vmul_arm64_standalone") {
++source_set("qu8-gemm_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -59071,8 +61041,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-scalar-u4.c",
-+    "src/src/qs8-vmul/gen/qs8-vmul-minmax-rndnu-neon-ld64-u16.c"
++    "src/src/qu8-gemm/gen/qu8-gemm-1x16-minmax-rndnu-neon-mlal-lane.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-1x16-minmax-rndnu16-neon-mlal-lane.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-1x8-minmax-rndnu-neon-mlal-lane.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-3x8-minmax-rndnu-neon-mlal-lane.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-neon-mlal-lane.c"
 +  ]
  
 -  source_set("s8-ibilinear_sse4.1-no-sse4.2") {
@@ -59124,8 +61099,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/s8-ibilinear/gen/s8-ibilinear-sse41-c16.c",
 -    ]
-+source_set("qs8-vmulc_arm64") {
++source_set("qu8-igemm_arch=armv8.2-a+fp16+dotprod") {
 +  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -59137,13 +61114,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
++  asmflags = cflags
  
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-scalar-u4.c",
-+    "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-rndnu-neon-ld64-u16.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S",
++    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75-prfm.S",
++    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75.S",
++    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu16-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -59181,20 +61160,24 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-vmulc_arm64_standalone") {
++source_set("qu8-igemm_arch=armv8.2-a+fp16+dotprod_standalone") {
 +  cflags = [
++    "-march=armv8.2-a+fp16+dotprod"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  ]
++  asmflags = cflags
  
 -  # This is a target that cannot depend on //base.
 -  source_set("s8-ibilinear_x64_standalone") {
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-scalar-u4.c",
-+    "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-rndnu-neon-ld64-u16.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S",
++    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75-prfm.S",
++    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75.S",
++    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu16-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S"
 +  ]
  
 -    sources = [
@@ -59238,7 +61221,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse4.2",
 -      "-msse4.1",
 -    ]
-+source_set("qu8-dwconv_arm64") {
++source_set("qu8-igemm_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -59259,14 +61242,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-rndnu-neon-mul8.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-rndnu-neon-mul8.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-1x16-minmax-rndnu-neon-mlal-lane.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-1x16-minmax-rndnu16-neon-mlal-lane.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-1x8-minmax-rndnu-neon-mlal-lane.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-3x8-minmax-rndnu-neon-mlal-lane.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-neon-mlal-lane.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -59306,7 +61288,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-dwconv_arm64_standalone") {
++source_set("qu8-igemm_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -59318,14 +61300,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-rndnu-neon-mul8.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-rndnu-neon-mul8.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-1x16-minmax-rndnu-neon-mlal-lane.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-1x16-minmax-rndnu16-neon-mlal-lane.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-1x8-minmax-rndnu-neon-mlal-lane.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-3x8-minmax-rndnu-neon-mlal-lane.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-neon-mlal-lane.c"
 +  ]
  
 -  source_set("s8-maxpool_x64") {
@@ -59371,7 +61352,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-scalar-u1.c",
 -    ]
-+source_set("qu8-f32-vcvt_arm64") {
++source_set("qu8-rdsum_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -59389,9 +61370,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-neon-u32.c",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u1.c",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c"
++    "src/src/qu8-rdsum/gen/qu8-rdsum-7p7x-neon-u16.c",
++    "src/src/qu8-rdsum/gen/qu8-rdsum-7p7x-neon-u32.c",
++    "src/src/qu8-rdsum/gen/qu8-rdsum-scalar.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -59433,7 +61414,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-f32-vcvt_arm64_standalone") {
++source_set("qu8-rdsum_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -59448,9 +61429,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-neon-u32.c",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u1.c",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c"
++    "src/src/qu8-rdsum/gen/qu8-rdsum-7p7x-neon-u16.c",
++    "src/src/qu8-rdsum/gen/qu8-rdsum-7p7x-neon-u32.c",
++    "src/src/qu8-rdsum/gen/qu8-rdsum-scalar.c"
 +  ]
  
 -    sources = [
@@ -59492,17 +61473,15 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("s8-rdminmax_x64") {
 -    cflags = []
-+source_set("qu8-gemm_arch=armv8.2-a+fp16+dotprod") {
++source_set("qu8-rsum_arm64") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-scalar-c2.c",
 -      "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-scalar-c2.c",
 -    ]
-+  asmflags = cflags
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -59516,10 +61495,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S",
-+    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75-prfm.S",
-+    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75.S",
-+    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu16-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S"
++    "src/src/qu8-rsum/gen/qu8-rsum-neon-u32-acc2.c",
++    "src/src/qu8-rsum/gen/qu8-rsum-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -59557,13 +61534,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
++source_set("qu8-rsum_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  asmflags = cflags
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
@@ -59571,10 +61546,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S",
-+    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75-prfm.S",
-+    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75.S",
-+    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu16-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S"
++    "src/src/qu8-rsum/gen/qu8-rsum-neon-u32-acc2.c",
++    "src/src/qu8-rsum/gen/qu8-rsum-scalar-u4.c"
 +  ]
  
 -  source_set("s8-rminmax_sse4.1-no-sse4.2") {
@@ -59623,7 +61596,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse4.2",
 -      "-msse4.1",
 -    ]
-+source_set("qu8-gemm_arm64") {
++source_set("qu8-vadd_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -59646,15 +61619,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x16-minmax-rndnu-neon-mlal-lane.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x16-minmax-rndnu16-neon-mlal-lane.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x8-minmax-rndnu-neon-mlal-lane.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-3x8-minmax-rndnu-neon-mlal-lane.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-neon-mlal-lane.c"
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-neon-ld64-u16.c",
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-neon-ld64-u32.c",
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u1.c",
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -59685,7 +61653,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/s8-rminmax/gen/s8-rminmax-scalar-u2-acc2.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-gemm_arm64_standalone") {
++source_set("qu8-vadd_arm64_standalone") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -59704,15 +61672,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x16-minmax-rndnu-neon-mlal-lane.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x16-minmax-rndnu16-neon-mlal-lane.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x8-minmax-rndnu-neon-mlal-lane.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-3x8-minmax-rndnu-neon-mlal-lane.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-neon-mlal-lane.c"
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-neon-ld64-u16.c",
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-neon-ld64-u32.c",
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u1.c",
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u4.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -59763,16 +61726,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("qu8-igemm_arch=armv8.2-a+fp16+dotprod") {
++source_set("qu8-vaddc_arm64") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
-+  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/s8-vclamp/s8-vclamp-avx2-u128.c",
 -    ]
-+  asmflags = cflags
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -59786,10 +61747,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S",
-+    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75-prfm.S",
-+    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75.S",
-+    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu16-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S"
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-neon-ld64-u16.c",
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-neon-ld64-u32.c",
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -59830,13 +61791,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-igemm_arch=armv8.2-a+fp16+dotprod_standalone") {
++source_set("qu8-vaddc_arm64_standalone") {
 +  cflags = [
-+    "-march=armv8.2-a+fp16+dotprod"
++
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  asmflags = cflags
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-neon-ld64-u16.c",
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-neon-ld64-u32.c",
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
@@ -59853,21 +61820,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S",
-+    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75-prfm.S",
-+    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75.S",
-+    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu16-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/s8-vclamp/s8-vclamp-avx512skx-u256.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -59879,16 +61845,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+  public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
 -  source_set(
@@ -59902,16 +61865,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/s8-vclamp/s8-vclamp-avx512skx-u256.c",
 -    ]
-+source_set("qu8-igemm_arm64") {
++source_set("qu8-vcvt_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -59929,15 +61888,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x16-minmax-rndnu-neon-mlal-lane.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x16-minmax-rndnu16-neon-mlal-lane.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x8-minmax-rndnu-neon-mlal-lane.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-3x8-minmax-rndnu-neon-mlal-lane.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-neon-mlal-lane.c"
++    "src/src/qu8-vcvt/gen/qu8-vcvt-neon-u32.c",
++    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -59978,7 +61930,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-igemm_arm64_standalone") {
++source_set("qu8-vcvt_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -59993,15 +61945,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x16-minmax-rndnu-neon-mlal-lane.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x16-minmax-rndnu16-neon-mlal-lane.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x8-minmax-rndnu-neon-mlal-lane.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-3x8-minmax-rndnu-neon-mlal-lane.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-neon-mlal-lane.c"
++    "src/src/qu8-vcvt/gen/qu8-vcvt-neon-u32.c",
++    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c"
 +  ]
  
 -    sources = [
@@ -60045,7 +61990,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse4.2",
 -      "-msse4.1",
 -    ]
-+source_set("qu8-rdsum_arm64") {
++source_set("qu8-vlrelu_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -60066,9 +62011,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rdsum/gen/qu8-rdsum-7p7x-neon-u16.c",
-+    "src/src/qu8-rdsum/gen/qu8-rdsum-7p7x-neon-u32.c",
-+    "src/src/qu8-rdsum/gen/qu8-rdsum-scalar.c"
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-neon-u32.c",
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -60108,7 +62052,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-rdsum_arm64_standalone") {
++source_set("qu8-vlrelu_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -60120,9 +62064,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rdsum/gen/qu8-rdsum-7p7x-neon-u16.c",
-+    "src/src/qu8-rdsum/gen/qu8-rdsum-7p7x-neon-u32.c",
-+    "src/src/qu8-rdsum/gen/qu8-rdsum-scalar.c"
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-neon-u32.c",
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c"
 +  ]
  
 -  source_set("s8-vclamp_x64") {
@@ -60163,7 +62106,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  # This is a target that cannot depend on //base.
 -  source_set("s8-vclamp_x64_standalone") {
 -    cflags = []
-+source_set("qu8-rsum_arm64") {
++source_set("qu8-vmul_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -60184,8 +62127,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rsum/gen/qu8-rsum-neon-u32-acc2.c",
-+    "src/src/qu8-rsum/gen/qu8-rsum-scalar-u4.c"
++    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c",
++    "src/src/qu8-vmul/gen/qu8-vmul-minmax-rndnu-neon-ld64-u16.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -60228,6 +62171,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/subgraph/static-resize-bilinear-2d.c",
 -      "src/src/subgraph/static-slice.c",
 -      "src/src/subgraph/static-transpose.c",
+-      "src/src/subgraph/subgraph-utils.c",
 -      "src/src/subgraph/unary.c",
 -      "src/src/subgraph/unpooling-2d.c",
 -      "src/src/subgraph/validation.c",
@@ -60255,7 +62199,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("qu8-rsum_arm64_standalone") {
++source_set("qu8-vmul_arm64_standalone") {
 +  cflags = [
  
 -  # This is a target that cannot depend on //base.
@@ -60289,6 +62233,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/subgraph/static-resize-bilinear-2d.c",
 -      "src/src/subgraph/static-slice.c",
 -      "src/src/subgraph/static-transpose.c",
+-      "src/src/subgraph/subgraph-utils.c",
 -      "src/src/subgraph/unary.c",
 -      "src/src/subgraph/unpooling-2d.c",
 -      "src/src/subgraph/validation.c",
@@ -60307,8 +62252,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rsum/gen/qu8-rsum-neon-u32-acc2.c",
-+    "src/src/qu8-rsum/gen/qu8-rsum-scalar-u4.c"
++    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c",
++    "src/src/qu8-vmul/gen/qu8-vmul-minmax-rndnu-neon-ld64-u16.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -60377,7 +62322,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/tables/exp2minus-k-over-8.c",
 -      "src/src/tables/vlog.c",
 -    ]
-+source_set("qu8-vadd_arm64") {
++source_set("qu8-vmulc_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -60395,10 +62340,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-neon-ld64-u16.c",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-neon-ld64-u32.c",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u1.c",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u4.c"
++    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c",
++    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-rndnu-neon-ld64-u16.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -60439,7 +62382,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vadd_arm64_standalone") {
++source_set("qu8-vmulc_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -60454,10 +62397,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-neon-ld64-u16.c",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-neon-ld64-u32.c",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u1.c",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u4.c"
++    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c",
++    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-rndnu-neon-ld64-u16.c"
 +  ]
  
 -    sources = [
@@ -60501,7 +62442,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse4.2",
 -      "-msse4.1",
 -    ]
-+source_set("qu8-vaddc_arm64") {
++source_set("qu8-vprelu_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -60522,10 +62463,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-neon-ld64-u16.c",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-neon-ld64-u32.c",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c"
++    "src/src/qu8-vprelu/gen/qu8-vprelu-scalar-u8.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -60565,7 +62503,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vaddc_arm64_standalone") {
++source_set("qu8-vprelu_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -60577,10 +62515,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-neon-ld64-u16.c",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-neon-ld64-u32.c",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c"
++    "src/src/qu8-vprelu/gen/qu8-vprelu-scalar-u8.c"
 +  ]
  
 -  source_set("u8-ibilinear_x64") {
@@ -60626,7 +62561,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/u8-ibilinear/gen/u8-ibilinear-scalar-c1.c",
 -    ]
-+source_set("qu8-vcvt_arm64") {
++source_set("qu8-vpreluc_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -60644,9 +62579,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-neon-u32.c",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u1.c",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c"
++    "src/src/qu8-vpreluc/gen/qu8-vpreluc-scalar-u8.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -60684,7 +62617,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vcvt_arm64_standalone") {
++source_set("qu8-vpreluc_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -60696,9 +62629,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-neon-u32.c",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u1.c",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c"
++    "src/src/qu8-vpreluc/gen/qu8-vpreluc-scalar-u8.c"
 +  ]
  
 -    sources = [
@@ -60742,7 +62673,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+source_set("qu8-vlrelu_arm64") {
++source_set("qu8-vrpreluc_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -60763,9 +62694,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-neon-u32.c",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-select-u4.c"
++    "src/src/qu8-vrpreluc/gen/qu8-vrpreluc-scalar-u8.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -60805,7 +62734,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vlrelu_arm64_standalone") {
++source_set("qu8-vrpreluc_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -60817,9 +62746,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-neon-u32.c",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-select-u4.c"
++    "src/src/qu8-vrpreluc/gen/qu8-vrpreluc-scalar-u8.c"
 +  ]
  
 -  source_set("u8-maxpool_x64") {
@@ -60865,7 +62792,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-scalar-u1.c",
 -    ]
-+source_set("qu8-vmul_arm64") {
++source_set("reference_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -60883,8 +62810,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c",
-+    "src/src/qu8-vmul/gen/qu8-vmul-minmax-rndnu-neon-ld64-u16.c"
++    "src/src/reference/binary-elementwise.cc",
++    "src/src/reference/packing.cc",
++    "src/src/reference/unary-elementwise.cc"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -60926,7 +62854,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vmul_arm64_standalone") {
++source_set("reference_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -60941,8 +62869,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c",
-+    "src/src/qu8-vmul/gen/qu8-vmul-minmax-rndnu-neon-ld64-u16.c"
++    "src/src/reference/binary-elementwise.cc",
++    "src/src/reference/packing.cc",
++    "src/src/reference/unary-elementwise.cc"
 +  ]
  
 -    sources = [
@@ -60984,7 +62913,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("u8-rdminmax_x64") {
 -    cflags = []
-+source_set("qu8-vmulc_arm64") {
++source_set("s8-ibilinear_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -61006,8 +62935,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c",
-+    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-rndnu-neon-ld64-u16.c"
++    "src/src/s8-ibilinear/gen/s8-ibilinear-neon-c16.c",
++    "src/src/s8-ibilinear/gen/s8-ibilinear-neon-c8.c",
++    "src/src/s8-ibilinear/gen/s8-ibilinear-scalar-c1.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -61045,7 +62975,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vmulc_arm64_standalone") {
++source_set("s8-ibilinear_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -61057,8 +62987,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c",
-+    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-rndnu-neon-ld64-u16.c"
++    "src/src/s8-ibilinear/gen/s8-ibilinear-neon-c16.c",
++    "src/src/s8-ibilinear/gen/s8-ibilinear-neon-c8.c",
++    "src/src/s8-ibilinear/gen/s8-ibilinear-scalar-c1.c"
 +  ]
  
 -  source_set("u8-rminmax_sse2-no-sse3") {
@@ -61107,7 +63038,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+source_set("reference_arm64") {
++source_set("s8-maxpool_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -61130,9 +63061,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/reference/binary-elementwise.cc",
-+    "src/src/reference/packing.cc",
-+    "src/src/reference/unary-elementwise.cc"
++    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-neon-u16.c",
++    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-scalar-u1.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -61163,7 +63093,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/u8-rminmax/gen/u8-rminmax-scalar-u2-acc2.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("reference_arm64_standalone") {
++source_set("s8-maxpool_arm64_standalone") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -61182,9 +63112,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/reference/binary-elementwise.cc",
-+    "src/src/reference/packing.cc",
-+    "src/src/reference/unary-elementwise.cc"
++    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-neon-u16.c",
++    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-scalar-u1.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -61235,7 +63164,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("s8-ibilinear_arm64") {
++source_set("s8-rdminmax_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -61256,9 +63185,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-ibilinear/gen/s8-ibilinear-neon-c16.c",
-+    "src/src/s8-ibilinear/gen/s8-ibilinear-neon-c8.c",
-+    "src/src/s8-ibilinear/gen/s8-ibilinear-scalar-c1.c"
++    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-neon-c32.c",
++    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-scalar-c2.c",
++    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-neon-c32.c",
++    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-scalar-c2.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -61299,7 +63229,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("s8-ibilinear_arm64_standalone") {
++source_set("s8-rdminmax_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -61322,9 +63252,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-ibilinear/gen/s8-ibilinear-neon-c16.c",
-+    "src/src/s8-ibilinear/gen/s8-ibilinear-neon-c8.c",
-+    "src/src/s8-ibilinear/gen/s8-ibilinear-scalar-c1.c"
++    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-neon-c32.c",
++    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-scalar-c2.c",
++    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-neon-c32.c",
++    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-scalar-c2.c"
 +  ]
  
 -    sources = [
@@ -61377,7 +63308,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/u8-vclamp/u8-vclamp-avx512skx-u256.c",
 -    ]
-+source_set("s8-maxpool_arm64") {
++source_set("s8-rminmax_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -61395,8 +63326,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-neon-u16.c",
-+    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-scalar-u1.c"
++    "src/src/s8-rminmax/gen/s8-rmax-neon-u32-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rmax-scalar-u2-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rmin-neon-u32-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rmin-scalar-u2-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rminmax-neon-u32-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rminmax-scalar-u2-acc2.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -61437,7 +63372,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("s8-maxpool_arm64_standalone") {
++source_set("s8-rminmax_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -61452,8 +63387,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-neon-u16.c",
-+    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-scalar-u1.c"
++    "src/src/s8-rminmax/gen/s8-rmax-neon-u32-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rmax-scalar-u2-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rmin-neon-u32-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rmin-scalar-u2-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rminmax-neon-u32-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rminmax-scalar-u2-acc2.c"
 +  ]
  
 -    sources = [
@@ -61494,7 +63433,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("u8-vclamp_x64") {
 -    cflags = []
-+source_set("s8-rdminmax_arm64") {
++source_set("s8-vclamp_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -61515,10 +63454,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-neon-c32.c",
-+    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-scalar-c2.c",
-+    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-neon-c32.c",
-+    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-scalar-c2.c"
++    "src/src/s8-vclamp/s8-vclamp-neon-u64.c",
++    "src/src/s8-vclamp/s8-vclamp-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -61555,7 +63492,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("s8-rdminmax_arm64_standalone") {
++source_set("s8-vclamp_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -61567,10 +63504,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-neon-c32.c",
-+    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-scalar-c2.c",
-+    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-neon-c32.c",
-+    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-scalar-c2.c"
++    "src/src/s8-vclamp/s8-vclamp-neon-u64.c",
++    "src/src/s8-vclamp/s8-vclamp-scalar-u4.c"
 +  ]
  
 -  source_set("x16-packw_f16c-fma-avx2") {
@@ -61624,7 +63559,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/x16-packw/gen/x16-packw-x16-gemm-goi-avx2-u16-prfm.c",
 -    ]
-+source_set("s8-rminmax_arm64") {
++source_set("subgraph_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -61642,12 +63577,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-rminmax/gen/s8-rmax-neon-u32-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rmax-scalar-u2-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rmin-neon-u32-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rmin-scalar-u2-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rminmax-neon-u32-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rminmax-scalar-u2-acc2.c"
++    "src/src/subgraph/argmax-pooling-2d.c",
++    "src/src/subgraph/average-pooling-2d.c",
++    "src/src/subgraph/batch-matrix-multiply.c",
++    "src/src/subgraph/binary.c",
++    "src/src/subgraph/concatenate.c",
++    "src/src/subgraph/convolution-2d.c",
++    "src/src/subgraph/copy.c",
++    "src/src/subgraph/deconvolution-2d.c",
++    "src/src/subgraph/deprecated.c",
++    "src/src/subgraph/depth-to-space-2d.c",
++    "src/src/subgraph/depthwise-convolution-2d.c",
++    "src/src/subgraph/even-split.c",
++    "src/src/subgraph/fully-connected-sparse.c",
++    "src/src/subgraph/fully-connected.c",
++    "src/src/subgraph/max-pooling-2d.c",
++    "src/src/subgraph/pack-lh.c",
++    "src/src/subgraph/reshape-helpers.c",
++    "src/src/subgraph/rope.c",
++    "src/src/subgraph/softmax.c",
++    "src/src/subgraph/space-to-depth-2d.c",
++    "src/src/subgraph/static-constant-pad.c",
++    "src/src/subgraph/static-reduce.c",
++    "src/src/subgraph/static-resize-bilinear-2d.c",
++    "src/src/subgraph/static-slice.c",
++    "src/src/subgraph/static-transpose.c",
++    "src/src/subgraph/subgraph-utils.c",
++    "src/src/subgraph/unary.c",
++    "src/src/subgraph/unpooling-2d.c",
++    "src/src/subgraph/validation.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -61689,7 +63647,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("s8-rminmax_arm64_standalone") {
++source_set("subgraph_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -61705,12 +63663,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-rminmax/gen/s8-rmax-neon-u32-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rmax-scalar-u2-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rmin-neon-u32-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rmin-scalar-u2-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rminmax-neon-u32-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rminmax-scalar-u2-acc2.c"
++    "src/src/subgraph/argmax-pooling-2d.c",
++    "src/src/subgraph/average-pooling-2d.c",
++    "src/src/subgraph/batch-matrix-multiply.c",
++    "src/src/subgraph/binary.c",
++    "src/src/subgraph/concatenate.c",
++    "src/src/subgraph/convolution-2d.c",
++    "src/src/subgraph/copy.c",
++    "src/src/subgraph/deconvolution-2d.c",
++    "src/src/subgraph/deprecated.c",
++    "src/src/subgraph/depth-to-space-2d.c",
++    "src/src/subgraph/depthwise-convolution-2d.c",
++    "src/src/subgraph/even-split.c",
++    "src/src/subgraph/fully-connected-sparse.c",
++    "src/src/subgraph/fully-connected.c",
++    "src/src/subgraph/max-pooling-2d.c",
++    "src/src/subgraph/pack-lh.c",
++    "src/src/subgraph/reshape-helpers.c",
++    "src/src/subgraph/rope.c",
++    "src/src/subgraph/softmax.c",
++    "src/src/subgraph/space-to-depth-2d.c",
++    "src/src/subgraph/static-constant-pad.c",
++    "src/src/subgraph/static-reduce.c",
++    "src/src/subgraph/static-resize-bilinear-2d.c",
++    "src/src/subgraph/static-slice.c",
++    "src/src/subgraph/static-transpose.c",
++    "src/src/subgraph/subgraph-utils.c",
++    "src/src/subgraph/unary.c",
++    "src/src/subgraph/unpooling-2d.c",
++    "src/src/subgraph/validation.c"
 +  ]
  
 -    sources = [
@@ -61754,7 +63735,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+source_set("s8-vclamp_arm64") {
++source_set("tables_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -61775,8 +63756,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-vclamp/s8-vclamp-neon-u64.c",
-+    "src/src/s8-vclamp/s8-vclamp-scalar-u4.c"
++    "src/src/tables/exp2-k-over-2048.c",
++    "src/src/tables/exp2-k-over-64.c",
++    "src/src/tables/exp2minus-k-over-16.c",
++    "src/src/tables/exp2minus-k-over-2048.c",
++    "src/src/tables/exp2minus-k-over-32.c",
++    "src/src/tables/exp2minus-k-over-4.c",
++    "src/src/tables/exp2minus-k-over-64.c",
++    "src/src/tables/exp2minus-k-over-8.c",
++    "src/src/tables/vlog.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -61816,7 +63804,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("s8-vclamp_arm64_standalone") {
++source_set("tables_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -61828,8 +63816,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-vclamp/s8-vclamp-neon-u64.c",
-+    "src/src/s8-vclamp/s8-vclamp-scalar-u4.c"
++    "src/src/tables/exp2-k-over-2048.c",
++    "src/src/tables/exp2-k-over-64.c",
++    "src/src/tables/exp2minus-k-over-16.c",
++    "src/src/tables/exp2minus-k-over-2048.c",
++    "src/src/tables/exp2minus-k-over-32.c",
++    "src/src/tables/exp2minus-k-over-4.c",
++    "src/src/tables/exp2minus-k-over-64.c",
++    "src/src/tables/exp2minus-k-over-8.c",
++    "src/src/tables/vlog.c"
 +  ]
  
 -  source_set("x16-transposec_x64") {
@@ -61875,7 +63870,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c",
 -    ]
-+source_set("subgraph_arm64") {
++source_set("u8-ibilinear_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -61893,34 +63888,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/subgraph/argmax-pooling-2d.c",
-+    "src/src/subgraph/average-pooling-2d.c",
-+    "src/src/subgraph/batch-matrix-multiply.c",
-+    "src/src/subgraph/binary.c",
-+    "src/src/subgraph/concatenate.c",
-+    "src/src/subgraph/convolution-2d.c",
-+    "src/src/subgraph/copy.c",
-+    "src/src/subgraph/deconvolution-2d.c",
-+    "src/src/subgraph/deprecated.c",
-+    "src/src/subgraph/depth-to-space-2d.c",
-+    "src/src/subgraph/depthwise-convolution-2d.c",
-+    "src/src/subgraph/even-split.c",
-+    "src/src/subgraph/fully-connected-sparse.c",
-+    "src/src/subgraph/fully-connected.c",
-+    "src/src/subgraph/max-pooling-2d.c",
-+    "src/src/subgraph/pack-lh.c",
-+    "src/src/subgraph/reshape-helpers.c",
-+    "src/src/subgraph/rope.c",
-+    "src/src/subgraph/softmax.c",
-+    "src/src/subgraph/space-to-depth-2d.c",
-+    "src/src/subgraph/static-constant-pad.c",
-+    "src/src/subgraph/static-reduce.c",
-+    "src/src/subgraph/static-resize-bilinear-2d.c",
-+    "src/src/subgraph/static-slice.c",
-+    "src/src/subgraph/static-transpose.c",
-+    "src/src/subgraph/unary.c",
-+    "src/src/subgraph/unpooling-2d.c",
-+    "src/src/subgraph/validation.c"
++    "src/src/u8-ibilinear/gen/u8-ibilinear-neon-c16.c",
++    "src/src/u8-ibilinear/gen/u8-ibilinear-neon-c8.c",
++    "src/src/u8-ibilinear/gen/u8-ibilinear-scalar-c1.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -61959,7 +63929,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("subgraph_arm64_standalone") {
++source_set("u8-ibilinear_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -61971,34 +63941,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/subgraph/argmax-pooling-2d.c",
-+    "src/src/subgraph/average-pooling-2d.c",
-+    "src/src/subgraph/batch-matrix-multiply.c",
-+    "src/src/subgraph/binary.c",
-+    "src/src/subgraph/concatenate.c",
-+    "src/src/subgraph/convolution-2d.c",
-+    "src/src/subgraph/copy.c",
-+    "src/src/subgraph/deconvolution-2d.c",
-+    "src/src/subgraph/deprecated.c",
-+    "src/src/subgraph/depth-to-space-2d.c",
-+    "src/src/subgraph/depthwise-convolution-2d.c",
-+    "src/src/subgraph/even-split.c",
-+    "src/src/subgraph/fully-connected-sparse.c",
-+    "src/src/subgraph/fully-connected.c",
-+    "src/src/subgraph/max-pooling-2d.c",
-+    "src/src/subgraph/pack-lh.c",
-+    "src/src/subgraph/reshape-helpers.c",
-+    "src/src/subgraph/rope.c",
-+    "src/src/subgraph/softmax.c",
-+    "src/src/subgraph/space-to-depth-2d.c",
-+    "src/src/subgraph/static-constant-pad.c",
-+    "src/src/subgraph/static-reduce.c",
-+    "src/src/subgraph/static-resize-bilinear-2d.c",
-+    "src/src/subgraph/static-slice.c",
-+    "src/src/subgraph/static-transpose.c",
-+    "src/src/subgraph/unary.c",
-+    "src/src/subgraph/unpooling-2d.c",
-+    "src/src/subgraph/validation.c"
++    "src/src/u8-ibilinear/gen/u8-ibilinear-neon-c16.c",
++    "src/src/u8-ibilinear/gen/u8-ibilinear-neon-c8.c",
++    "src/src/u8-ibilinear/gen/u8-ibilinear-scalar-c1.c"
 +  ]
  
 -    sources = [
@@ -62043,7 +63988,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse4.1",
 -      "-mssse3",
 -    ]
-+source_set("tables_arm64") {
++source_set("u8-lut32norm_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -62064,15 +64009,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/tables/exp2-k-over-2048.c",
-+    "src/src/tables/exp2-k-over-64.c",
-+    "src/src/tables/exp2minus-k-over-16.c",
-+    "src/src/tables/exp2minus-k-over-2048.c",
-+    "src/src/tables/exp2minus-k-over-32.c",
-+    "src/src/tables/exp2minus-k-over-4.c",
-+    "src/src/tables/exp2minus-k-over-64.c",
-+    "src/src/tables/exp2minus-k-over-8.c",
-+    "src/src/tables/vlog.c"
++    "src/src/u8-lut32norm/u8-lut32norm-scalar.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -62112,7 +64049,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("tables_arm64_standalone") {
++source_set("u8-lut32norm_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -62124,15 +64061,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/tables/exp2-k-over-2048.c",
-+    "src/src/tables/exp2-k-over-64.c",
-+    "src/src/tables/exp2minus-k-over-16.c",
-+    "src/src/tables/exp2minus-k-over-2048.c",
-+    "src/src/tables/exp2minus-k-over-32.c",
-+    "src/src/tables/exp2minus-k-over-4.c",
-+    "src/src/tables/exp2minus-k-over-64.c",
-+    "src/src/tables/exp2minus-k-over-8.c",
-+    "src/src/tables/vlog.c"
++    "src/src/u8-lut32norm/u8-lut32norm-scalar.c"
 +  ]
  
 -  source_set("x24-transposec_x64") {
@@ -62173,7 +64102,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  # This is a target that cannot depend on //base.
 -  source_set("x24-transposec_x64_standalone") {
 -    cflags = []
-+source_set("u8-ibilinear_arm64") {
++source_set("u8-maxpool_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -62194,9 +64123,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-ibilinear/gen/u8-ibilinear-neon-c16.c",
-+    "src/src/u8-ibilinear/gen/u8-ibilinear-neon-c8.c",
-+    "src/src/u8-ibilinear/gen/u8-ibilinear-scalar-c1.c"
++    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-neon-u16.c",
++    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-scalar-u1.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -62234,7 +64162,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/x32-packw/gen/x32-packw-x8-gemm-goi-avx-u4.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("u8-ibilinear_arm64_standalone") {
++source_set("u8-maxpool_arm64_standalone") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -62253,9 +64181,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-ibilinear/gen/u8-ibilinear-neon-c16.c",
-+    "src/src/u8-ibilinear/gen/u8-ibilinear-neon-c8.c",
-+    "src/src/u8-ibilinear/gen/u8-ibilinear-scalar-c1.c"
++    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-neon-u16.c",
++    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-scalar-u1.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -62318,7 +64245,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/x32-packw/gen/x32-packw-x32-gemm-gio-avx512f-u8.c",
 -      "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-avx512f-u4-prfm.c",
 -    ]
-+source_set("u8-lut32norm_arm64") {
++source_set("u8-rdminmax_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -62337,7 +64264,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-lut32norm/u8-lut32norm-scalar.c"
++    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-neon-c32.c",
++    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-scalar-c2.c",
++    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-neon-c32.c",
++    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-scalar-c2.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -62381,7 +64311,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    }
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("u8-lut32norm_arm64_standalone") {
++source_set("u8-rdminmax_arm64_standalone") {
 +  cflags = [
  
 -  source_set("x32-packw_sse2-no-sse3") {
@@ -62398,7 +64328,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-lut32norm/u8-lut32norm-scalar.c"
++    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-neon-c32.c",
++    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-scalar-c2.c",
++    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-neon-c32.c",
++    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-scalar-c2.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -62414,21 +64347,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
  
 -  # This is a target that cannot depend on //base.
 -  source_set("x32-packw_sse2-no-sse3_standalone") {
@@ -62436,15 +64363,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+source_set("u8-maxpool_arm64") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/x32-packw/gen/x32-packw-x2c4-gemm-goi-sse2-u4.c",
 -      "src/src/x32-packw/gen/x32-packw-x8-gemm-goi-sse2-u4.c",
 -    ]
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -62456,16 +64385,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-neon-u16.c",
-+    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-scalar-u1.c"
-+  ]
++source_set("u8-rminmax_arm64") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
@@ -62484,11 +64408,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/x32-packw/gen/x32-packw-x4-gemm-goi-scalar-float-u4.c",
 -      "src/src/x32-packw/gen/x32-packw-x64-gemm-goi-scalar-int-u2.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-rminmax/gen/u8-rmax-neon-u32-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rmax-scalar-u2-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rmin-neon-u32-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rmin-scalar-u2-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rminmax-neon-u32-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rminmax-scalar-u2-acc2.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -62501,14 +64428,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+# This is a target that cannot depend on //base.
-+source_set("u8-maxpool_arm64_standalone") {
-+  cflags = [
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set("x32-packw_x64_standalone") {
@@ -62523,7 +64454,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/x32-packw/gen/x32-packw-x4-gemm-goi-scalar-float-u4.c",
 -      "src/src/x32-packw/gen/x32-packw-x64-gemm-goi-scalar-int-u2.c",
 -    ]
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -62535,33 +64467,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-neon-u16.c",
-+    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-scalar-u1.c"
-+  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++# This is a target that cannot depend on //base.
++source_set("u8-rminmax_arm64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-rminmax/gen/u8-rmax-neon-u32-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rmax-scalar-u2-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rmin-neon-u32-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rmin-scalar-u2-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rminmax-neon-u32-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rminmax-scalar-u2-acc2.c"
++  ]
  
 -  source_set("x32-transposec_avx-no-avx2-no-f16c-no-fma") {
 -    cflags = [
@@ -62570,13 +64495,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
-+source_set("u8-rdminmax_arm64") {
-+  cflags = [
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/x32-transposec/gen/x32-transposec-8x8-reuse-multi-avx.c",
 -    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -62589,19 +64520,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-neon-c32.c",
-+    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-scalar-c2.c",
-+    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-neon-c32.c",
-+    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-scalar-c2.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
 -  source_set("x32-transposec_avx-no-avx2-no-f16c-no-fma_standalone") {
@@ -62611,19 +64536,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-f16c",
 -      "-mno-fma",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/x32-transposec/gen/x32-transposec-8x8-reuse-multi-avx.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++source_set("u8-vclamp_arm64") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -62635,45 +64554,42 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("u8-rdminmax_arm64_standalone") {
-+  cflags = [
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-vclamp/u8-vclamp-neon-u64.c",
++    "src/src/u8-vclamp/u8-vclamp-scalar-u4.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-neon-c32.c",
-+    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-scalar-c2.c",
-+    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-neon-c32.c",
-+    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-scalar-c2.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  source_set("x32-transposec_sse-no-sse2") {
 -    cflags = [
 -      "-mno-sse2",
 -      "-msse",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/x32-transposec/x32-transposec-4x4-sse.c",
--    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/x32-transposec/x32-transposec-4x4-sse.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -62684,13 +64600,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++# This is a target that cannot depend on //base.
++source_set("u8-vclamp_arm64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set("x32-transposec_sse-no-sse2_standalone") {
@@ -62698,13 +64614,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse2",
 -      "-msse",
 -    ]
--
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-vclamp/u8-vclamp-neon-u64.c",
++    "src/src/u8-vclamp/u8-vclamp-scalar-u4.c"
++  ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/x32-transposec/x32-transposec-4x4-sse.c",
 -    ]
-+source_set("u8-rminmax_arm64") {
-+  cflags = [
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -62716,42 +64638,39 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/u8-rminmax/gen/u8-rmax-neon-u32-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rmax-scalar-u2-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rmin-neon-u32-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rmin-scalar-u2-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rminmax-neon-u32-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rminmax-scalar-u2-acc2.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  source_set("x32-transposec_x64") {
 -    cflags = []
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++source_set("x16-pack-lh_arch=armv8.2-a+sve+sve2") {
++  cflags = [
++    "-march=armv8.2-a+sve+sve2"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/x32-transposec/gen/x32-transposec-2x4-scalar-int.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x16-pack-lh/x16-packlh-neonsme2.c"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -62763,34 +64682,34 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("u8-rminmax_arm64_standalone") {
-+  cflags = [
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set("x32-transposec_x64_standalone") {
 -    cflags = []
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/u8-rminmax/gen/u8-rmax-neon-u32-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rmax-scalar-u2-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rmin-neon-u32-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rmin-scalar-u2-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rminmax-neon-u32-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rminmax-scalar-u2-acc2.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/x32-transposec/gen/x32-transposec-2x4-scalar-int.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++# This is a target that cannot depend on //base.
++source_set("x16-pack-lh_arch=armv8.2-a+sve+sve2_standalone") {
++  cflags = [
++    "-march=armv8.2-a+sve+sve2"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -62802,14 +64721,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x16-pack-lh/x16-packlh-neonsme2.c"
++  ]
++
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
- 
--    public_configs = [ ":xnnpack_config" ]
++
 +  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -62825,7 +64753,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+source_set("u8-vclamp_arm64") {
++source_set("x16-packw_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -62846,8 +64774,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-vclamp/u8-vclamp-neon-u64.c",
-+    "src/src/u8-vclamp/u8-vclamp-scalar-u4.c"
++    "src/src/x16-packw/gen/x16-packw-x16-gemm-goi-neon-ld4lane-u8-prfm.c",
++    "src/src/x16-packw/gen/x16-packw-x8-gemm-goi-neon-ld4lane-u8-prfm.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -62887,7 +64815,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("u8-vclamp_arm64_standalone") {
++source_set("x16-packw_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -62899,8 +64827,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-vclamp/u8-vclamp-neon-u64.c",
-+    "src/src/u8-vclamp/u8-vclamp-scalar-u4.c"
++    "src/src/x16-packw/gen/x16-packw-x16-gemm-goi-neon-ld4lane-u8-prfm.c",
++    "src/src/x16-packw/gen/x16-packw-x8-gemm-goi-neon-ld4lane-u8-prfm.c"
 +  ]
  
 -  source_set("x32-unpool_x64") {
@@ -62946,7 +64874,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/x32-unpool/x32-unpool-scalar.c",
 -    ]
-+source_set("x16-packw_arm64") {
++source_set("x16-transposec_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -62964,8 +64892,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x16-packw/gen/x16-packw-x16-gemm-goi-neon-ld4lane-u8-prfm.c",
-+    "src/src/x16-packw/gen/x16-packw-x8-gemm-goi-neon-ld4lane-u8-prfm.c"
++    "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c",
++    "src/src/x16-transposec/gen/x16-transposec-8x8-reuse-dec-zip-neon.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -63008,7 +64936,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x16-packw_arm64_standalone") {
++source_set("x16-transposec_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -63025,8 +64953,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x16-packw/gen/x16-packw-x16-gemm-goi-neon-ld4lane-u8-prfm.c",
-+    "src/src/x16-packw/gen/x16-packw-x8-gemm-goi-neon-ld4lane-u8-prfm.c"
++    "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c",
++    "src/src/x16-transposec/gen/x16-transposec-8x8-reuse-dec-zip-neon.c"
 +  ]
  
 -    sources = [
@@ -63070,7 +64998,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+source_set("x16-transposec_arm64") {
++source_set("x16-x32-packw_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -63091,8 +65019,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c",
-+    "src/src/x16-transposec/gen/x16-transposec-8x8-reuse-dec-zip-neon.c"
++    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-gio-scalar.c",
++    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-goi-scalar.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -63132,7 +65060,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x16-transposec_arm64_standalone") {
++source_set("x16-x32-packw_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -63144,8 +65072,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c",
-+    "src/src/x16-transposec/gen/x16-transposec-8x8-reuse-dec-zip-neon.c"
++    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-gio-scalar.c",
++    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-goi-scalar.c"
 +  ]
  
 -  source_set("x64-transposec_x64") {
@@ -63191,7 +65119,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c",
 -    ]
-+source_set("x16-x32-packw_arm64") {
++source_set("x24-transposec_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -63209,8 +65137,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-gio-scalar.c",
-+    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-goi-scalar.c"
++    "src/src/x24-transposec/gen/x24-transposec-1x2-scalar.c",
++    "src/src/x24-transposec/x24-transposec-2x2-neon-tbl64.c",
++    "src/src/x24-transposec/x24-transposec-4x4-aarch64-neon-tbl128.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -63253,7 +65182,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x16-x32-packw_arm64_standalone") {
++source_set("x24-transposec_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -63270,8 +65199,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-gio-scalar.c",
-+    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-goi-scalar.c"
++    "src/src/x24-transposec/gen/x24-transposec-1x2-scalar.c",
++    "src/src/x24-transposec/x24-transposec-2x2-neon-tbl64.c",
++    "src/src/x24-transposec/x24-transposec-4x4-aarch64-neon-tbl128.c"
 +  ]
  
 -    sources = [
@@ -63316,13 +65246,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+source_set("x24-transposec_arm64") {
-+  cflags = [
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/x8-lut/gen/x8-lut-avx2-u128.c",
 -    ]
++source_set("x32-pack-lh_arch=armv8.2-a+sve+sve2") {
++  cflags = [
++    "-march=armv8.2-a+sve+sve2"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -63337,9 +65268,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x24-transposec/gen/x24-transposec-1x2-scalar.c",
-+    "src/src/x24-transposec/x24-transposec-2x2-neon-tbl64.c",
-+    "src/src/x24-transposec/x24-transposec-4x4-aarch64-neon-tbl128.c"
++    "src/src/x32-pack-lh/x32-packlh-neonsme2.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -63380,10 +65309,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x24-transposec_arm64_standalone") {
++source_set("x32-pack-lh_arch=armv8.2-a+sve+sve2_standalone") {
 +  cflags = [
++    "-march=armv8.2-a+sve+sve2"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-pack-lh/x32-packlh-neonsme2.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -63401,20 +65335,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mf16c",
 -      "-mfma",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x24-transposec/gen/x24-transposec-1x2-scalar.c",
-+    "src/src/x24-transposec/x24-transposec-2x2-neon-tbl64.c",
-+    "src/src/x24-transposec/x24-transposec-4x4-aarch64-neon-tbl128.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/x8-lut/gen/x8-lut-avx512skx-vpshufb-u64.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -63426,13 +65360,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
 +  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -63718,15 +65645,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  # This is a target that cannot depend on //base.
 -  source_set("x8-packq_x64_standalone") {
 -    cflags = []
-+source_set("x32-unpool_arm64") {
-+  cflags = [
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/x8-packq/x8-packq-scalar-f32qp8-u1.c",
 -    ]
-+  ]
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -63737,16 +65661,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x32-unpool/x32-unpool-neon.c",
-+    "src/src/x32-unpool/x32-unpool-scalar.c"
-+  ]
++source_set("x32-unpool_arm64") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
@@ -63763,97 +65682,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/x8-packw/gen/x8-packw-x4-gemm-goi-scalar-u2.c",
 -      "src/src/x8-packw/gen/x8-packw-x8-gemm-goi-scalar-u2.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
-+# This is a target that cannot depend on //base.
-+source_set("x32-unpool_arm64_standalone") {
-+  cflags = [
- 
--  # This is a target that cannot depend on //base.
--  source_set("x8-packw_x64_standalone") {
--    cflags = []
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/x8-packw/gen/x8-packw-x16-gemm-goi-scalar-u2.c",
--      "src/src/x8-packw/gen/x8-packw-x32-gemm-goi-scalar-u2.c",
--      "src/src/x8-packw/gen/x8-packw-x4-gemm-goi-scalar-u2.c",
--      "src/src/x8-packw/gen/x8-packw-x8-gemm-goi-scalar-u2.c",
--    ]
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/x32-unpool/x32-unpool-neon.c",
 +    "src/src/x32-unpool/x32-unpool-scalar.c"
 +  ]
-+
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
-+  public_configs = [ ":xnnpack_config" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
- 
--  source_set("x8-transposec_f16c-fma-avx2") {
--    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
--    ]
-+source_set("x64-transposec_arm64") {
-+  cflags = [
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/x8-transposec/gen/x8-transposec-32x32-reuse-switch-avx2.c",
--    ]
-+  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -63865,26 +65698,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x64-transposec/gen/x64-transposec-2x2-multi-dec-zip-neon.c",
-+    "src/src/x64-transposec/gen/x64-transposec-2x2-reuse-dec-zip-neon.c",
-+    "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("x8-transposec_f16c-fma-avx2_standalone") {
--    cflags = [
--      "-mavx2",
--      "-mf16c",
--      "-mfma",
--    ]
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -63892,9 +65711,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-  # This is a target that cannot depend on //base.
+-  source_set("x8-packw_x64_standalone") {
+-    cflags = []
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x8-transposec/gen/x8-transposec-32x32-reuse-switch-avx2.c",
+-      "src/src/x8-packw/gen/x8-packw-x16-gemm-goi-scalar-u2.c",
+-      "src/src/x8-packw/gen/x8-packw-x32-gemm-goi-scalar-u2.c",
+-      "src/src/x8-packw/gen/x8-packw-x4-gemm-goi-scalar-u2.c",
+-      "src/src/x8-packw/gen/x8-packw-x8-gemm-goi-scalar-u2.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -63910,7 +65736,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x64-transposec_arm64_standalone") {
++source_set("x32-unpool_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -63922,15 +65748,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x64-transposec/gen/x64-transposec-2x2-multi-dec-zip-neon.c",
-+    "src/src/x64-transposec/gen/x64-transposec-2x2-reuse-dec-zip-neon.c",
-+    "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c"
++    "src/src/x32-unpool/x32-unpool-neon.c",
++    "src/src/x32-unpool/x32-unpool-scalar.c"
 +  ]
  
--  source_set("x8-transposec_sse2-no-sse3") {
+-  source_set("x8-transposec_f16c-fma-avx2") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -63938,7 +65764,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x8-transposec/gen/x8-transposec-16x16-reuse-mov-sse2.c",
+-      "src/src/x8-transposec/gen/x8-transposec-32x32-reuse-switch-avx2.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -63966,17 +65792,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("x8-transposec_sse2-no-sse3_standalone") {
+-  source_set("x8-transposec_f16c-fma-avx2_standalone") {
 -    cflags = [
--      "-mno-sse3",
--      "-msse2",
+-      "-mavx2",
+-      "-mf16c",
+-      "-mfma",
 -    ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x8-transposec/gen/x8-transposec-16x16-reuse-mov-sse2.c",
+-      "src/src/x8-transposec/gen/x8-transposec-32x32-reuse-switch-avx2.c",
 -    ]
-+source_set("x8-lut_arm64") {
++source_set("x64-transposec_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -63994,8 +65821,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x8-lut/gen/x8-lut-aarch64-neon-tbx128x4-u64.c",
-+    "src/src/x8-lut/gen/x8-lut-scalar-u4.c"
++    "src/src/x64-transposec/gen/x64-transposec-2x2-multi-dec-zip-neon.c",
++    "src/src/x64-transposec/gen/x64-transposec-2x2-reuse-dec-zip-neon.c",
++    "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -64006,8 +65834,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("x8-transposec_x64") {
--    cflags = []
+-  source_set("x8-transposec_sse2-no-sse3") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -64017,7 +65848,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x8-transposec/gen/x8-transposec-2x4-scalar-int.c",
+-      "src/src/x8-transposec/gen/x8-transposec-16x16-reuse-mov-sse2.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -64033,7 +65864,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x8-lut_arm64_standalone") {
++source_set("x64-transposec_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -64041,17 +65872,21 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("x8-transposec_x64_standalone") {
--    cflags = []
+-  source_set("x8-transposec_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x8-lut/gen/x8-lut-aarch64-neon-tbx128x4-u64.c",
-+    "src/src/x8-lut/gen/x8-lut-scalar-u4.c"
++    "src/src/x64-transposec/gen/x64-transposec-2x2-multi-dec-zip-neon.c",
++    "src/src/x64-transposec/gen/x64-transposec-2x2-reuse-dec-zip-neon.c",
++    "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x8-transposec/gen/x8-transposec-2x4-scalar-int.c",
+-      "src/src/x8-transposec/gen/x8-transposec-16x16-reuse-mov-sse2.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -64085,14 +65920,14 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("xx-copy_x64") {
+-  source_set("x8-transposec_x64") {
 -    cflags = []
-+source_set("x8-packq_arm64") {
++source_set("x8-lut_arm64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/xx-copy/xx-copy-scalar-memcpy.c",
+-      "src/src/x8-transposec/gen/x8-transposec-2x4-scalar-int.c",
 -    ]
 +  ]
  
@@ -64108,8 +65943,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x8-packq/x8-packq-aarch64-neon-f32qp8-u2.c",
-+    "src/src/x8-packq/x8-packq-scalar-f32qp8-u1.c"
++    "src/src/x8-lut/gen/x8-lut-aarch64-neon-tbx128x4-u64.c",
++    "src/src/x8-lut/gen/x8-lut-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -64119,7 +65954,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("xx-copy_x64_standalone") {
+-  source_set("x8-transposec_x64_standalone") {
 -    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -64130,7 +65965,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/xx-copy/xx-copy-scalar-memcpy.c",
+-      "src/src/x8-transposec/gen/x8-transposec-2x4-scalar-int.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -64146,7 +65981,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x8-packq_arm64_standalone") {
++source_set("x8-lut_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -64158,22 +65993,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x8-packq/x8-packq-aarch64-neon-f32qp8-u2.c",
-+    "src/src/x8-packq/x8-packq-scalar-f32qp8-u1.c"
++    "src/src/x8-lut/gen/x8-lut-aarch64-neon-tbx128x4-u64.c",
++    "src/src/x8-lut/gen/x8-lut-scalar-u4.c"
 +  ]
  
--  source_set("xx-fill_sse2-no-sse3") {
--    cflags = [
--      "-mno-sse3",
--      "-msse2",
--    ]
+-  source_set("xx-copy_x64") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/xx-fill/xx-fill-sse2-u64.c",
+-      "src/src/xx-copy/xx-copy-scalar-memcpy.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -64201,18 +66033,110 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("xx-fill_sse2-no-sse3_standalone") {
+-  source_set("xx-copy_x64_standalone") {
+-    cflags = []
++source_set("x8-pack-lh_arch=armv8.2-a+sve+sve2") {
++  cflags = [
++    "-march=armv8.2-a+sve+sve2"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/xx-copy/xx-copy-scalar-memcpy.c",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-pack-lh/x8--packlh-neonsme2.c"
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-  source_set("xx-fill_sse2-no-sse3") {
 -    cflags = [
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
++# This is a target that cannot depend on //base.
++source_set("x8-pack-lh_arch=armv8.2-a+sve+sve2_standalone") {
++  cflags = [
++    "-march=armv8.2-a+sve+sve2"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/xx-fill/xx-fill-sse2-u64.c",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-pack-lh/x8--packlh-neonsme2.c"
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
 -
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("xx-fill_sse2-no-sse3_standalone") {
+-    cflags = [
+-      "-mno-sse3",
+-      "-msse2",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/xx-fill/xx-fill-sse2-u64.c",
 -    ]
-+source_set("x8-packw_arm64") {
-+  cflags = [
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -64224,27 +66148,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
++source_set("x8-packq_arm64") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x8-packw/gen/x8-packw-x16-gemm-goi-scalar-u2.c",
-+    "src/src/x8-packw/gen/x8-packw-x32-gemm-goi-scalar-u2.c",
-+    "src/src/x8-packw/gen/x8-packw-x4-gemm-goi-scalar-u2.c",
-+    "src/src/x8-packw/gen/x8-packw-x8-gemm-goi-scalar-u2.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-packq/x8-packq-aarch64-neon-f32qp8-u2.c",
++    "src/src/x8-packq/x8-packq-scalar-f32qp8-u1.c"
++  ]
+ 
+-  source_set("xx-fill_x64") {
+-    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("xx-fill_x64") {
--    cflags = []
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/xx-fill/xx-fill-scalar-u16.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -64252,13 +66181,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/xx-fill/xx-fill-scalar-u16.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -64269,32 +66191,29 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("x8-packw_arm64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  ]
++# This is a target that cannot depend on //base.
++source_set("x8-packq_arm64_standalone") {
++  cflags = [
  
 -  # This is a target that cannot depend on //base.
 -  source_set("xx-fill_x64_standalone") {
 -    cflags = []
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x8-packw/gen/x8-packw-x16-gemm-goi-scalar-u2.c",
-+    "src/src/x8-packw/gen/x8-packw-x32-gemm-goi-scalar-u2.c",
-+    "src/src/x8-packw/gen/x8-packw-x4-gemm-goi-scalar-u2.c",
-+    "src/src/x8-packw/gen/x8-packw-x8-gemm-goi-scalar-u2.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/xx-fill/xx-fill-scalar-u16.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-packq/x8-packq-aarch64-neon-f32qp8-u2.c",
++    "src/src/x8-packq/x8-packq-scalar-f32qp8-u1.c"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -64306,6 +66225,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -64329,7 +66252,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "-mno-sse3",
 -      "-msse2",
 -    ]
-+source_set("x8-transposec_arm64") {
++source_set("x8-packw_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -64350,8 +66273,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x8-transposec/gen/x8-transposec-16x16-reuse-dec-zip-neon.c",
-+    "src/src/x8-transposec/gen/x8-transposec-2x4-scalar-int.c"
++    "src/src/x8-packw/gen/x8-packw-x16-gemm-goi-scalar-u2.c",
++    "src/src/x8-packw/gen/x8-packw-x32-gemm-goi-scalar-u2.c",
++    "src/src/x8-packw/gen/x8-packw-x4-gemm-goi-scalar-u2.c",
++    "src/src/x8-packw/gen/x8-packw-x8-gemm-goi-scalar-u2.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -64391,7 +66316,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x8-transposec_arm64_standalone") {
++source_set("x8-packw_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -64403,8 +66328,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x8-transposec/gen/x8-transposec-16x16-reuse-dec-zip-neon.c",
-+    "src/src/x8-transposec/gen/x8-transposec-2x4-scalar-int.c"
++    "src/src/x8-packw/gen/x8-packw-x16-gemm-goi-scalar-u2.c",
++    "src/src/x8-packw/gen/x8-packw-x32-gemm-goi-scalar-u2.c",
++    "src/src/x8-packw/gen/x8-packw-x4-gemm-goi-scalar-u2.c",
++    "src/src/x8-packw/gen/x8-packw-x8-gemm-goi-scalar-u2.c"
 +  ]
  
 -  source_set("xx-pad_x64") {
@@ -64450,7 +66377,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/xx-pad/xx-pad-p4-scalar-u16.c",
 -    ]
-+source_set("xx-copy_arm64") {
++source_set("x8-transposec_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -64468,7 +66395,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/xx-copy/xx-copy-scalar-memcpy.c"
++    "src/src/x8-transposec/gen/x8-transposec-16x16-reuse-dec-zip-neon.c",
++    "src/src/x8-transposec/gen/x8-transposec-2x4-scalar-int.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -64506,7 +66434,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("xx-copy_arm64_standalone") {
++source_set("x8-transposec_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -64518,7 +66446,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/xx-copy/xx-copy-scalar-memcpy.c"
++    "src/src/x8-transposec/gen/x8-transposec-16x16-reuse-dec-zip-neon.c",
++    "src/src/x8-transposec/gen/x8-transposec-2x4-scalar-int.c"
 +  ]
  
 -    sources = [
@@ -64588,7 +66517,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/configs/xx-fill-config.c",
 -      "src/src/configs/xx-pad-config.c",
 -    ]
-+source_set("xx-fill_arm64") {
++source_set("xx-copy_arm64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -64607,8 +66536,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/xx-fill/xx-fill-neon-u64.c",
-+    "src/src/xx-fill/xx-fill-scalar-u16.c"
++    "src/src/xx-copy/xx-copy-scalar-memcpy.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -64684,7 +66612,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/enums/operator-type.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("xx-fill_arm64_standalone") {
++source_set("xx-copy_arm64_standalone") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -64703,8 +66631,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/xx-fill/xx-fill-neon-u64.c",
-+    "src/src/xx-fill/xx-fill-scalar-u16.c"
++    "src/src/xx-copy/xx-copy-scalar-memcpy.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -64753,7 +66680,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f16-avgpool_arch=armv8.2-a+fp16") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
-+source_set("xx-pad_arm64") {
++source_set("xx-fill_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -64774,8 +66701,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/xx-pad/xx-pad-p16-neon-u16.c",
-+    "src/src/xx-pad/xx-pad-p4-scalar-u16.c"
++    "src/src/xx-fill/xx-fill-neon-u64.c",
++    "src/src/xx-fill/xx-fill-scalar-u16.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -64812,7 +66739,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("xx-pad_arm64_standalone") {
++source_set("xx-fill_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -64824,8 +66751,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/xx-pad/xx-pad-p16-neon-u16.c",
-+    "src/src/xx-pad/xx-pad-p4-scalar-u16.c"
++    "src/src/xx-fill/xx-fill-neon-u64.c",
++    "src/src/xx-fill/xx-fill-scalar-u16.c"
 +  ]
  
 -  source_set("f16-conv-hwc2chw_arch=armv8.2-a+fp16") {
@@ -64866,7 +66793,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  # This is a target that cannot depend on //base.
 -  source_set("f16-conv-hwc2chw_arch=armv8.2-a+fp16_standalone") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
-+source_set("xx-transposev_arm64") {
++source_set("xx-pad_arm64") {
 +  cflags = [
  
 -    sources = [
@@ -64887,7 +66814,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/xx-transposev/xx-transposev-1x1-scalar-memcpy.c"
++    "src/src/xx-pad/xx-pad-p16-neon-u16.c",
++    "src/src/xx-pad/xx-pad-p4-scalar-u16.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -64933,7 +66861,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("xx-transposev_arm64_standalone") {
++source_set("xx-pad_arm64_standalone") {
 +  cflags = [
  
 -  # This is a target that cannot depend on //base.
@@ -64961,7 +66889,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/xx-transposev/xx-transposev-1x1-scalar-memcpy.c"
++    "src/src/xx-pad/xx-pad-p16-neon-u16.c",
++    "src/src/xx-pad/xx-pad-p4-scalar-u16.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -65035,14 +66964,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 -
 -    public_configs = [ ":xnnpack_config" ]
-+}
- 
+-
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+if (current_cpu == "ppc64") {
-+source_set("configs_ppc64") {
++source_set("xx-transposev_arm64") {
 +  cflags = [
  
 -  source_set("f16-f32-vcvt_arm64") {
@@ -65057,30 +66984,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/configs/argmaxpool-config.c",
-+    "src/src/configs/avgpool-config.c",
-+    "src/src/configs/binary-elementwise-config.c",
-+    "src/src/configs/cmul-config.c",
-+    "src/src/configs/conv-hwc2chw-config.c",
-+    "src/src/configs/dwconv-config.c",
-+    "src/src/configs/dwconv2d-chw-config.c",
-+    "src/src/configs/gemm-config.c",
-+    "src/src/configs/hardware-config.c",
-+    "src/src/configs/ibilinear-chw-config.c",
-+    "src/src/configs/ibilinear-config.c",
-+    "src/src/configs/lut32norm-config.c",
-+    "src/src/configs/maxpool-config.c",
-+    "src/src/configs/pack-lh-config.c",
-+    "src/src/configs/raddstoreexpminusmax-config.c",
-+    "src/src/configs/reduce-config.c",
-+    "src/src/configs/spmm-config.c",
-+    "src/src/configs/transpose-config.c",
-+    "src/src/configs/unary-elementwise-config.c",
-+    "src/src/configs/unpool-config.c",
-+    "src/src/configs/vmulcaddc-config.c",
-+    "src/src/configs/x8-lut-config.c",
-+    "src/src/configs/xx-fill-config.c",
-+    "src/src/configs/xx-pad-config.c"
++    "src/src/xx-transposev/xx-transposev-1x1-scalar-memcpy.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -65130,7 +67034,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("configs_ppc64_standalone") {
++source_set("xx-transposev_arm64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -65142,30 +67046,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/configs/argmaxpool-config.c",
-+    "src/src/configs/avgpool-config.c",
-+    "src/src/configs/binary-elementwise-config.c",
-+    "src/src/configs/cmul-config.c",
-+    "src/src/configs/conv-hwc2chw-config.c",
-+    "src/src/configs/dwconv-config.c",
-+    "src/src/configs/dwconv2d-chw-config.c",
-+    "src/src/configs/gemm-config.c",
-+    "src/src/configs/hardware-config.c",
-+    "src/src/configs/ibilinear-chw-config.c",
-+    "src/src/configs/ibilinear-config.c",
-+    "src/src/configs/lut32norm-config.c",
-+    "src/src/configs/maxpool-config.c",
-+    "src/src/configs/pack-lh-config.c",
-+    "src/src/configs/raddstoreexpminusmax-config.c",
-+    "src/src/configs/reduce-config.c",
-+    "src/src/configs/spmm-config.c",
-+    "src/src/configs/transpose-config.c",
-+    "src/src/configs/unary-elementwise-config.c",
-+    "src/src/configs/unpool-config.c",
-+    "src/src/configs/vmulcaddc-config.c",
-+    "src/src/configs/x8-lut-config.c",
-+    "src/src/configs/xx-fill-config.c",
-+    "src/src/configs/xx-pad-config.c"
++    "src/src/xx-transposev/xx-transposev-1x1-scalar-memcpy.c"
 +  ]
  
 -  source_set("f16-f32acc-rdsum_arch=armv8.2-a+fp16") {
@@ -65206,14 +67087,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  # This is a target that cannot depend on //base.
 -  source_set("f16-f32acc-rdsum_arch=armv8.2-a+fp16_standalone") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
-+source_set("enums_ppc64") {
-+  cflags = [
++}
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f16-f32acc-rdsum/gen/f16-f32acc-rdsum-7p7x-minmax-neonfp16arith-c16.c",
 -    ]
-+  ]
++if (current_cpu == "ppc64") {
++source_set("configs_ppc64") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -65225,24 +67107,47 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/enums/allocation-type.c",
-+    "src/src/enums/datatype-strings.c",
-+    "src/src/enums/microkernel-type.c",
-+    "src/src/enums/node-type.c",
-+    "src/src/enums/operator-type.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/configs/argmaxpool-config.c",
++    "src/src/configs/avgpool-config.c",
++    "src/src/configs/binary-elementwise-config.c",
++    "src/src/configs/cmul-config.c",
++    "src/src/configs/conv-hwc2chw-config.c",
++    "src/src/configs/dwconv-config.c",
++    "src/src/configs/dwconv2d-chw-config.c",
++    "src/src/configs/gemm-config.c",
++    "src/src/configs/hardware-config.c",
++    "src/src/configs/ibilinear-chw-config.c",
++    "src/src/configs/ibilinear-config.c",
++    "src/src/configs/lut32norm-config.c",
++    "src/src/configs/maxpool-config.c",
++    "src/src/configs/pack-lh-config.c",
++    "src/src/configs/raddstoreexpminusmax-config.c",
++    "src/src/configs/reduce-config.c",
++    "src/src/configs/spmm-config.c",
++    "src/src/configs/transpose-config.c",
++    "src/src/configs/unary-elementwise-config.c",
++    "src/src/configs/unpool-config.c",
++    "src/src/configs/vmulcaddc-config.c",
++    "src/src/configs/x8-lut-config.c",
++    "src/src/configs/xx-fill-config.c",
++    "src/src/configs/xx-pad-config.c"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  source_set("f16-f32acc-rsum_arch=armv8.2-a+fp16") {
+-    cflags = [ "-march=armv8.2-a+fp16" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -65250,18 +67155,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  source_set("f16-f32acc-rsum_arch=armv8.2-a+fp16") {
--    cflags = [ "-march=armv8.2-a+fp16" ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f16-f32acc-rsum/gen/f16-f32acc-rsum-neonfp16arith-u32-acc4.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("enums_ppc64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -65273,36 +67172,52 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  ]
++# This is a target that cannot depend on //base.
++source_set("configs_ppc64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/enums/allocation-type.c",
-+    "src/src/enums/datatype-strings.c",
-+    "src/src/enums/microkernel-type.c",
-+    "src/src/enums/node-type.c",
-+    "src/src/enums/operator-type.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set("f16-f32acc-rsum_arch=armv8.2-a+fp16_standalone") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/configs/argmaxpool-config.c",
++    "src/src/configs/avgpool-config.c",
++    "src/src/configs/binary-elementwise-config.c",
++    "src/src/configs/cmul-config.c",
++    "src/src/configs/conv-hwc2chw-config.c",
++    "src/src/configs/dwconv-config.c",
++    "src/src/configs/dwconv2d-chw-config.c",
++    "src/src/configs/gemm-config.c",
++    "src/src/configs/hardware-config.c",
++    "src/src/configs/ibilinear-chw-config.c",
++    "src/src/configs/ibilinear-config.c",
++    "src/src/configs/lut32norm-config.c",
++    "src/src/configs/maxpool-config.c",
++    "src/src/configs/pack-lh-config.c",
++    "src/src/configs/raddstoreexpminusmax-config.c",
++    "src/src/configs/reduce-config.c",
++    "src/src/configs/spmm-config.c",
++    "src/src/configs/transpose-config.c",
++    "src/src/configs/unary-elementwise-config.c",
++    "src/src/configs/unpool-config.c",
++    "src/src/configs/vmulcaddc-config.c",
++    "src/src/configs/x8-lut-config.c",
++    "src/src/configs/xx-fill-config.c",
++    "src/src/configs/xx-pad-config.c"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/f16-f32acc-rsum/gen/f16-f32acc-rsum-neonfp16arith-u32-acc4.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -65314,13 +67229,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
@@ -65337,8 +67254,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-neonfp16arith-ld64.c",
 -      "src/src/f16-gemm/gen/f16-gemm-6x8-minmax-neonfp16arith-ld64.c",
 -    ]
-+source_set("f16-f32-vcvt_ppc64") {
-+  cflags = [
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -65350,13 +67269,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  ]
++source_set("enums_ppc64") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-f32-vcvt/gen/f16-f32-vcvt-scalar-u4.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -65370,9 +67287,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-neonfp16arith-ld64.c",
 -      "src/src/f16-gemm/gen/f16-gemm-6x8-minmax-neonfp16arith-ld64.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/enums/allocation-type.c",
++    "src/src/enums/datatype-strings.c",
++    "src/src/enums/microkernel-type.c",
++    "src/src/enums/node-type.c",
++    "src/src/enums/operator-type.c"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -65384,6 +67306,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -65391,10 +67318,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
@@ -65414,9 +67337,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a75.S",
 -      "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f16-f32-vcvt_ppc64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -65428,13 +67350,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  ]
++# This is a target that cannot depend on //base.
++source_set("enums_ppc64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f16-f32-vcvt/gen/f16-f32-vcvt-scalar-u4.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -65452,9 +67373,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a75.S",
 -      "src/src/f16-gemm/gen/f16-gemm-6x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/enums/allocation-type.c",
++    "src/src/enums/datatype-strings.c",
++    "src/src/enums/microkernel-type.c",
++    "src/src/enums/node-type.c",
++    "src/src/enums/operator-type.c"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -65466,14 +67392,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
- 
--    public_configs = [ ":xnnpack_config" ]
++
 +  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -65486,7 +67416,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f16-ibilinear-chw_arch=armv8.2-a+fp16") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
-+source_set("f16-qs8-vcvt_ppc64") {
++source_set("f16-f32-vcvt_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -65507,7 +67437,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-qs8-vcvt/gen/f16-qs8-vcvt-scalar-imagic-u4.c"
++    "src/src/f16-f32-vcvt/gen/f16-f32-vcvt-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -65544,7 +67474,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-qs8-vcvt_ppc64_standalone") {
++source_set("f16-f32-vcvt_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -65556,7 +67486,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-qs8-vcvt/gen/f16-qs8-vcvt-scalar-imagic-u4.c"
++    "src/src/f16-f32-vcvt/gen/f16-f32-vcvt-scalar-u4.c"
 +  ]
  
 -  source_set("f16-ibilinear_arch=armv8.2-a+fp16") {
@@ -65597,7 +67527,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  # This is a target that cannot depend on //base.
 -  source_set("f16-ibilinear_arch=armv8.2-a+fp16_standalone") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
-+source_set("f16-qu8-vcvt_ppc64") {
++source_set("f16-qs8-vcvt_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -65618,7 +67548,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-qu8-vcvt/gen/f16-qu8-vcvt-scalar-imagic-u4.c"
++    "src/src/f16-qs8-vcvt/gen/f16-qs8-vcvt-scalar-imagic-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -65664,7 +67594,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("f16-qu8-vcvt_ppc64_standalone") {
++source_set("f16-qs8-vcvt_ppc64_standalone") {
 +  cflags = [
  
 -  # This is a target that cannot depend on //base.
@@ -65692,7 +67622,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-qu8-vcvt/gen/f16-qu8-vcvt-scalar-imagic-u4.c"
++    "src/src/f16-qs8-vcvt/gen/f16-qs8-vcvt-scalar-imagic-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -65759,7 +67689,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-cortex-a75.S",
 -      "src/src/f16-igemm/f16-igemm-6x16-minmax-asm-aarch64-neonfp16arith-ld64.S",
 -    ]
-+source_set("f16-rdminmax_ppc64") {
++source_set("f16-qu8-vcvt_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -65777,8 +67707,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-rdminmax/gen/f16-rdmax-2p2x-scalar-c2.c",
-+    "src/src/f16-rdminmax/gen/f16-rdmin-2p2x-scalar-c2.c"
++    "src/src/f16-qu8-vcvt/gen/f16-qu8-vcvt-scalar-imagic-u4.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -65816,7 +67745,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-rdminmax_ppc64_standalone") {
++source_set("f16-qu8-vcvt_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -65828,8 +67757,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = [ "-march=armv8.2-a+fp16" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-rdminmax/gen/f16-rdmax-2p2x-scalar-c2.c",
-+    "src/src/f16-rdminmax/gen/f16-rdmin-2p2x-scalar-c2.c"
++    "src/src/f16-qu8-vcvt/gen/f16-qu8-vcvt-scalar-imagic-u4.c"
 +  ]
  
 -    sources = [
@@ -65870,7 +67798,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f16-qs8-vcvt_arch=armv8.2-a+fp16") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
-+source_set("f16-rminmax_ppc64") {
++source_set("f16-rdminmax_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -65891,9 +67819,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-rminmax/gen/f16-rmax-scalar-u2-acc2.c",
-+    "src/src/f16-rminmax/gen/f16-rmin-scalar-u2-acc2.c",
-+    "src/src/f16-rminmax/gen/f16-rminmax-scalar-u2-acc2.c"
++    "src/src/f16-rdminmax/gen/f16-rdmax-2p2x-scalar-c2.c",
++    "src/src/f16-rdminmax/gen/f16-rdmin-2p2x-scalar-c2.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -65930,7 +67857,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-rminmax_ppc64_standalone") {
++source_set("f16-rdminmax_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -65942,9 +67869,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-rminmax/gen/f16-rmax-scalar-u2-acc2.c",
-+    "src/src/f16-rminmax/gen/f16-rmin-scalar-u2-acc2.c",
-+    "src/src/f16-rminmax/gen/f16-rminmax-scalar-u2-acc2.c"
++    "src/src/f16-rdminmax/gen/f16-rdmax-2p2x-scalar-c2.c",
++    "src/src/f16-rdminmax/gen/f16-rdmin-2p2x-scalar-c2.c"
 +  ]
  
 -  source_set("f16-qs8-vcvt_arm64") {
@@ -65990,7 +67916,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/f16-qs8-vcvt/gen/f16-qs8-vcvt-scalar-imagic-u4.c",
 -    ]
-+source_set("f16-vapproxgelu_ppc64") {
++source_set("f16-rminmax_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -66008,7 +67934,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vapproxgelu/gen/f16-vapproxgelu-scalar-rational-6-4-div.c"
++    "src/src/f16-rminmax/gen/f16-rmax-scalar-u2-acc2.c",
++    "src/src/f16-rminmax/gen/f16-rmin-scalar-u2-acc2.c",
++    "src/src/f16-rminmax/gen/f16-rminmax-scalar-u2-acc2.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -66046,7 +67974,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-vapproxgelu_ppc64_standalone") {
++source_set("f16-rminmax_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -66058,7 +67986,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vapproxgelu/gen/f16-vapproxgelu-scalar-rational-6-4-div.c"
++    "src/src/f16-rminmax/gen/f16-rmax-scalar-u2-acc2.c",
++    "src/src/f16-rminmax/gen/f16-rmin-scalar-u2-acc2.c",
++    "src/src/f16-rminmax/gen/f16-rminmax-scalar-u2-acc2.c"
 +  ]
  
 -    sources = [
@@ -66099,7 +68029,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f16-raddstoreexpminusmax_arch=armv8.2-a+fp16") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
-+source_set("f16-vcos_ppc64") {
++source_set("f16-vapproxgelu_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -66120,7 +68050,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vcos/gen/f16-vcos-scalar-rational-3-2-div.c"
++    "src/src/f16-vapproxgelu/gen/f16-vapproxgelu-scalar-rational-6-4-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -66157,7 +68087,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-vcos_ppc64_standalone") {
++source_set("f16-vapproxgelu_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -66169,7 +68099,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vcos/gen/f16-vcos-scalar-rational-3-2-div.c"
++    "src/src/f16-vapproxgelu/gen/f16-vapproxgelu-scalar-rational-6-4-div.c"
 +  ]
  
 -  source_set("f16-rdminmax_arch=armv8.2-a+fp16") {
@@ -66217,7 +68147,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f16-rdminmax/gen/f16-rdmax-2p2x-neonfp16arith-c32.c",
 -      "src/src/f16-rdminmax/gen/f16-rdmin-2p2x-neonfp16arith-c32.c",
 -    ]
-+source_set("f16-vexp_ppc64") {
++source_set("f16-vcos_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -66235,7 +68165,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vexp/gen/f16-vexp-scalar-poly-3.c"
++    "src/src/f16-vcos/gen/f16-vcos-scalar-rational-3-2-div.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -66274,7 +68204,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-vexp_ppc64_standalone") {
++source_set("f16-vcos_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -66286,7 +68216,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vexp/gen/f16-vexp-scalar-poly-3.c"
++    "src/src/f16-vcos/gen/f16-vcos-scalar-rational-3-2-div.c"
 +  ]
  
 -    sources = [
@@ -66328,7 +68258,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f16-rminmax_arch=armv8.2-a+fp16") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
-+source_set("f16-vgelu_ppc64") {
++source_set("f16-vexp_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -66351,7 +68281,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vgelu/gen/f16-vgelu-scalar-rational-6-4-div.c"
++    "src/src/f16-vexp/gen/f16-vexp-scalar-poly-3.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -66392,7 +68322,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-vgelu_ppc64_standalone") {
++source_set("f16-vexp_ppc64_standalone") {
 +  cflags = [
  
 -    if (!(is_android && use_order_profiling)) {
@@ -66405,7 +68335,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vgelu/gen/f16-vgelu-scalar-rational-6-4-div.c"
++    "src/src/f16-vexp/gen/f16-vexp-scalar-poly-3.c"
 +  ]
  
 -    sources = [
@@ -66453,7 +68383,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f16-rminmax/gen/f16-rmin-scalar-u2-acc2.c",
 -      "src/src/f16-rminmax/gen/f16-rminmax-scalar-u2-acc2.c",
 -    ]
-+source_set("f16-vsin_ppc64") {
++source_set("f16-vgelu_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -66471,7 +68401,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vsin/gen/f16-vsin-scalar-rational-3-2-div.c"
++    "src/src/f16-vgelu/gen/f16-vgelu-scalar-rational-6-4-div.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -66509,7 +68439,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f16-vsin_ppc64_standalone") {
++source_set("f16-vgelu_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -66521,7 +68451,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = [ "-march=armv8.2-a+fp16" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f16-vsin/gen/f16-vsin-scalar-rational-3-2-div.c"
++    "src/src/f16-vgelu/gen/f16-vgelu-scalar-rational-6-4-div.c"
 +  ]
  
 -    sources = [
@@ -66562,7 +68492,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f16-vapproxgelu_arch=armv8.2-a+fp16") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
-+source_set("f32-argmaxpool_ppc64") {
++source_set("f16-vsin_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -66583,7 +68513,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c"
++    "src/src/f16-vsin/gen/f16-vsin-scalar-rational-3-2-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -66620,7 +68550,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-argmaxpool_ppc64_standalone") {
++source_set("f16-vsin_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -66632,7 +68562,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c"
++    "src/src/f16-vsin/gen/f16-vsin-scalar-rational-3-2-div.c"
 +  ]
  
 -  source_set("f16-vapproxgelu_arm64") {
@@ -66689,7 +68619,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("f32-avgpool_ppc64") {
++source_set("f32-argmaxpool_ppc64") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -66730,7 +68660,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-scalar-u1.c"
++    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -66799,7 +68729,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-avgpool_ppc64_standalone") {
++source_set("f32-argmaxpool_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -66811,7 +68741,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-scalar-u1.c"
++    "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c"
 +  ]
  
 -  source_set("f16-vclamp_arch=armv8.2-a+fp16") {
@@ -66857,7 +68787,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/f16-vclamp/gen/f16-vclamp-neonfp16arith-u16.c",
 -    ]
-+source_set("f32-conv-hwc2chw_ppc64") {
++source_set("f32-avgpool_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -66875,7 +68805,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-scalar-1x1.c"
++    "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-scalar-u1.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -66913,7 +68843,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-conv-hwc2chw_ppc64_standalone") {
++source_set("f32-avgpool_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -66925,7 +68855,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = [ "-march=armv8.2-a+fp16" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-scalar-1x1.c"
++    "src/src/f32-avgpool/gen/f32-avgpool-9p-minmax-scalar-u1.c"
 +  ]
  
 -    sources = [
@@ -66966,7 +68896,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f16-vcos_arch=armv8.2-a+fp16") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
-+source_set("f32-dwconv2d-chw_ppc64") {
++source_set("f32-conv-hwc2chw_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -66987,14 +68917,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-2x1-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-4x1.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-1x1-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-2x1-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-1x1-acc5.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-2x1-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-1x1-acc5.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-2x1-acc2.c"
++    "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-scalar-1x1.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -67031,7 +68954,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-dwconv2d-chw_ppc64_standalone") {
++source_set("f32-conv-hwc2chw_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -67043,14 +68966,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-2x1-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-4x1.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-1x1-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-2x1-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-1x1-acc5.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-2x1-acc2.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-1x1-acc5.c",
-+    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-2x1-acc2.c"
++    "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-scalar-1x1.c"
 +  ]
  
 -  source_set("f16-vcos_arm64") {
@@ -67096,7 +69012,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/f16-vcos/gen/f16-vcos-scalar-rational-3-2-div.c",
 -    ]
-+source_set("f32-dwconv_ppc64") {
++source_set("f32-dwconv2d-chw_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -67114,16 +69030,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p1c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p1c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p2c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p2c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-3p1c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-3p1c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-4p1c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-4p1c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-9p1c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-9p1c-scalar-acc2.c"
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-2x1-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-4x1.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-1x1-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-2x1-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-1x1-acc5.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-2x1-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-1x1-acc5.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-2x1-acc2.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -67161,7 +69075,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-dwconv_ppc64_standalone") {
++source_set("f32-dwconv2d-chw_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -67173,16 +69087,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = [ "-march=armv8.2-a+fp16" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p1c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p1c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p2c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-25p2c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-3p1c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-3p1c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-4p1c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-4p1c-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-9p1c-minmax-scalar-acc2.c",
-+    "src/src/f32-dwconv/gen/f32-dwconv-9p1c-scalar-acc2.c"
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-2x1-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3p1-minmax-scalar-4x1.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-1x1-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-3x3s2p1-minmax-scalar-2x1-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-1x1-acc5.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5p2-minmax-scalar-2x1-acc2.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-1x1-acc5.c",
++    "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-2x1-acc2.c"
 +  ]
  
 -    sources = [
@@ -67223,7 +69135,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f16-vexp_arch=armv8.2-a+fp16") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
-+source_set("f32-f16-vcvt_ppc64") {
++source_set("f32-dwconv_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -67244,7 +69156,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-scalar-fabsf-u2.c"
++    "src/src/f32-dwconv/gen/f32-dwconv-25p1c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-25p1c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-25p2c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-25p2c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-3p1c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-3p1c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-4p1c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-4p1c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-9p1c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-9p1c-scalar-acc2.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -67281,7 +69202,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-f16-vcvt_ppc64_standalone") {
++source_set("f32-dwconv_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -67293,7 +69214,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-scalar-fabsf-u2.c"
++    "src/src/f32-dwconv/gen/f32-dwconv-25p1c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-25p1c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-25p2c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-25p2c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-3p1c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-3p1c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-4p1c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-4p1c-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-9p1c-minmax-scalar-acc2.c",
++    "src/src/f32-dwconv/gen/f32-dwconv-9p1c-scalar-acc2.c"
 +  ]
  
 -  source_set("f16-vexp_arm64") {
@@ -67339,7 +69269,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/f16-vexp/gen/f16-vexp-scalar-poly-3.c",
 -    ]
-+source_set("f32-gemm_ppc64") {
++source_set("f32-f16-vcvt_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -67357,17 +69287,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-gemm/gen/f32-gemm-1x4-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-1x4-relu-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-1x4-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-relu-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x2-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x4-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x4-relu-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x4-scalar.c"
++    "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-scalar-fabsf-u2.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -67405,7 +69325,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-gemm_ppc64_standalone") {
++source_set("f32-f16-vcvt_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -67417,17 +69337,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = [ "-march=armv8.2-a+fp16" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-gemm/gen/f32-gemm-1x4-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-1x4-relu-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-1x4-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-relu-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-2x4-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x2-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x4-minmax-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x4-relu-scalar.c",
-+    "src/src/f32-gemm/gen/f32-gemm-4x4-scalar.c"
++    "src/src/f32-f16-vcvt/gen/f32-f16-vcvt-scalar-fabsf-u2.c"
 +  ]
  
 -    sources = [
@@ -67468,7 +69378,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f16-vgelu_arm64") {
 -    cflags = []
-+source_set("f32-ibilinear-chw_ppc64") {
++source_set("f32-gemm_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -67489,7 +69399,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-scalar-p4.c"
++    "src/src/f32-gemm/gen/f32-gemm-1x4-minmax-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-1x4-relu-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-1x4-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x2-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x4-minmax-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x4-relu-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x4-scalar.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -67526,7 +69443,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-ibilinear-chw_ppc64_standalone") {
++source_set("f32-gemm_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -67538,7 +69455,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-scalar-p4.c"
++    "src/src/f32-gemm/gen/f32-gemm-1x4-minmax-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-1x4-relu-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-1x4-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x2-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x4-minmax-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x4-relu-scalar.c",
++    "src/src/f32-gemm/gen/f32-gemm-4x4-scalar.c"
 +  ]
  
 -  source_set("f16-vhswish_arch=armv8.2-a+fp16") {
@@ -67584,7 +69508,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/f16-vhswish/gen/f16-vhswish-neonfp16arith-u16.c",
 -    ]
-+source_set("f32-ibilinear_ppc64") {
++source_set("f32-ibilinear-chw_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -67602,7 +69526,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-ibilinear/gen/f32-ibilinear-scalar-c2.c"
++    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-scalar-p4.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -67640,7 +69564,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-ibilinear_ppc64_standalone") {
++source_set("f32-ibilinear-chw_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -67652,7 +69576,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = [ "-march=armv8.2-a+fp16" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-ibilinear/gen/f32-ibilinear-scalar-c2.c"
++    "src/src/f32-ibilinear-chw/gen/f32-ibilinear-chw-scalar-p4.c"
 +  ]
  
 -    sources = [
@@ -67709,7 +69633,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("f32-igemm_ppc64") {
++source_set("f32-ibilinear_ppc64") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -67721,17 +69645,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = [ "-march=armv8.2-a+fp16" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-igemm/gen/f32-igemm-1x4-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x4-relu-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x4-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-relu-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x2-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x4-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x4-relu-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x4-scalar.c"
++    "src/src/f32-ibilinear/gen/f32-ibilinear-scalar-c2.c"
 +  ]
  
 -    sources = [
@@ -67779,7 +69693,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f16-vrnd/gen/f16-vrndz-neonfp16arith-u16.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-igemm_ppc64_standalone") {
++source_set("f32-ibilinear_ppc64_standalone") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -67798,17 +69712,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-igemm/gen/f32-igemm-1x4-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x4-relu-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-1x4-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-relu-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-2x4-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x2-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x4-minmax-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x4-relu-scalar.c",
-+    "src/src/f32-igemm/gen/f32-igemm-4x4-scalar.c"
++    "src/src/f32-ibilinear/gen/f32-ibilinear-scalar-c2.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -67856,7 +69760,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f16-vrsqrt_arch=armv8.2-a+fp16") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
-+source_set("f32-maxpool_ppc64") {
++source_set("f32-igemm_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -67877,7 +69781,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-maxpool/gen/f32-maxpool-9p-minmax-scalar-u1.c"
++    "src/src/f32-igemm/gen/f32-igemm-1x4-minmax-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x4-relu-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x4-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x2-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x4-minmax-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x4-relu-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x4-scalar.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -67914,7 +69825,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-maxpool_ppc64_standalone") {
++source_set("f32-igemm_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -67926,7 +69837,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-maxpool/gen/f32-maxpool-9p-minmax-scalar-u1.c"
++    "src/src/f32-igemm/gen/f32-igemm-1x4-minmax-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x4-relu-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-1x4-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x2-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x4-minmax-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x4-relu-scalar.c",
++    "src/src/f32-igemm/gen/f32-igemm-4x4-scalar.c"
 +  ]
  
 -  source_set("f16-vsigmoid_arch=armv8.2-a+fp16") {
@@ -67974,7 +69892,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f16-vsigmoid/gen/f16-vsigmoid-neonfp16arith-rr2-p2-nr1fma-u32.c",
 -      "src/src/f16-vsigmoid/gen/f16-vsigmoid-neonfp16arith-rr2-p2-nr1recps-u16.c",
 -    ]
-+source_set("f32-qc4w-gemm_ppc64") {
++source_set("f32-maxpool_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -67992,8 +69910,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-1x4-minmax-scalar.c",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-4x4-minmax-scalar.c"
++    "src/src/f32-maxpool/gen/f32-maxpool-9p-minmax-scalar-u1.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -68031,7 +69948,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-qc4w-gemm_ppc64_standalone") {
++source_set("f32-maxpool_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -68043,8 +69960,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = [ "-march=armv8.2-a+fp16" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-1x4-minmax-scalar.c",
-+    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-4x4-minmax-scalar.c"
++    "src/src/f32-maxpool/gen/f32-maxpool-9p-minmax-scalar-u1.c"
 +  ]
  
 -    sources = [
@@ -68085,7 +70001,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f16-vsin_arm64") {
 -    cflags = []
-+source_set("f32-qc8w-gemm_ppc64") {
++source_set("f32-qc4w-gemm_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -68106,8 +70022,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x4-minmax-scalar.c",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-4x4-minmax-scalar.c"
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-1x4-minmax-scalar.c",
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-4x4-minmax-scalar.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -68144,7 +70060,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-qc8w-gemm_ppc64_standalone") {
++source_set("f32-qc4w-gemm_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -68156,8 +70072,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x4-minmax-scalar.c",
-+    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-4x4-minmax-scalar.c"
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-1x4-minmax-scalar.c",
++    "src/src/f32-qc4w-gemm/gen/f32-qc4w-gemm-4x4-minmax-scalar.c"
 +  ]
  
 -  source_set("f16-vsqrt_arch=armv8.2-a+fp16") {
@@ -68205,7 +70121,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f16-vsqrt/gen/f16-vsqrt-aarch64-neonfp16arith-sqrt-u8.c",
 -      "src/src/f16-vsqrt/gen/f16-vsqrt-neonfp16arith-nr1fma1adj-u8.c",
 -    ]
-+source_set("f32-qs8-vcvt_ppc64") {
++source_set("f32-qc8w-gemm_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -68223,9 +70139,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u1.c",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u4.c",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-lrintf-u4.c"
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x4-minmax-scalar.c",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-4x4-minmax-scalar.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -68264,7 +70179,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-qs8-vcvt_ppc64_standalone") {
++source_set("f32-qc8w-gemm_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -68276,9 +70191,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = [ "-march=armv8.2-a+fp16" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u1.c",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u4.c",
-+    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-lrintf-u4.c"
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-1x4-minmax-scalar.c",
++    "src/src/f32-qc8w-gemm/gen/f32-qc8w-gemm-4x4-minmax-scalar.c"
 +  ]
  
 -    sources = [
@@ -68317,11 +70231,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    assert_no_deps = [ "//base" ]
    }
 +}
++
++source_set("f32-qs8-vcvt_ppc64") {
++  cflags = [
  
 -  source_set("f16-vunary_arch=armv8.2-a+fp16") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
-+source_set("f32-qu8-vcvt_ppc64") {
-+  cflags = [
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
@@ -68329,6 +70245,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f16-vunary/gen/f16-vneg-neonfp16arith-u16.c",
 -      "src/src/f16-vunary/gen/f16-vsqr-neonfp16arith-u16.c",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u4.c",
++    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-lrintf-u4.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -68341,19 +70261,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u1.c",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u4.c",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-lrintf-u4.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
 -  # This is a target that cannot depend on //base.
 -  source_set("f16-vunary_arch=armv8.2-a+fp16_standalone") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
@@ -68364,12 +70284,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f16-vunary/gen/f16-vneg-neonfp16arith-u16.c",
 -      "src/src/f16-vunary/gen/f16-vsqr-neonfp16arith-u16.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -68381,11 +70297,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
-+
 +# This is a target that cannot depend on //base.
-+source_set("f32-qu8-vcvt_ppc64_standalone") {
++source_set("f32-qs8-vcvt_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -68397,9 +70310,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u1.c",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u4.c",
-+    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-lrintf-u4.c"
++    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u4.c",
++    "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-lrintf-u4.c"
 +  ]
  
 -  source_set("f32-argmaxpool_arm64") {
@@ -68447,7 +70359,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-neon-c4.c",
 -      "src/src/f32-argmaxpool/f32-argmaxpool-9p8x-scalar-c1.c",
 -    ]
-+source_set("f32-raddstoreexpminusmax_ppc64") {
++source_set("f32-qu8-vcvt_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -68465,7 +70377,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-scalar-rr2-p5-u4-acc2.c"
++    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u4.c",
++    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-lrintf-u4.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -68504,7 +70417,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-raddstoreexpminusmax_ppc64_standalone") {
++source_set("f32-qu8-vcvt_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -68516,7 +70429,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-scalar-rr2-p5-u4-acc2.c"
++    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u4.c",
++    "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-lrintf-u4.c"
 +  ]
  
 -    sources = [
@@ -68565,7 +70479,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-neon-2x2.c",
 -      "src/src/f32-conv-hwc2chw/f32-conv-hwc2chw-3x3s2p1c3x4-scalar-1x1.c",
 -    ]
-+source_set("f32-rdminmax_ppc64") {
++source_set("f32-raddstoreexpminusmax_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -68584,8 +70498,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-scalar-c2.c",
-+    "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-scalar-c2.c"
++    "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-scalar-rr2-p5-u4-acc2.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -68651,7 +70564,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-dwconv2d-chw/gen/f32-dwconv2d-chw-5x5s2p2-minmax-scalar-2x1-acc2.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-rdminmax_ppc64_standalone") {
++source_set("f32-raddstoreexpminusmax_ppc64_standalone") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -68670,8 +70583,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-scalar-c2.c",
-+    "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-scalar-c2.c"
++    "src/src/f32-raddstoreexpminusmax/gen/f32-raddstoreexpminusmax-scalar-rr2-p5-u4-acc2.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -68731,7 +70643,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f32-dwconv_arch=armv8.2-a+fp16+dotprod") {
 -    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
-+source_set("f32-rdsum_ppc64") {
++source_set("f32-rdminmax_ppc64") {
 +  cflags = [
  
 -    asmflags = cflags
@@ -68743,7 +70655,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-scalar.c"
++    "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-scalar-c2.c",
++    "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-scalar-c2.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -68777,7 +70690,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    asmflags = cflags
 +# This is a target that cannot depend on //base.
-+source_set("f32-rdsum_ppc64_standalone") {
++source_set("f32-rdminmax_ppc64_standalone") {
 +  cflags = [
  
 -    sources = [
@@ -68798,7 +70711,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-scalar.c"
++    "src/src/f32-rdminmax/gen/f32-rdmax-2p2x-scalar-c2.c",
++    "src/src/f32-rdminmax/gen/f32-rdmin-2p2x-scalar-c2.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -68896,7 +70810,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("f32-rminmax_ppc64") {
++source_set("f32-rdsum_ppc64") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -68908,9 +70822,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rminmax/gen/f32-rmax-scalar-u4-acc4.c",
-+    "src/src/f32-rminmax/gen/f32-rmin-scalar-u4-acc4.c",
-+    "src/src/f32-rminmax/gen/f32-rminmax-scalar-u4-acc4.c"
++    "src/src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-scalar.c"
 +  ]
  
 -  source_set("f32-f16-vcvt_arm64") {
@@ -68948,7 +70860,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("f32-rminmax_ppc64_standalone") {
++source_set("f32-rdsum_ppc64_standalone") {
 +  cflags = [
  
 -  # This is a target that cannot depend on //base.
@@ -68975,9 +70887,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rminmax/gen/f32-rmax-scalar-u4-acc4.c",
-+    "src/src/f32-rminmax/gen/f32-rmin-scalar-u4-acc4.c",
-+    "src/src/f32-rminmax/gen/f32-rminmax-scalar-u4-acc4.c"
++    "src/src/f32-rdsum/gen/f32-rdsum-7p7x-minmax-scalar.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -69081,7 +70991,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("f32-rsum_ppc64") {
++source_set("f32-rminmax_ppc64") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -69103,9 +71013,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-aarch64-neonfma-lane-ld128.c",
 -      "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-neon-lane-ld64.c",
 -      "src/src/f32-gemm/gen/f32-gemm-1x8s4-minmax-neonfma.c",
--      "src/src/f32-gemm/gen/f32-gemm-2x4-minmax-scalar.c",
--      "src/src/f32-gemm/gen/f32-gemm-2x4-relu-scalar.c",
--      "src/src/f32-gemm/gen/f32-gemm-2x4-scalar.c",
 -      "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-aarch64-neonfma-lane-ld64.c",
 -      "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-neon-lane-ld64.c",
 -      "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-scalar.c",
@@ -69120,7 +71027,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rsum/gen/f32-rsum-scalar-u4-acc4.c"
++    "src/src/f32-rminmax/gen/f32-rmax-scalar-u4-acc4.c",
++    "src/src/f32-rminmax/gen/f32-rmin-scalar-u4-acc4.c",
++    "src/src/f32-rminmax/gen/f32-rminmax-scalar-u4-acc4.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -69158,9 +71067,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-aarch64-neonfma-lane-ld128.c",
 -      "src/src/f32-gemm/gen/f32-gemm-1x8-minmax-neon-lane-ld64.c",
 -      "src/src/f32-gemm/gen/f32-gemm-1x8s4-minmax-neonfma.c",
--      "src/src/f32-gemm/gen/f32-gemm-2x4-minmax-scalar.c",
--      "src/src/f32-gemm/gen/f32-gemm-2x4-relu-scalar.c",
--      "src/src/f32-gemm/gen/f32-gemm-2x4-scalar.c",
 -      "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-aarch64-neonfma-lane-ld64.c",
 -      "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-neon-lane-ld64.c",
 -      "src/src/f32-gemm/gen/f32-gemm-4x2-minmax-scalar.c",
@@ -69187,7 +71093,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-rsum_ppc64_standalone") {
++source_set("f32-rminmax_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -69199,7 +71105,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-rsum/gen/f32-rsum-scalar-u4-acc4.c"
++    "src/src/f32-rminmax/gen/f32-rmax-scalar-u4-acc4.c",
++    "src/src/f32-rminmax/gen/f32-rmin-scalar-u4-acc4.c",
++    "src/src/f32-rminmax/gen/f32-rminmax-scalar-u4-acc4.c"
 +  ]
  
 -  source_set("f32-ibilinear-chw_arm64") {
@@ -69260,7 +71168,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("f32-spmm_ppc64") {
++source_set("f32-rsum_ppc64") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -69272,9 +71180,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-spmm/gen/f32-spmm-8x1-minmax-scalar.c",
-+    "src/src/f32-spmm/gen/f32-spmm-8x2-minmax-scalar.c",
-+    "src/src/f32-spmm/gen/f32-spmm-8x4-minmax-scalar.c"
++    "src/src/f32-rsum/gen/f32-rsum-scalar-u4-acc4.c"
 +  ]
  
 -  source_set("f32-ibilinear_arm64") {
@@ -69312,7 +71218,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("f32-spmm_ppc64_standalone") {
++source_set("f32-rsum_ppc64_standalone") {
 +  cflags = [
  
 -  # This is a target that cannot depend on //base.
@@ -69339,9 +71245,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-spmm/gen/f32-spmm-8x1-minmax-scalar.c",
-+    "src/src/f32-spmm/gen/f32-spmm-8x2-minmax-scalar.c",
-+    "src/src/f32-spmm/gen/f32-spmm-8x4-minmax-scalar.c"
++    "src/src/f32-rsum/gen/f32-rsum-scalar-u4-acc4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -69443,7 +71347,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("f32-vapproxgelu_ppc64") {
++source_set("f32-spmm_ppc64") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -69465,9 +71369,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-aarch64-neonfma-lane-ld128.c",
 -      "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-neon-lane-ld64.c",
 -      "src/src/f32-igemm/gen/f32-igemm-1x8s4-minmax-neonfma.c",
--      "src/src/f32-igemm/gen/f32-igemm-2x4-minmax-scalar.c",
--      "src/src/f32-igemm/gen/f32-igemm-2x4-relu-scalar.c",
--      "src/src/f32-igemm/gen/f32-igemm-2x4-scalar.c",
 -      "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-aarch64-neonfma-lane-ld64.c",
 -      "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-neon-lane-ld64.c",
 -      "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-scalar.c",
@@ -69482,7 +71383,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-scalar-rational-12-10-div.c"
++    "src/src/f32-spmm/gen/f32-spmm-8x1-minmax-scalar.c",
++    "src/src/f32-spmm/gen/f32-spmm-8x2-minmax-scalar.c",
++    "src/src/f32-spmm/gen/f32-spmm-8x4-minmax-scalar.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -69520,9 +71423,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-aarch64-neonfma-lane-ld128.c",
 -      "src/src/f32-igemm/gen/f32-igemm-1x8-minmax-neon-lane-ld64.c",
 -      "src/src/f32-igemm/gen/f32-igemm-1x8s4-minmax-neonfma.c",
--      "src/src/f32-igemm/gen/f32-igemm-2x4-minmax-scalar.c",
--      "src/src/f32-igemm/gen/f32-igemm-2x4-relu-scalar.c",
--      "src/src/f32-igemm/gen/f32-igemm-2x4-scalar.c",
 -      "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-aarch64-neonfma-lane-ld64.c",
 -      "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-neon-lane-ld64.c",
 -      "src/src/f32-igemm/gen/f32-igemm-4x2-minmax-scalar.c",
@@ -69549,7 +71449,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vapproxgelu_ppc64_standalone") {
++source_set("f32-spmm_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -69561,7 +71461,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-scalar-rational-12-10-div.c"
++    "src/src/f32-spmm/gen/f32-spmm-8x1-minmax-scalar.c",
++    "src/src/f32-spmm/gen/f32-spmm-8x2-minmax-scalar.c",
++    "src/src/f32-spmm/gen/f32-spmm-8x4-minmax-scalar.c"
 +  ]
  
 -  source_set("f32-maxpool_arm64") {
@@ -69622,7 +71524,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 -
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("f32-vbinary_ppc64") {
++source_set("f32-vapproxgelu_ppc64") {
 +  cflags = [
  
 -    if (!(is_android && use_order_profiling)) {
@@ -69657,25 +71559,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vbinary/gen/f32-vadd-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vaddc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vdiv-scalar-u2.c",
-+    "src/src/f32-vbinary/gen/f32-vdivc-scalar-u2.c",
-+    "src/src/f32-vbinary/gen/f32-vmax-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmaxc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmin-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vminc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmul-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmulc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vprelu-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vpreluc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vrdivc-scalar-u2.c",
-+    "src/src/f32-vbinary/gen/f32-vrpreluc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vrsubc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsqrdiff-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsqrdiffc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsub-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsubc-scalar-u8.c"
++    "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-scalar-rational-12-10-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -69720,7 +71604,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vbinary_ppc64_standalone") {
++source_set("f32-vapproxgelu_ppc64_standalone") {
 +  cflags = [
  
 -    if (!(is_android && use_order_profiling)) {
@@ -69733,25 +71617,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vbinary/gen/f32-vadd-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vaddc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vdiv-scalar-u2.c",
-+    "src/src/f32-vbinary/gen/f32-vdivc-scalar-u2.c",
-+    "src/src/f32-vbinary/gen/f32-vmax-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmaxc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmin-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vminc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmul-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vmulc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vprelu-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vpreluc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vrdivc-scalar-u2.c",
-+    "src/src/f32-vbinary/gen/f32-vrpreluc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vrsubc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsqrdiff-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsqrdiffc-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsub-scalar-u8.c",
-+    "src/src/f32-vbinary/gen/f32-vsubc-scalar-u8.c"
++    "src/src/f32-vapproxgelu/gen/f32-vapproxgelu-scalar-rational-12-10-div.c"
 +  ]
  
 -    asmflags = cflags
@@ -69795,7 +71661,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
 -
 -    asmflags = cflags
-+source_set("f32-vclamp_ppc64") {
++source_set("f32-vbinary_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -69818,7 +71684,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vclamp/gen/f32-vclamp-scalar-u4.c"
++    "src/src/f32-vbinary/gen/f32-vadd-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vaddc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vdiv-scalar-u2.c",
++    "src/src/f32-vbinary/gen/f32-vdivc-scalar-u2.c",
++    "src/src/f32-vbinary/gen/f32-vmax-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmaxc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmin-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vminc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmul-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmulc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vprelu-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vpreluc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vrdivc-scalar-u2.c",
++    "src/src/f32-vbinary/gen/f32-vrpreluc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vrsubc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsqrdiff-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsqrdiffc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsub-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsubc-scalar-u8.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -69866,7 +71750,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("f32-vclamp_ppc64_standalone") {
++source_set("f32-vbinary_ppc64_standalone") {
 +  cflags = [
  
 -  # This is a target that cannot depend on //base.
@@ -69896,7 +71780,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vclamp/gen/f32-vclamp-scalar-u4.c"
++    "src/src/f32-vbinary/gen/f32-vadd-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vaddc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vdiv-scalar-u2.c",
++    "src/src/f32-vbinary/gen/f32-vdivc-scalar-u2.c",
++    "src/src/f32-vbinary/gen/f32-vmax-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmaxc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmin-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vminc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmul-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vmulc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vprelu-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vpreluc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vrdivc-scalar-u2.c",
++    "src/src/f32-vbinary/gen/f32-vrpreluc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vrsubc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsqrdiff-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsqrdiffc-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsub-scalar-u8.c",
++    "src/src/f32-vbinary/gen/f32-vsubc-scalar-u8.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -69916,7 +71818,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-neon-u32.c",
 -      "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-neonv8-u32.c",
--      "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u1.c",
 -      "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u4.c",
 -      "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-lrintf-u4.c",
 -    ]
@@ -69953,7 +71854,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-neon-u32.c",
 -      "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-neonv8-u32.c",
--      "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u1.c",
 -      "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-imagic-u4.c",
 -      "src/src/f32-qs8-vcvt/gen/f32-qs8-vcvt-scalar-lrintf-u4.c",
 -    ]
@@ -69968,7 +71868,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("f32-vcmul_ppc64") {
++source_set("f32-vclamp_ppc64") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -69986,13 +71886,12 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-neon-u32.c",
 -      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-neonv8-u32.c",
--      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u1.c",
 -      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u4.c",
 -      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-lrintf-u4.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vcmul/gen/f32-vcmul-scalar-u4.c"
++    "src/src/f32-vclamp/gen/f32-vclamp-scalar.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -70026,7 +71925,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-neon-u32.c",
 -      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-neonv8-u32.c",
--      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u1.c",
 -      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-imagic-u4.c",
 -      "src/src/f32-qu8-vcvt/gen/f32-qu8-vcvt-scalar-lrintf-u4.c",
 -    ]
@@ -70044,7 +71942,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vcmul_ppc64_standalone") {
++source_set("f32-vclamp_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -70056,7 +71954,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vcmul/gen/f32-vcmul-scalar-u4.c"
++    "src/src/f32-vclamp/gen/f32-vclamp-scalar.c"
 +  ]
  
 -  source_set("f32-raddstoreexpminusmax_arm64") {
@@ -70115,7 +72013,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("f32-vcopysign_ppc64") {
++source_set("f32-vcmul_ppc64") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -70138,9 +72036,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vcopysign/gen/f32-vcopysign-scalar.c",
-+    "src/src/f32-vcopysign/gen/f32-vcopysignc-scalar.c",
-+    "src/src/f32-vcopysign/gen/f32-vrcopysignc-scalar.c"
++    "src/src/f32-vcmul/gen/f32-vcmul-scalar-u4.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -70191,7 +72087,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vcopysign_ppc64_standalone") {
++source_set("f32-vcmul_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -70203,9 +72099,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vcopysign/gen/f32-vcopysign-scalar.c",
-+    "src/src/f32-vcopysign/gen/f32-vcopysignc-scalar.c",
-+    "src/src/f32-vcopysign/gen/f32-vrcopysignc-scalar.c"
++    "src/src/f32-vcmul/gen/f32-vcmul-scalar-u4.c"
 +  ]
  
 -  source_set("f32-rdsum_arm64") {
@@ -70264,7 +72158,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("f32-vcos_ppc64") {
++source_set("f32-vcopysign_ppc64") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -70289,7 +72183,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vcos/gen/f32-vcos-scalar-rational-5-4-div.c"
++    "src/src/f32-vcopysign/gen/f32-vcopysign-scalar.c",
++    "src/src/f32-vcopysign/gen/f32-vcopysignc-scalar.c",
++    "src/src/f32-vcopysign/gen/f32-vrcopysignc-scalar.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -70342,7 +72238,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vcos_ppc64_standalone") {
++source_set("f32-vcopysign_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -70354,7 +72250,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vcos/gen/f32-vcos-scalar-rational-5-4-div.c"
++    "src/src/f32-vcopysign/gen/f32-vcopysign-scalar.c",
++    "src/src/f32-vcopysign/gen/f32-vcopysignc-scalar.c",
++    "src/src/f32-vcopysign/gen/f32-vrcopysignc-scalar.c"
 +  ]
  
 -  source_set("f32-rsum_arm64") {
@@ -70396,7 +72294,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  # This is a target that cannot depend on //base.
 -  source_set("f32-rsum_arm64_standalone") {
 -    cflags = []
-+source_set("f32-velu_ppc64") {
++source_set("f32-vcos_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -70418,8 +72316,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u2.c",
-+    "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u4.c"
++    "src/src/f32-vcos/gen/f32-vcos-scalar-rational-5-4-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -70468,7 +72365,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("f32-velu_ppc64_standalone") {
++source_set("f32-vcos_ppc64_standalone") {
 +  cflags = [
  
 -  # This is a target that cannot depend on //base.
@@ -70499,8 +72396,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u2.c",
-+    "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u4.c"
++    "src/src/f32-vcos/gen/f32-vcos-scalar-rational-5-4-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -70543,7 +72439,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("f32-vexp_ppc64") {
++source_set("f32-velu_ppc64") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -70555,7 +72451,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vexp/gen/f32-vexp-scalar-rational-3-2-div.c"
++    "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u4.c"
 +  ]
  
 -    sources = [
@@ -70638,7 +72534,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-vbinary/gen/f32-vsubc-scalar-u8.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vexp_ppc64_standalone") {
++source_set("f32-velu_ppc64_standalone") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -70657,7 +72553,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vexp/gen/f32-vexp-scalar-rational-3-2-div.c"
++    "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u4.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -70739,13 +72635,13 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f32-vclamp_arm64") {
 -    cflags = []
-+source_set("f32-vgelu_ppc64") {
++source_set("f32-vexp_ppc64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vclamp/gen/f32-vclamp-neon-u16.c",
--      "src/src/f32-vclamp/gen/f32-vclamp-scalar-u4.c",
+-      "src/src/f32-vclamp/gen/f32-vclamp-neon.c",
+-      "src/src/f32-vclamp/gen/f32-vclamp-scalar.c",
 -    ]
 +  ]
  
@@ -70761,7 +72657,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vgelu/gen/f32-vgelu-scalar-rational-12-10-div.c"
++    "src/src/f32-vexp/gen/f32-vexp-scalar-rational-3-2-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -70782,8 +72678,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vclamp/gen/f32-vclamp-neon-u16.c",
--      "src/src/f32-vclamp/gen/f32-vclamp-scalar-u4.c",
+-      "src/src/f32-vclamp/gen/f32-vclamp-neon.c",
+-      "src/src/f32-vclamp/gen/f32-vclamp-scalar.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -70799,7 +72695,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vgelu_ppc64_standalone") {
++source_set("f32-vexp_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -70811,7 +72707,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vgelu/gen/f32-vgelu-scalar-rational-12-10-div.c"
++    "src/src/f32-vexp/gen/f32-vexp-scalar-rational-3-2-div.c"
 +  ]
  
 -  source_set("f32-vcmul_arm64") {
@@ -70853,7 +72749,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  # This is a target that cannot depend on //base.
 -  source_set("f32-vcmul_arm64_standalone") {
 -    cflags = []
-+source_set("f32-vhswish_ppc64") {
++source_set("f32-vgelu_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -70875,7 +72771,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vhswish/gen/f32-vhswish-scalar-u4.c"
++    "src/src/f32-vgelu/gen/f32-vgelu-scalar-rational-12-10-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -70923,7 +72819,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("f32-vhswish_ppc64_standalone") {
++source_set("f32-vgelu_ppc64_standalone") {
 +  cflags = [
  
 -  # This is a target that cannot depend on //base.
@@ -70953,7 +72849,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vhswish/gen/f32-vhswish-scalar-u4.c"
++    "src/src/f32-vgelu/gen/f32-vgelu-scalar-rational-12-10-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -70996,7 +72892,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("f32-vlog_ppc64") {
++source_set("f32-vhswish_ppc64") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -71008,7 +72904,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vlog/gen/f32-vlog-scalar-rational-3-3-div.c"
++    "src/src/f32-vhswish/gen/f32-vhswish-scalar.c"
 +  ]
  
 -    sources = [
@@ -71054,11 +72950,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-velu/gen/f32-velu-neon-rr2-lut16-p3-u8.c",
 -      "src/src/f32-velu/gen/f32-velu-neonfma-rr1-lut16-p3-u16.c",
 -      "src/src/f32-velu/gen/f32-velu-neonfma-rr1-p6-u8.c",
--      "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u2.c",
 -      "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u4.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vlog_ppc64_standalone") {
++source_set("f32-vhswish_ppc64_standalone") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -71077,7 +72972,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vlog/gen/f32-vlog-scalar-rational-3-3-div.c"
++    "src/src/f32-vhswish/gen/f32-vhswish-scalar.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -71089,7 +72984,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-velu/gen/f32-velu-neon-rr2-lut16-p3-u8.c",
 -      "src/src/f32-velu/gen/f32-velu-neonfma-rr1-lut16-p3-u16.c",
 -      "src/src/f32-velu/gen/f32-velu-neonfma-rr1-p6-u8.c",
--      "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u2.c",
 -      "src/src/f32-velu/gen/f32-velu-scalar-rr2-lut16-p3-u4.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -71126,7 +73020,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f32-vexp_arm64") {
 -    cflags = []
-+source_set("f32-vlrelu_ppc64") {
++source_set("f32-vlog_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -71148,7 +73042,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vlrelu/gen/f32-vlrelu-scalar-u4.c"
++    "src/src/f32-vlog/gen/f32-vlog-scalar-rational-3-3-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -71186,7 +73080,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vlrelu_ppc64_standalone") {
++source_set("f32-vlog_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -71198,7 +73092,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vlrelu/gen/f32-vlrelu-scalar-u4.c"
++    "src/src/f32-vlog/gen/f32-vlog-scalar-rational-3-3-div.c"
 +  ]
  
 -  source_set("f32-vgelu_arm64") {
@@ -71246,7 +73140,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-vgelu/gen/f32-vgelu-neon-rational-12-10-div.c",
 -      "src/src/f32-vgelu/gen/f32-vgelu-scalar-rational-12-10-div.c",
 -    ]
-+source_set("f32-vmulcaddc_ppc64") {
++source_set("f32-vlrelu_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -71264,7 +73158,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c1-minmax-scalar-2x.c"
++    "src/src/f32-vlrelu/gen/f32-vlrelu-scalar-u4.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -71286,8 +73180,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vhswish/gen/f32-vhswish-neon-u16.c",
--      "src/src/f32-vhswish/gen/f32-vhswish-scalar-u4.c",
+-      "src/src/f32-vhswish/gen/f32-vhswish-neon.c",
+-      "src/src/f32-vhswish/gen/f32-vhswish-scalar.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -71303,7 +73197,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vmulcaddc_ppc64_standalone") {
++source_set("f32-vlrelu_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -71315,13 +73209,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c1-minmax-scalar-2x.c"
++    "src/src/f32-vlrelu/gen/f32-vlrelu-scalar-u4.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vhswish/gen/f32-vhswish-neon-u16.c",
--      "src/src/f32-vhswish/gen/f32-vhswish-scalar-u4.c",
+-      "src/src/f32-vhswish/gen/f32-vhswish-neon.c",
+-      "src/src/f32-vhswish/gen/f32-vhswish-scalar.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -71357,7 +73251,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f32-vlog_arm64") {
 -    cflags = []
-+source_set("f32-vrelu_ppc64") {
++source_set("f32-vmulcaddc_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -71379,7 +73273,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrelu/gen/f32-vrelu-scalar-u8.c"
++    "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c1-minmax-scalar-2x.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -71417,7 +73311,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vrelu_ppc64_standalone") {
++source_set("f32-vmulcaddc_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -71429,7 +73323,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrelu/gen/f32-vrelu-scalar-u8.c"
++    "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c1-minmax-scalar-2x.c"
 +  ]
  
 -  source_set("f32-vlrelu_arm64") {
@@ -71471,7 +73365,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  # This is a target that cannot depend on //base.
 -  source_set("f32-vlrelu_arm64_standalone") {
 -    cflags = []
-+source_set("f32-vrnd_ppc64") {
++source_set("f32-vrelu_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -71493,14 +73387,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u4.c"
++    "src/src/f32-vrelu/gen/f32-vrelu-scalar.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -71531,7 +73418,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-vmulcaddc/gen/f32-vmulcaddc-c4-minmax-neonfma-2x.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vrnd_ppc64_standalone") {
++source_set("f32-vrelu_ppc64_standalone") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -71550,14 +73437,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u4.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c",
-+    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u4.c"
++    "src/src/f32-vrelu/gen/f32-vrelu-scalar.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -71607,7 +73487,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrelu/gen/f32-vrelu-scalar-u8.c",
+-      "src/src/f32-vrelu/gen/f32-vrelu-neon.c",
+-      "src/src/f32-vrelu/gen/f32-vrelu-scalar.c",
 -    ]
 -
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -71620,7 +73501,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("f32-vrsqrt_ppc64") {
++source_set("f32-vrnd_ppc64") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -71632,13 +73513,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u4.c"
++    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrelu/gen/f32-vrelu-scalar-u8.c",
+-      "src/src/f32-vrelu/gen/f32-vrelu-neon.c",
+-      "src/src/f32-vrelu/gen/f32-vrelu-scalar.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -71678,22 +73562,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-vrnd/gen/f32-vrndd-neon-u8.c",
 -      "src/src/f32-vrnd/gen/f32-vrndd-neonv8-u8.c",
 -      "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
--      "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u4.c",
 -      "src/src/f32-vrnd/gen/f32-vrndne-neon-u8.c",
 -      "src/src/f32-vrnd/gen/f32-vrndne-neonv8-u8.c",
 -      "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
--      "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u4.c",
 -      "src/src/f32-vrnd/gen/f32-vrndu-neon-u8.c",
 -      "src/src/f32-vrnd/gen/f32-vrndu-neonv8-u8.c",
 -      "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
--      "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u4.c",
 -      "src/src/f32-vrnd/gen/f32-vrndz-neon-u8.c",
 -      "src/src/f32-vrnd/gen/f32-vrndz-neonv8-u8.c",
 -      "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c",
--      "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u4.c",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vrsqrt_ppc64_standalone") {
++source_set("f32-vrnd_ppc64_standalone") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -71712,8 +73592,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
-+    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u4.c"
++    "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
++    "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -71725,19 +73607,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-vrnd/gen/f32-vrndd-neon-u8.c",
 -      "src/src/f32-vrnd/gen/f32-vrndd-neonv8-u8.c",
 -      "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u1.c",
--      "src/src/f32-vrnd/gen/f32-vrndd-scalar-libm-u4.c",
 -      "src/src/f32-vrnd/gen/f32-vrndne-neon-u8.c",
 -      "src/src/f32-vrnd/gen/f32-vrndne-neonv8-u8.c",
 -      "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u1.c",
--      "src/src/f32-vrnd/gen/f32-vrndne-scalar-libm-u4.c",
 -      "src/src/f32-vrnd/gen/f32-vrndu-neon-u8.c",
 -      "src/src/f32-vrnd/gen/f32-vrndu-neonv8-u8.c",
 -      "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u1.c",
--      "src/src/f32-vrnd/gen/f32-vrndu-scalar-libm-u4.c",
 -      "src/src/f32-vrnd/gen/f32-vrndz-neon-u8.c",
 -      "src/src/f32-vrnd/gen/f32-vrndz-neonv8-u8.c",
 -      "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u1.c",
--      "src/src/f32-vrnd/gen/f32-vrndz-scalar-libm-u4.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -71770,23 +73648,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    assert_no_deps = [ "//base" ]
    }
 +}
-+
-+source_set("f32-vsigmoid_ppc64") {
-+  cflags = [
  
 -  source_set("f32-vrsqrt_arm64") {
 -    cflags = []
-+  ]
++source_set("f32-vrsqrt_ppc64") {
++  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrsqrt/gen/f32-vrsqrt-neon-rsqrt-u16.c",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-neon-rsqrt.c",
 -      "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
--      "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u4.c",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-sqrt.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -71799,18 +73672,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-sqrt.c"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set("f32-vrsqrt_arm64_standalone") {
@@ -71818,12 +73690,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vrsqrt/gen/f32-vrsqrt-neon-rsqrt-u16.c",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-neon-rsqrt.c",
 -      "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
--      "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u4.c",
+-      "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-sqrt.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -71835,27 +73711,27 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("f32-vsigmoid_ppc64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
-+  ]
++# This is a target that cannot depend on //base.
++source_set("f32-vrsqrt_ppc64_standalone") {
++  cflags = [
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c"
 +  ]
  
 -  source_set("f32-vsigmoid_arm64") {
 -    cflags = []
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-rsqrt-u1.c",
++    "src/src/f32-vrsqrt/gen/f32-vrsqrt-scalar-sqrt.c"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
@@ -71863,12 +73739,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-vsigmoid/gen/f32-vsigmoid-neonfma-rr1-lut64-p2-nr2recps-u16.c",
 -      "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -71880,9 +73753,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
++
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
    }
@@ -71898,7 +73778,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/f32-vsigmoid/gen/f32-vsigmoid-neonfma-rr1-lut64-p2-nr2recps-u16.c",
 -      "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c",
 -    ]
-+source_set("f32-vsin_ppc64") {
++source_set("f32-vsigmoid_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -71916,7 +73796,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsin/gen/f32-vsin-scalar-rational-5-4-div.c"
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -71955,7 +73835,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsin_ppc64_standalone") {
++source_set("f32-vsigmoid_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -71967,7 +73847,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsin/gen/f32-vsin-scalar-rational-5-4-div.c"
++    "src/src/f32-vsigmoid/gen/f32-vsigmoid-scalar-rr2-lut64-p2-div-u2.c"
 +  ]
  
 -    sources = [
@@ -72009,13 +73889,13 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("f32-vsqrt_arm64") {
 -    cflags = []
-+source_set("f32-vsqrt_ppc64") {
++source_set("f32-vsin_ppc64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsqrt/gen/f32-vsqrt-aarch64-neon-sqrt-u4.c",
--      "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt-u1.c",
+-      "src/src/f32-vsqrt/gen/f32-vsqrt-aarch64-neon-sqrt.c",
+-      "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt.c",
 -    ]
 +  ]
  
@@ -72031,7 +73911,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt-u1.c"
++    "src/src/f32-vsin/gen/f32-vsin-scalar-rational-5-4-div.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -72052,8 +73932,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/f32-vsqrt/gen/f32-vsqrt-aarch64-neon-sqrt-u4.c",
--      "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt-u1.c",
+-      "src/src/f32-vsqrt/gen/f32-vsqrt-aarch64-neon-sqrt.c",
+-      "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -72069,7 +73949,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vsqrt_ppc64_standalone") {
++source_set("f32-vsin_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -72081,7 +73961,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt-u1.c"
++    "src/src/f32-vsin/gen/f32-vsin-scalar-rational-5-4-div.c"
 +  ]
  
 -  source_set("f32-vtanh_arm64") {
@@ -72123,7 +74003,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  # This is a target that cannot depend on //base.
 -  source_set("f32-vtanh_arm64_standalone") {
 -    cflags = []
-+source_set("f32-vtanh_ppc64") {
++source_set("f32-vsqrt_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -72145,7 +74025,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vtanh/gen/f32-vtanh-scalar-rational-9-8-div.c"
++    "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -72193,7 +74073,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("f32-vtanh_ppc64_standalone") {
++source_set("f32-vsqrt_ppc64_standalone") {
 +  cflags = [
  
 -  # This is a target that cannot depend on //base.
@@ -72223,7 +74103,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vtanh/gen/f32-vtanh-scalar-rational-9-8-div.c"
++    "src/src/f32-vsqrt/gen/f32-vsqrt-scalar-sqrt.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -72316,7 +74196,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/operators/unary-elementwise-nc.c",
 -      "src/src/operators/unpooling-nhwc.c",
 -    ]
-+source_set("f32-vunary_ppc64") {
++source_set("f32-vtanh_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -72334,9 +74214,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vunary/gen/f32-vabs-scalar.c",
-+    "src/src/f32-vunary/gen/f32-vneg-scalar.c",
-+    "src/src/f32-vunary/gen/f32-vsqr-scalar.c"
++    "src/src/f32-vtanh/gen/f32-vtanh-scalar-rational-9-8-div.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -72347,8 +74225,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+dotprod+fp16") {
--    cflags = [ "-march=armv8.2-a+dotprod+fp16" ]
+-  source_set("pf16-gemm_arch=armv8.2-a+sve+sve2") {
+-    cflags = [ "-march=armv8.2-a+sve+sve2" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -72358,8 +74236,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16c4-minmax-neondotfp16arith.c",
--      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-4x16c4-minmax-neondotfp16arith.c",
+-      "src/src/pf16-gemm/pf16-gemm-1x32c2-minmax-neonsme2.c",
+-      "src/src/pf16-gemm/pf16-gemm-32x32c2-minmax-neonsme2.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -72375,7 +74253,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("f32-vunary_ppc64_standalone") {
++source_set("f32-vtanh_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -72383,19 +74261,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+dotprod+fp16_standalone") {
--    cflags = [ "-march=armv8.2-a+dotprod+fp16" ]
+-  source_set("pf16-gemm_arch=armv8.2-a+sve+sve2_standalone") {
+-    cflags = [ "-march=armv8.2-a+sve+sve2" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/f32-vunary/gen/f32-vabs-scalar.c",
-+    "src/src/f32-vunary/gen/f32-vneg-scalar.c",
-+    "src/src/f32-vunary/gen/f32-vsqr-scalar.c"
++    "src/src/f32-vtanh/gen/f32-vtanh-scalar-rational-9-8-div.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16c4-minmax-neondotfp16arith.c",
--      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-4x16c4-minmax-neondotfp16arith.c",
+-      "src/src/pf16-gemm/pf16-gemm-1x32c2-minmax-neonsme2.c",
+-      "src/src/pf16-gemm/pf16-gemm-32x32c2-minmax-neonsme2.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -72429,16 +74305,103 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+fp16") {
--    cflags = [ "-march=armv8.2-a+fp16" ]
-+source_set("operators_ppc64") {
+-  source_set("pf32-gemm_arch=armv8.2-a+sve+sve2") {
+-    cflags = [ "-march=armv8.2-a+sve+sve2" ]
++source_set("f32-vunary_ppc64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16-minmax-neonfp16arith-mlal-lane.c",
--      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-6x16-minmax-neonfp16arith-mlal-lane.c",
+-      "src/src/pf32-gemm/pf32-gemm-1x32-minmax-neonsme2.c",
+-      "src/src/pf32-gemm/pf32-gemm-32x32-minmax-neonsme2.c",
+-    ]
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vunary/gen/f32-vabs-scalar.c",
++    "src/src/f32-vunary/gen/f32-vneg-scalar.c",
++    "src/src/f32-vunary/gen/f32-vsqr-scalar.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("pf32-gemm_arch=armv8.2-a+sve+sve2_standalone") {
+-    cflags = [ "-march=armv8.2-a+sve+sve2" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/pf32-gemm/pf32-gemm-1x32-minmax-neonsme2.c",
+-      "src/src/pf32-gemm/pf32-gemm-32x32-minmax-neonsme2.c",
 -    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("f32-vunary_ppc64_standalone") {
++  cflags = [
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/f32-vunary/gen/f32-vabs-scalar.c",
++    "src/src/f32-vunary/gen/f32-vneg-scalar.c",
++    "src/src/f32-vunary/gen/f32-vsqr-scalar.c"
++  ]
+ 
+-  source_set("pqs8-qc8w-gemm_arch=armv8.2-a+sve+sve2") {
+-    cflags = [ "-march=armv8.2-a+sve+sve2" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/pqs8-qc8w-gemm/pqs8-qc8w-gemm-1x32c4-minmax-neonsme2.c",
+-      "src/src/pqs8-qc8w-gemm/pqs8-qc8w-gemm-32x32c4-minmax-neonsme2.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -72451,6 +74414,39 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("pqs8-qc8w-gemm_arch=armv8.2-a+sve+sve2_standalone") {
+-    cflags = [ "-march=armv8.2-a+sve+sve2" ]
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/pqs8-qc8w-gemm/pqs8-qc8w-gemm-1x32c4-minmax-neonsme2.c",
+-      "src/src/pqs8-qc8w-gemm/pqs8-qc8w-gemm-32x32c4-minmax-neonsme2.c",
+-    ]
++source_set("operators_ppc64") {
++  cflags = [
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/operators/argmax-pooling-nhwc.c",
@@ -72476,15 +74472,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/operators/unpooling-nhwc.c"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 -  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+fp16_standalone") {
--    cflags = [ "-march=armv8.2-a+fp16" ]
+-  source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+dotprod+fp16") {
+-    cflags = [ "-march=armv8.2-a+dotprod+fp16" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -72494,8 +74491,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16-minmax-neonfp16arith-mlal-lane.c",
--      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-6x16-minmax-neonfp16arith-mlal-lane.c",
+-      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16c4-minmax-neondotfp16arith.c",
+-      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-4x16c4-minmax-neondotfp16arith.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -72508,19 +74505,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
 +source_set("operators_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
+-  }
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+dotprod+fp16_standalone") {
+-    cflags = [ "-march=armv8.2-a+dotprod+fp16" ]
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/operators/argmax-pooling-nhwc.c",
@@ -72546,6 +74543,122 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/operators/unpooling-nhwc.c"
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16c4-minmax-neondotfp16arith.c",
+-      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-4x16c4-minmax-neondotfp16arith.c",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+fp16") {
+-    cflags = [ "-march=armv8.2-a+fp16" ]
++source_set("qd8-f32-qb4w-gemm_ppc64") {
++  cflags = [
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16-minmax-neonfp16arith-mlal-lane.c",
+-      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-6x16-minmax-neonfp16arith-mlal-lane.c",
+-    ]
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4-minmax-scalar.c",
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4-minmax-scalar.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+fp16_standalone") {
+-    cflags = [ "-march=armv8.2-a+fp16" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16-minmax-neonfp16arith-mlal-lane.c",
+-      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-6x16-minmax-neonfp16arith-mlal-lane.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("qd8-f32-qb4w-gemm_ppc64_standalone") {
++  cflags = [
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4-minmax-scalar.c",
++    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4-minmax-scalar.c"
++  ]
+ 
 -  source_set("qd8-f16-qb4w-gemm_arch=armv8.2-a+i8mm+fp16") {
 -    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
@@ -72591,7 +74704,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-1x16c8-minmax-neoni8mm.c",
 -      "src/src/qd8-f16-qb4w-gemm/gen/qd8-f16-qb4w-gemm-4x16c8-minmax-neoni8mm.c",
 -    ]
-+source_set("qd8-f32-qb4w-gemm_ppc64") {
++source_set("qd8-f32-qc4w-gemm_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -72609,8 +74722,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4-minmax-scalar.c"
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4-minmax-scalar.c",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4-minmax-scalar.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -72649,7 +74762,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qb4w-gemm_ppc64_standalone") {
++source_set("qd8-f32-qc4w-gemm_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -72661,8 +74774,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = [ "-march=armv8.2-a+dotprod+fp16" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4-minmax-scalar.c"
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4-minmax-scalar.c",
++    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4-minmax-scalar.c"
 +  ]
  
 -    sources = [
@@ -72704,7 +74817,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("qd8-f16-qc4w-gemm_arch=armv8.2-a+fp16") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
-+source_set("qd8-f32-qc4w-gemm_ppc64") {
++source_set("qd8-f32-qc8w-gemm_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -72726,8 +74839,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4-minmax-scalar.c"
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x2-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4-minmax-scalar.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -72765,7 +74879,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qc4w-gemm_ppc64_standalone") {
++source_set("qd8-f32-qc8w-gemm_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -72777,8 +74891,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4-minmax-scalar.c"
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x2-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4-minmax-scalar.c"
 +  ]
  
 -  source_set("qd8-f16-qc4w-gemm_arch=armv8.2-a+i8mm+fp16") {
@@ -72837,7 +74952,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("qd8-f32-qc8w-gemm_ppc64") {
++source_set("qd8-f32-qc8w-igemm_ppc64") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -72860,10 +74975,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x2-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-2x2-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4-minmax-scalar.c"
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x2-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4-minmax-scalar.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -72914,7 +75028,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qc8w-gemm_ppc64_standalone") {
++source_set("qd8-f32-qc8w-igemm_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -72926,10 +75040,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x2-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-2x2-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4-minmax-scalar.c"
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x2-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4-minmax-scalar.c",
++    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4-minmax-scalar.c"
 +  ]
  
 -  source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+fp16") {
@@ -72977,9 +75090,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x8c2s4-minmax-neonfp16arith.c",
 -      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-2x8c2s4-minmax-neonfp16arith.c",
 -    ]
-+source_set("qd8-f32-qc8w-igemm_ppc64") {
-+  cflags = [
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -72990,91 +75101,44 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
- 
+-
 -    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x2-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-2x2-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4-minmax-scalar.c"
-+  ]
++source_set("qs8-dwconv_ppc64") {
++  cflags = [
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  ]
  
 -  source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+fp16+dotprod") {
 -    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qs8-dwconv/gen/qs8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
++    "src/src/qs8-dwconv/gen/qs8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -    asmflags = cflags
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S",
 -      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondotfp16arith-cortex-a55.S",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qd8-f32-qc8w-igemm_ppc64_standalone") {
-+  cflags = [
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x2-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-2x2-minmax-scalar.c",
-+    "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4-minmax-scalar.c"
-+  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
--    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    asmflags = cflags
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondotfp16arith-cortex-a55.S",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -73083,39 +75147,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
++  public_configs = [ ":xnnpack_config" ]
 +}
  
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("qs8-dwconv_ppc64") {
+-  }
++# This is a target that cannot depend on //base.
++source_set("qs8-dwconv_ppc64_standalone") {
 +  cflags = [
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
+-    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
 +  ]
  
--  source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+i8mm+fp16") {
--    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
+-    asmflags = cflags
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-dwconv/gen/qs8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
 +    "src/src/qs8-dwconv/gen/qs8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x16c8-minmax-neoni8mm.c",
--      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c8-minmax-neoni8mm.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondotfp16arith-cortex-a55.S",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -73129,81 +75189,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
 +  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
 +}
  
--  # This is a target that cannot depend on //base.
--  source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
+-  source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+i8mm+fp16") {
 -    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
-+# This is a target that cannot depend on //base.
-+source_set("qs8-dwconv_ppc64_standalone") {
-+  cflags = [
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x16c8-minmax-neoni8mm.c",
 -      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c8-minmax-neoni8mm.c",
 -    ]
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
--
--  source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+dotprod+fp16") {
--    cflags = [ "-march=armv8.2-a+dotprod+fp16" ]
 -
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x16c4-minmax-neondotfp16arith.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c4-minmax-neondotfp16arith.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-neondotfp16arith.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x8c4-minmax-neondotfp16arith.c",
--    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -73214,27 +75228,29 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++source_set("qs8-f32-vcvt_ppc64") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+dotprod+fp16_standalone") {
--    cflags = [ "-march=armv8.2-a+dotprod+fp16" ]
--
+-  source_set("qd8-f16-qc8w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
+-    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u4.c"
++  ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x16c4-minmax-neondotfp16arith.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c4-minmax-neondotfp16arith.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-neondotfp16arith.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x8c4-minmax-neondotfp16arith.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-1x16c8-minmax-neoni8mm.c",
+-      "src/src/qd8-f16-qc8w-gemm/gen/qd8-f16-qc8w-gemm-4x16c8-minmax-neoni8mm.c",
 -    ]
-+source_set("qs8-f32-vcvt_ppc64") {
-+  cflags = [
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -73246,25 +75262,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u1.c",
-+    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u4.c"
-+  ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--  source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+fp16") {
--    cflags = [ "-march=armv8.2-a+fp16" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -73272,13 +75269,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
+-
+-  source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+dotprod+fp16") {
+-    cflags = [ "-march=armv8.2-a+dotprod+fp16" ]
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c2s4-minmax-neonfp16arith-mlal.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-2x8c2s4-minmax-neonfp16arith-mlal.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x16c4-minmax-neondotfp16arith.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c4-minmax-neondotfp16arith.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-neondotfp16arith.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x8c4-minmax-neondotfp16arith.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++# This is a target that cannot depend on //base.
++source_set("qs8-f32-vcvt_ppc64_standalone") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -73290,27 +75302,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qs8-f32-vcvt_ppc64_standalone") {
-+  cflags = [
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+fp16_standalone") {
--    cflags = [ "-march=armv8.2-a+fp16" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u1.c",
 +    "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u4.c"
 +  ]
  
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+dotprod+fp16_standalone") {
+-    cflags = [ "-march=armv8.2-a+dotprod+fp16" ]
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c2s4-minmax-neonfp16arith-mlal.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-2x8c2s4-minmax-neonfp16arith-mlal.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x16c4-minmax-neondotfp16arith.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c4-minmax-neondotfp16arith.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-neondotfp16arith.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x8c4-minmax-neondotfp16arith.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -73344,15 +75354,13 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+fp16+dotprod") {
--    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
--
--    asmflags = cflags
+-  source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+fp16") {
+-    cflags = [ "-march=armv8.2-a+fp16" ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c2s4-minmax-neonfp16arith-mlal.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-2x8c2s4-minmax-neonfp16arith-mlal.c",
 -    ]
 +source_set("qs8-packw_ppc64") {
 +  cflags = [
@@ -73377,13 +75385,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+fp16+dotprod_standalone") {
--    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
+-  source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+fp16_standalone") {
+-    cflags = [ "-march=armv8.2-a+fp16" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    asmflags = cflags
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x8c2s4-minmax-neonfp16arith-mlal.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-2x8c2s4-minmax-neonfp16arith-mlal.c",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -73391,14 +75403,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -73409,32 +75413,36 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
 +source_set("qs8-packw_ppc64_standalone") {
 +  cflags = [
  
--    public_configs = [ ":xnnpack_config" ]
-+  ]
- 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
++  ]
+ 
+-  source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+fp16+dotprod") {
+-    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/qs8-packw/gen/qs8-packw-x16c8-gemm-goi-scalar.c"
 +  ]
  
--  source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+i8mm+fp16") {
--    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
+-    asmflags = cflags
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x16c8-minmax-neoni8mm.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c8-minmax-neoni8mm.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -73462,13 +75470,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+i8mm+fp16_standalone") {
--    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
+-  source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+fp16+dotprod_standalone") {
+-    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
+-
+-    asmflags = cflags
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x16c8-minmax-neoni8mm.c",
--      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c8-minmax-neoni8mm.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S",
 -    ]
 +source_set("qs8-qc4w-gemm_ppc64") {
 +  cflags = [
@@ -73499,8 +75509,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qd8-f32-qb4w-gemm_arch=armv8.2-a+dotprod") {
--    cflags = [ "-march=armv8.2-a+dotprod" ]
+-  source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+i8mm+fp16") {
+-    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -73510,8 +75520,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c4-minmax-neondot.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x16c4-minmax-neondot.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x16c8-minmax-neoni8mm.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c8-minmax-neoni8mm.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -73535,8 +75545,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qb4w-gemm_arch=armv8.2-a+dotprod_standalone") {
--    cflags = [ "-march=armv8.2-a+dotprod" ]
+-  source_set("qd8-f16-qc8w-igemm_arch=armv8.2-a+i8mm+fp16_standalone") {
+-    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-1x2-minmax-fp32-scalar-lrintf.c"
@@ -73544,8 +75554,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c4-minmax-neondot.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x16c4-minmax-neondot.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-1x16c8-minmax-neoni8mm.c",
+-      "src/src/qd8-f16-qc8w-igemm/gen/qd8-f16-qc8w-igemm-4x16c8-minmax-neoni8mm.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -73579,15 +75589,15 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qd8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16") {
--    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
+-  source_set("qd8-f32-qb4w-gemm_arch=armv8.2-a+dotprod") {
+-    cflags = [ "-march=armv8.2-a+dotprod" ]
 +source_set("qs8-qc8w-dwconv_ppc64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c8-minmax-neoni8mm.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x16c8-minmax-neoni8mm.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c4-minmax-neondot.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x16c4-minmax-neondot.c",
 -    ]
 +  ]
  
@@ -73604,13 +75614,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-lrintf.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
  
@@ -73621,8 +75628,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
--    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
+-  source_set("qd8-f32-qb4w-gemm_arch=armv8.2-a+dotprod_standalone") {
+-    cflags = [ "-march=armv8.2-a+dotprod" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -73632,8 +75639,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c8-minmax-neoni8mm.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x16c8-minmax-neoni8mm.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c4-minmax-neondot.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x16c4-minmax-neondot.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -73659,30 +75666,34 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set("qd8-f32-qb4w-gemm_arm64") {
--    cflags = []
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16-minmax-neon-mlal-lane.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4-minmax-scalar.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4-minmax-scalar.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-6x16-minmax-neon-mlal-lane.c",
--    ]
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-lrintf.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
  
+-  source_set("qd8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16") {
+-    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c8-minmax-neoni8mm.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x16c8-minmax-neoni8mm.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -73693,17 +75704,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
 +  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -73713,19 +75713,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qb4w-gemm_arm64_standalone") {
--    cflags = []
+-  source_set("qd8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
+-    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16-minmax-neon-mlal-lane.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4-minmax-scalar.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4-minmax-scalar.c",
--      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-6x16-minmax-neon-mlal-lane.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16c8-minmax-neoni8mm.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x16c8-minmax-neoni8mm.c",
 -    ]
-+source_set("qs8-qc8w-gemm_ppc64") {
-+  cflags = [
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -73736,27 +75732,49 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
++source_set("qs8-qc8w-gemm_ppc64") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
+-
+-  source_set("qd8-f32-qb4w-gemm_arm64") {
+-    cflags = []
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16-minmax-neon-mlal-lane.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4-minmax-scalar.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4-minmax-scalar.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-6x16-minmax-neon-mlal-lane.c",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4-minmax-fp32-scalar-lrintf.c"
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qd8-f32-qc4w-gemm_arch=armv8.2-a+dotprod") {
--    cflags = [ "-march=armv8.2-a+dotprod" ]
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -73764,10 +75782,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f32-qb4w-gemm_arm64_standalone") {
+-    cflags = []
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c4-minmax-neondot.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x16c4-minmax-neondot.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x16-minmax-neon-mlal-lane.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-1x4-minmax-scalar.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-4x4-minmax-scalar.c",
+-      "src/src/qd8-f32-qb4w-gemm/gen/qd8-f32-qb4w-gemm-6x16-minmax-neon-mlal-lane.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -73780,35 +75804,42 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
 +source_set("qs8-qc8w-gemm_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc4w-gemm_arch=armv8.2-a+dotprod_standalone") {
--    cflags = [ "-march=armv8.2-a+dotprod" ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x2-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x2-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
  
+-  source_set("qd8-f32-qc4w-gemm_arch=armv8.2-a+dotprod") {
+-    cflags = [ "-march=armv8.2-a+dotprod" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c4-minmax-neondot.c",
 -      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x16c4-minmax-neondot.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -73818,35 +75849,26 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
+-    public_configs = [ ":xnnpack_config" ]
 +  if (!(is_android && use_order_profiling)) {
 +    assert_no_deps = [ "//base" ]
    }
 +}
  
--  source_set("qd8-f32-qc4w-gemm_arch=armv8.2-a+i8mm+fp16") {
--    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f32-qc4w-gemm_arch=armv8.2-a+dotprod_standalone") {
+-    cflags = [ "-march=armv8.2-a+dotprod" ]
 +source_set("qs8-qc8w-igemm_ppc64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-neoni8mm.c",
--      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x16c8-minmax-neoni8mm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c4-minmax-neondot.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x16c4-minmax-neondot.c",
 -    ]
 +  ]
  
@@ -73858,25 +75880,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x2-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x2-minmax-fp32-scalar-imagic.c",
 +    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qd8-f32-qc4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
--    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -73884,13 +75904,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-  source_set("qd8-f32-qc4w-gemm_arch=armv8.2-a+i8mm+fp16") {
+-    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-neoni8mm.c",
 -      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x16c8-minmax-neoni8mm.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++# This is a target that cannot depend on //base.
++source_set("qs8-qc8w-igemm_ppc64_standalone") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -73900,15 +75926,55 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qs8-qc8w-igemm_ppc64_standalone") {
-+  cflags = [
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4-minmax-fp32-scalar-lrintf.c"
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qd8-f32-qc4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
+-    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-1x16c8-minmax-neoni8mm.c",
+-      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x16c8-minmax-neoni8mm.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
@@ -73924,13 +75990,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4-minmax-scalar.c",
 -      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-6x16-minmax-neon-mlal-lane.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4-minmax-fp32-scalar-lrintf.c"
-+  ]
++source_set("qs8-qu8-packw_ppc64") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -73942,17 +76003,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-scalar.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -73966,7 +76023,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-4x4-minmax-scalar.c",
 -      "src/src/qd8-f32-qc4w-gemm/gen/qd8-f32-qc4w-gemm-6x16-minmax-neon-mlal-lane.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -73978,14 +76037,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("qs8-qu8-packw_ppc64") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
@@ -74002,7 +76063,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-neondot.c",
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x8c4-minmax-neondot.c",
 -    ]
-+  ]
++# This is a target that cannot depend on //base.
++source_set("qs8-qu8-packw_ppc64_standalone") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -74014,16 +76077,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-scalar.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-scalar.c"
++  ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set("qd8-f32-qc8w-gemm_arch=armv8.2-a+dotprod_standalone") {
@@ -74036,12 +76097,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-neondot.c",
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x8c4-minmax-neondot.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -74053,43 +76111,36 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+# This is a target that cannot depend on //base.
-+source_set("qs8-qu8-packw_ppc64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  source_set("qd8-f32-qc8w-gemm_arch=armv8.2-a+fp16+dotprod") {
 -    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-scalar.c"
-+  ]
- 
+-
 -    asmflags = cflags
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++source_set("qs8-rdsum_ppc64") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -74101,21 +76152,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-rdsum/gen/qs8-rdsum-minmax-fp32-scalar-u1-acc1.c"
++  ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set("qd8-f32-qc8w-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
 -    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
-+source_set("qs8-rdsum_ppc64") {
-+  cflags = [
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    asmflags = cflags
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    sources = [
@@ -74123,10 +76181,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-rdsum/gen/qs8-rdsum-minmax-fp32-scalar-u1-acc1.c"
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -74138,41 +76194,38 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++# This is a target that cannot depend on //base.
++source_set("qs8-rdsum_ppc64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+# This is a target that cannot depend on //base.
-+source_set("qs8-rdsum_ppc64_standalone") {
-+  cflags = [
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-rdsum/gen/qs8-rdsum-minmax-fp32-scalar-u1-acc1.c"
++  ]
  
 -  source_set("qd8-f32-qc8w-gemm_arch=armv8.2-a+i8mm+fp16") {
 -    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x16c8-minmax-neoni8mm.c",
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x16c8-minmax-neoni8mm.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-rdsum/gen/qs8-rdsum-minmax-fp32-scalar-u1-acc1.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -74185,17 +76238,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
 +  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -74250,7 +76292,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x2-minmax-scalar.c",
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4-minmax-scalar.c",
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c2s4-minmax-neon-mlal.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-2x2-minmax-scalar.c",
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-2x8c2s4-minmax-neon-mlal.c",
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4-minmax-scalar.c",
 -    ]
@@ -74289,7 +76330,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x2-minmax-scalar.c",
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x4-minmax-scalar.c",
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-1x8c2s4-minmax-neon-mlal.c",
--      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-2x2-minmax-scalar.c",
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-2x8c2s4-minmax-neon-mlal.c",
 -      "src/src/qd8-f32-qc8w-gemm/gen/qd8-f32-qc8w-gemm-4x4-minmax-scalar.c",
 -    ]
@@ -74366,9 +76406,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-neondot.c",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x8c4-minmax-neondot.c",
 -    ]
-+source_set("qs8-vadd_ppc64") {
-+  cflags = [
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -74379,44 +76417,41 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
- 
+-
 -    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u1.c",
-+    "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u4.c"
-+  ]
++source_set("qs8-vadd_ppc64") {
++  cflags = [
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  ]
  
 -  source_set("qd8-f32-qc8w-igemm_arch=armv8.2-a+fp16+dotprod") {
 -    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u1.c",
++    "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u4.c"
 +  ]
  
 -    asmflags = cflags
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qs8-vadd_ppc64_standalone") {
-+  cflags = [
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -74428,29 +76463,25 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u1.c",
-+    "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u4.c"
-+  ]
++# This is a target that cannot depend on //base.
++source_set("qs8-vadd_ppc64_standalone") {
++  cflags = [
  
 -  # This is a target that cannot depend on //base.
 -  source_set("qd8-f32-qc8w-igemm_arch=armv8.2-a+fp16+dotprod_standalone") {
 -    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  ]
  
 -    asmflags = cflags
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u1.c",
++    "src/src/qs8-vadd/gen/qs8-vadd-minmax-scalar-u4.c"
 +  ]
  
 -    sources = [
@@ -74458,7 +76489,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-cortex-a55.S",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c4-minmax-asm-aarch64-neondot-ld128.S",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -74470,38 +76503,33 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("qs8-vaddc_ppc64") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  source_set("qd8-f32-qc8w-igemm_arch=armv8.2-a+i8mm+fp16") {
 -    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u1.c",
-+    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u4.c"
-+  ]
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c8-minmax-neoni8mm.c",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c8-minmax-neoni8mm.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -74512,31 +76540,30 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++source_set("qs8-vaddc_ppc64") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set("qd8-f32-qc8w-igemm_arch=armv8.2-a+i8mm+fp16_standalone") {
 -    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
-+# This is a target that cannot depend on //base.
-+source_set("qs8-vaddc_ppc64_standalone") {
-+  cflags = [
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u1.c",
++    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u4.c"
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x16c8-minmax-neoni8mm.c",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x16c8-minmax-neoni8mm.c",
 -    ]
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -74548,16 +76575,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u1.c",
-+    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u4.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
@@ -74573,17 +76600,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4-minmax-scalar.c",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8-minmax-neon-mlal-lane.c",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c2s4-minmax-neon-mlal.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-2x2-minmax-scalar.c",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-2x8c2s4-minmax-neon-mlal.c",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4-minmax-scalar.c",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x8-minmax-neon-mlal-lane.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++# This is a target that cannot depend on //base.
++source_set("qs8-vaddc_ppc64_standalone") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -74595,13 +76618,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u1.c",
++    "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u4.c"
++  ]
  
 -  # This is a target that cannot depend on //base.
 -  source_set("qd8-f32-qc8w-igemm_arm64_standalone") {
@@ -74613,90 +76638,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x4-minmax-scalar.c",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8-minmax-neon-mlal-lane.c",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-1x8c2s4-minmax-neon-mlal.c",
--      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-2x2-minmax-scalar.c",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-2x8c2s4-minmax-neon-mlal.c",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x4-minmax-scalar.c",
 -      "src/src/qd8-f32-qc8w-igemm/gen/qd8-f32-qc8w-igemm-4x8-minmax-neon-mlal-lane.c",
 -    ]
-+source_set("qs8-vcvt_ppc64") {
-+  cflags = [
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u1.c",
-+    "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c"
-+  ]
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--  source_set("qp8-f32-qb4w-gemm_arch=armv8.2-a+dotprod") {
--    cflags = [ "-march=armv8.2-a+dotprod" ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-1x4c16s2-aarch64-neondot.c",
--      "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-1x8c16s2-aarch64-neondot.c",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+# This is a target that cannot depend on //base.
-+source_set("qs8-vcvt_ppc64_standalone") {
-+  cflags = [
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
-+  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("qp8-f32-qb4w-gemm_arch=armv8.2-a+dotprod_standalone") {
--    cflags = [ "-march=armv8.2-a+dotprod" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u1.c",
-+    "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c"
-+  ]
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-1x4c16s2-aarch64-neondot.c",
--      "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-1x8c16s2-aarch64-neondot.c",
--    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -74729,14 +76674,15 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qp8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16") {
--    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
-+source_set("qs8-vlrelu_ppc64") {
+-  source_set("qp8-f32-qb4w-gemm_arch=armv8.2-a+dotprod") {
+-    cflags = [ "-march=armv8.2-a+dotprod" ]
++source_set("qs8-vcvt_ppc64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-16x4c16s2-mstep4-neoni8mm.c",
+-      "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-1x4c16s2-aarch64-neondot.c",
+-      "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-1x8c16s2-aarch64-neondot.c",
 -    ]
 +  ]
  
@@ -74752,8 +76698,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c",
-+    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-select-u4.c"
++    "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -74763,8 +76708,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qp8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
--    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
+-  source_set("qp8-f32-qb4w-gemm_arch=armv8.2-a+dotprod_standalone") {
+-    cflags = [ "-march=armv8.2-a+dotprod" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -74774,7 +76719,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-16x4c16s2-mstep4-neoni8mm.c",
+-      "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-1x4c16s2-aarch64-neondot.c",
+-      "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-1x8c16s2-aarch64-neondot.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -74790,7 +76736,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qs8-vlrelu_ppc64_standalone") {
++source_set("qs8-vcvt_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -74802,21 +76748,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c",
-+    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-select-u4.c"
++    "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c"
 +  ]
  
--  source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+dotprod") {
--    cflags = [ "-march=armv8.2-a+dotprod" ]
+-  source_set("qp8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16") {
+-    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-16x4c8s2-mstep4-aarch64-neondot.c",
--      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x4c8s2-aarch64-neondot.c",
--      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x8c16s2-aarch64-neondot.c",
+-      "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-16x4c16s2-mstep4-neoni8mm.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -74844,17 +76787,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+dotprod_standalone") {
--    cflags = [ "-march=armv8.2-a+dotprod" ]
--
+-  source_set("qp8-f32-qb4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
+-    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
++source_set("qs8-vlrelu_ppc64") {
++  cflags = [
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-16x4c8s2-mstep4-aarch64-neondot.c",
--      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x4c8s2-aarch64-neondot.c",
--      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x8c16s2-aarch64-neondot.c",
+-      "src/src/qp8-f32-qb4w-gemm/qp8-f32-qb4w-gemm-minmax-16x4c16s2-mstep4-neoni8mm.c",
 -    ]
-+source_set("qs8-vmul_ppc64") {
-+  cflags = [
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -74866,24 +76808,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-scalar-u4.c"
++    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c"
 +  ]
  
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
+-    public_configs = [ ":xnnpack_config" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+i8mm+fp16") {
--    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -74891,12 +76829,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-  source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+dotprod") {
+-    cflags = [ "-march=armv8.2-a+dotprod" ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-8x8c16s2-mstep2-neoni8mm.c",
+-      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-16x4c8s2-mstep4-aarch64-neondot.c",
+-      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x4c8s2-aarch64-neondot.c",
+-      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x8c16s2-aarch64-neondot.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++# This is a target that cannot depend on //base.
++source_set("qs8-vlrelu_ppc64_standalone") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -74908,25 +76854,24 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("qs8-vmul_ppc64_standalone") {
-+  cflags = [
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
--    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-scalar-u4.c"
++    "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c"
 +  ]
  
+-  # This is a target that cannot depend on //base.
+-  source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+dotprod_standalone") {
+-    cflags = [ "-march=armv8.2-a+dotprod" ]
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-8x8c16s2-mstep2-neoni8mm.c",
+-      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-16x4c8s2-mstep4-aarch64-neondot.c",
+-      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x4c8s2-aarch64-neondot.c",
+-      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x8c16s2-aarch64-neondot.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -74960,16 +76905,14 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qp8-f32-qc8w-gemm_arch=armv8.2-a+dotprod") {
--    cflags = [ "-march=armv8.2-a+dotprod" ]
-+source_set("qs8-vmulc_ppc64") {
+-  source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+i8mm+fp16") {
+-    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
++source_set("qs8-vmul_ppc64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-16x4c4-mstep4-aarch64-neondot.c",
--      "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-1x4c4-aarch64-neondot.c",
--      "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-1x4c8-aarch64-neondot.c",
+-      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-8x8c16s2-mstep2-neoni8mm.c",
 -    ]
 +  ]
  
@@ -74985,7 +76928,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-scalar-u4.c"
++    "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -74995,15 +76938,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qp8-f32-qc8w-gemm_arch=armv8.2-a+dotprod_standalone") {
--    cflags = [ "-march=armv8.2-a+dotprod" ]
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-16x4c4-mstep4-aarch64-neondot.c",
--      "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-1x4c4-aarch64-neondot.c",
--      "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-1x4c8-aarch64-neondot.c",
--    ]
+-  source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
+-    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -75011,6 +76947,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-8x8c16s2-mstep2-neoni8mm.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -75021,11 +76964,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
-+
 +# This is a target that cannot depend on //base.
-+source_set("qs8-vmulc_ppc64_standalone") {
++source_set("qs8-vmul_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -75037,18 +76977,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-scalar-u4.c"
++    "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-scalar-u4.c"
 +  ]
  
--  source_set("qp8-f32-qc8w-gemm_arch=armv8.2-a+i8mm+fp16") {
--    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
+-  source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+sve+sve2") {
+-    cflags = [ "-march=armv8.2-a+sve+sve2" ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-16x4c8-mstep4-neoni8mm.c",
+-      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x128c4-neonsme2.c",
+-      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-32x128c4-neonsme2.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -75076,14 +77017,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qp8-f32-qc8w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
--    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
-+source_set("qu8-dwconv_ppc64") {
+-  source_set("qp8-f32-qc4w-gemm_arch=armv8.2-a+sve+sve2_standalone") {
+-    cflags = [ "-march=armv8.2-a+sve+sve2" ]
++source_set("qs8-vmulc_ppc64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-16x4c8-mstep4-neoni8mm.c",
+-      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-1x128c4-neonsme2.c",
+-      "src/src/qp8-f32-qc4w-gemm/qp8-f32-qc4w-gemm-minmax-32x128c4-neonsme2.c",
 -    ]
 +  ]
  
@@ -75099,12 +77041,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
++    "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -75116,29 +77053,97 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-  source_set("qp8-f32-qc8w-gemm_arch=armv8.2-a+dotprod") {
+-    cflags = [ "-march=armv8.2-a+dotprod" ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-16x4c4-mstep4-aarch64-neondot.c",
+-      "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-1x4c4-aarch64-neondot.c",
+-      "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-1x4c8-aarch64-neondot.c",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("qs8-vmulc_ppc64_standalone") {
++  cflags = [
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
 -
--  source_set("qs8-dwconv_arm64") {
--    cflags = []
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-scalar-u4.c"
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qp8-f32-qc8w-gemm_arch=armv8.2-a+dotprod_standalone") {
+-    cflags = [ "-march=armv8.2-a+dotprod" ]
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p16c-minmax-rndnu-neon-mla8-ld64.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p8c-minmax-rndnu-neon-mla8-ld64.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p16c-minmax-rndnu-neon-mla8-ld64.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-16x4c4-mstep4-aarch64-neondot.c",
+-      "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-1x4c4-aarch64-neondot.c",
+-      "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-1x4c8-aarch64-neondot.c",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  source_set("qp8-f32-qc8w-gemm_arch=armv8.2-a+i8mm+fp16") {
+-    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-16x4c8-mstep4-neoni8mm.c",
+-    ]
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -75149,32 +77154,71 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++source_set("qs8-vprelu_ppc64") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+# This is a target that cannot depend on //base.
-+source_set("qu8-dwconv_ppc64_standalone") {
-+  cflags = [
++  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-dwconv_arm64_standalone") {
+-  source_set("qp8-f32-qc8w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
+-    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qs8-vprelu/gen/qs8-vprelu-scalar-u8.c"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qp8-f32-qc8w-gemm/qp8-f32-qc8w-gemm-minmax-16x4c8-mstep4-neoni8mm.c",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
+-
+-  source_set("qs8-dwconv_arm64") {
 -    cflags = []
 -
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qs8-dwconv/gen/qs8-dwconv-25p16c-minmax-rndnu-neon-mla8-ld64.c",
 -      "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
 -      "src/src/qs8-dwconv/gen/qs8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qs8-dwconv/gen/qs8-dwconv-25p8c-minmax-rndnu-neon-mla8-ld64.c",
 -      "src/src/qs8-dwconv/gen/qs8-dwconv-9p16c-minmax-rndnu-neon-mla8-ld64.c",
 -      "src/src/qs8-dwconv/gen/qs8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
 -      "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
 -    ]
-+  ]
++# This is a target that cannot depend on //base.
++source_set("qs8-vprelu_ppc64_standalone") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -75184,30 +77228,53 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
++    "src/src/qs8-vprelu/gen/qs8-vprelu-scalar-u8.c"
 +  ]
-+
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-dwconv_arm64_standalone") {
+-    cflags = []
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p16c-minmax-rndnu-neon-mla8-ld64.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-25p8c-minmax-rndnu-neon-mla8-ld64.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p16c-minmax-rndnu-neon-mla8-ld64.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
+-      "src/src/qs8-dwconv/gen/qs8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -75220,7 +77287,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("qs8-f16-vcvt_arch=armv8.2-a+fp16") {
 -    cflags = [ "-march=armv8.2-a+fp16" ]
-+source_set("qu8-f32-vcvt_ppc64") {
++source_set("qs8-vpreluc_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -75241,8 +77308,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u1.c",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c"
++    "src/src/qs8-vpreluc/gen/qs8-vpreluc-scalar-u8.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -75279,7 +77345,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-f32-vcvt_ppc64_standalone") {
++source_set("qs8-vpreluc_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -75291,8 +77357,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u1.c",
-+    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c"
++    "src/src/qs8-vpreluc/gen/qs8-vpreluc-scalar-u8.c"
 +  ]
  
 -  source_set("qs8-f32-vcvt_arm64") {
@@ -75304,7 +77369,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-neon-u32.c",
--      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u1.c",
 -      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u4.c",
 -    ]
 +  deps = [
@@ -75339,10 +77403,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-neon-u32.c",
--      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u1.c",
 -      "src/src/qs8-f32-vcvt/gen/qs8-f32-vcvt-scalar-u4.c",
 -    ]
-+source_set("qu8-gemm_ppc64") {
++source_set("qs8-vrpreluc_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -75360,10 +77423,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c"
++    "src/src/qs8-vrpreluc/gen/qs8-vrpreluc-scalar-u8.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -75401,7 +77461,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-gemm_ppc64_standalone") {
++source_set("qs8-vrpreluc_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -75413,10 +77473,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c"
++    "src/src/qs8-vrpreluc/gen/qs8-vrpreluc-scalar-u8.c"
 +  ]
  
 -    sources = [
@@ -75465,7 +77522,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-1x16-minmax-fp32-asm-aarch64-neondot-ld128.S",
 -      "src/src/qs8-qc4w-gemm/gen/qs8-qc4w-gemm-5x16-minmax-fp32-asm-aarch64-neondot-ld128.S",
 -    ]
-+source_set("qu8-igemm_ppc64") {
++source_set("qu8-dwconv_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -75484,10 +77541,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c"
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -75524,7 +77581,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-igemm_ppc64_standalone") {
++source_set("qu8-dwconv_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -75536,10 +77593,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-2x2-minmax-fp32-scalar-imagic.c",
-+    "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c"
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
++    "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -  source_set("qs8-qc4w-gemm_arm64") {
@@ -75580,7 +77637,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  # This is a target that cannot depend on //base.
 -  source_set("qs8-qc4w-gemm_arm64_standalone") {
 -    cflags = []
-+source_set("qu8-rdsum_ppc64") {
++source_set("qu8-f32-vcvt_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -75601,7 +77658,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rdsum/gen/qu8-rdsum-scalar.c"
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -75621,7 +77678,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p16c-minmax-fp32-neonv8-mla8-ld64.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-neon-mla8-ld64.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-neonv8-mla8-ld64.c",
@@ -75629,12 +77685,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-neonv8-mla8-ld128.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-neonv8-mla8-ld64.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-imagic.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-neon-mla8-ld64.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-neonv8-mla8-ld64.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
 -    ]
 +  deps = [
@@ -75660,7 +77714,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("qu8-rdsum_ppc64_standalone") {
++source_set("qu8-f32-vcvt_ppc64_standalone") {
 +  cflags = [
  
 -  # This is a target that cannot depend on //base.
@@ -75671,7 +77725,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p16c-minmax-fp32-neonv8-mla8-ld64.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-neon-mla8-ld64.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-25p8c-minmax-fp32-neonv8-mla8-ld64.c",
@@ -75679,12 +77732,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-neonv8-mla8-ld128.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p16c-minmax-fp32-neonv8-mla8-ld64.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-imagic.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-3p2c-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-neon-mla8-ld64.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p16c-minmax-fp32-neonv8-mla8-ld64.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
 -      "src/src/qs8-qc8w-dwconv/gen/qs8-qc8w-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
 -    ]
 +  ]
@@ -75701,7 +77752,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rdsum/gen/qu8-rdsum-scalar.c"
++    "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -75771,7 +77822,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("qu8-rsum_ppc64") {
++source_set("qu8-gemm_ppc64") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -75802,7 +77853,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rsum/gen/qu8-rsum-scalar-u4.c"
++    "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -75861,7 +77913,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-rsum_ppc64_standalone") {
++source_set("qu8-gemm_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -75873,7 +77925,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-rsum/gen/qu8-rsum-scalar-u4.c"
++    "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -  source_set("qs8-qc8w-gemm_arch=armv8.2-a+i8mm+fp16") {
@@ -75915,7 +77968,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  # This is a target that cannot depend on //base.
 -  source_set("qs8-qc8w-gemm_arch=armv8.2-a+i8mm+fp16_standalone") {
 -    cflags = [ "-march=armv8.2-a+i8mm+fp16" ]
-+source_set("qu8-vadd_ppc64") {
++source_set("qu8-igemm_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -75937,8 +77990,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u1.c",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u4.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -75957,11 +78010,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16-minmax-fp32-neonv8-mlal-lane.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x2-minmax-fp32-scalar-imagic.c",
 -      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c2s4-minmax-fp32-neon-mlal.c",
 -      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c2s4-minmax-fp32-neonv8-mlal.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x2-minmax-fp32-scalar-imagic.c",
 -      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c2s4-minmax-fp32-neon-mlal.c",
 -      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c2s4-minmax-fp32-neonv8-mlal.c",
 -      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4-minmax-fp32-scalar-lrintf.c",
@@ -75989,7 +78040,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vadd_ppc64_standalone") {
++source_set("qu8-igemm_ppc64_standalone") {
 +  cflags = [
  
 -  # This is a target that cannot depend on //base.
@@ -75999,11 +78050,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x16-minmax-fp32-neonv8-mlal-lane.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x2-minmax-fp32-scalar-imagic.c",
 -      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x4-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c2s4-minmax-fp32-neon-mlal.c",
 -      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-1x8c2s4-minmax-fp32-neonv8-mlal.c",
--      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x2-minmax-fp32-scalar-imagic.c",
 -      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c2s4-minmax-fp32-neon-mlal.c",
 -      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-2x8c2s4-minmax-fp32-neonv8-mlal.c",
 -      "src/src/qs8-qc8w-gemm/gen/qs8-qc8w-gemm-3x4-minmax-fp32-scalar-lrintf.c",
@@ -76022,8 +78071,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u1.c",
-+    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u4.c"
++    "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
++    "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -76093,7 +78142,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("qu8-vaddc_ppc64") {
++source_set("qu8-rdsum_ppc64") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -76124,8 +78173,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c"
++    "src/src/qu8-rdsum/gen/qu8-rdsum-scalar.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -76184,7 +78232,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vaddc_ppc64_standalone") {
++source_set("qu8-rdsum_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -76196,8 +78244,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
-+    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c"
++    "src/src/qu8-rdsum/gen/qu8-rdsum-scalar.c"
 +  ]
  
 -  source_set("qs8-qc8w-igemm_arch=armv8.2-a+i8mm+fp16") {
@@ -76256,7 +78303,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+source_set("qu8-vcvt_ppc64") {
++source_set("qu8-rsum_ppc64") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -76273,19 +78320,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16-minmax-fp32-neonv8-mlal-lane.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x2-minmax-fp32-scalar-imagic.c",
 -      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c2s4-minmax-fp32-neon-mlal.c",
 -      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c2s4-minmax-fp32-neonv8-mlal.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x2-minmax-fp32-scalar-imagic.c",
 -      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c2s4-minmax-fp32-neon-mlal.c",
 -      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c2s4-minmax-fp32-neonv8-mlal.c",
 -      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4-minmax-fp32-scalar-lrintf.c",
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u1.c",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c"
++    "src/src/qu8-rsum/gen/qu8-rsum-scalar-u4.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -76318,11 +78362,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x16-minmax-fp32-neonv8-mlal-lane.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x2-minmax-fp32-scalar-imagic.c",
 -      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x4-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c2s4-minmax-fp32-neon-mlal.c",
 -      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-1x8c2s4-minmax-fp32-neonv8-mlal.c",
--      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x2-minmax-fp32-scalar-imagic.c",
 -      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c2s4-minmax-fp32-neon-mlal.c",
 -      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-2x8c2s4-minmax-fp32-neonv8-mlal.c",
 -      "src/src/qs8-qc8w-igemm/gen/qs8-qc8w-igemm-3x4-minmax-fp32-scalar-lrintf.c",
@@ -76341,7 +78383,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vcvt_ppc64_standalone") {
++source_set("qu8-rsum_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -76353,8 +78395,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u1.c",
-+    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c"
++    "src/src/qu8-rsum/gen/qu8-rsum-scalar-u4.c"
 +  ]
  
 -  source_set("qs8-qu8-packw_arm64") {
@@ -76400,7 +78441,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/qs8-qu8-packw/gen/qs8-qu8-packw-x16c8-gemm-goi-scalar.c",
 -    ]
-+source_set("qu8-vlrelu_ppc64") {
++source_set("qu8-vadd_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -76418,8 +78459,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-select-u4.c"
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u1.c",
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u4.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -76458,7 +78499,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vlrelu_ppc64_standalone") {
++source_set("qu8-vadd_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -76470,8 +78511,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c",
-+    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-select-u4.c"
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u1.c",
++    "src/src/qu8-vadd/gen/qu8-vadd-minmax-scalar-u4.c"
 +  ]
  
 -    sources = [
@@ -76513,7 +78554,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("qs8-rsum_arch=armv8.2-a+dotprod") {
 -    cflags = [ "-march=armv8.2-a+dotprod" ]
-+source_set("qu8-vmul_ppc64") {
++source_set("qu8-vaddc_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -76534,7 +78575,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c"
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -76571,7 +78613,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vmul_ppc64_standalone") {
++source_set("qu8-vaddc_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -76583,7 +78625,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c"
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
++    "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c"
 +  ]
  
 -  source_set("qs8-rsum_arm64") {
@@ -76625,7 +78668,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  # This is a target that cannot depend on //base.
 -  source_set("qs8-rsum_arm64_standalone") {
 -    cflags = []
-+source_set("qu8-vmulc_ppc64") {
++source_set("qu8-vcvt_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -76647,7 +78690,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c"
++    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -76693,7 +78736,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("qu8-vmulc_ppc64_standalone") {
++source_set("qu8-vcvt_ppc64_standalone") {
 +  cflags = [
  
 -  # This is a target that cannot depend on //base.
@@ -76721,7 +78764,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c"
++    "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -76780,7 +78823,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u1.c",
 -      "src/src/qs8-vaddc/gen/qs8-vaddc-minmax-scalar-u4.c",
 -    ]
--
++source_set("qu8-vlrelu_ppc64") {
++  cflags = [
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -76791,31 +78836,112 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
--
++  ]
+ 
 -    public_configs = [ ":xnnpack_config" ]
--
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c"
++  ]
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+source_set("reference_ppc64") {
-+  cflags = [
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  source_set("qs8-vcvt_arm64") {
 -    cflags = []
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qs8-vcvt/gen/qs8-vcvt-neon-u32.c",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u1.c",
 -      "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c",
 -    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("qu8-vlrelu_ppc64_standalone") {
++  cflags = [
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-vcvt_arm64_standalone") {
+-    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/reference/binary-elementwise.cc",
-+    "src/src/reference/packing.cc",
-+    "src/src/reference/unary-elementwise.cc"
++    "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vcvt/gen/qs8-vcvt-neon-u32.c",
+-      "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  source_set("qs8-vlrelu_arm64") {
+-    cflags = []
++source_set("qu8-vmul_ppc64") {
++  cflags = [
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vlrelu/gen/qs8-vlrelu-neon-u32.c",
+-      "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c",
+-    ]
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -76828,12 +78954,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-vlrelu_arm64_standalone") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -76841,15 +78975,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qs8-vcvt_arm64_standalone") {
--    cflags = []
--
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-neon-u32.c",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u1.c",
--      "src/src/qs8-vcvt/gen/qs8-vcvt-scalar-u4.c",
+-      "src/src/qs8-vlrelu/gen/qs8-vlrelu-neon-u32.c",
+-      "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -76865,7 +78994,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("reference_ppc64_standalone") {
++source_set("qu8-vmul_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -76877,12 +79006,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/reference/binary-elementwise.cc",
-+    "src/src/reference/packing.cc",
-+    "src/src/reference/unary-elementwise.cc"
++    "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c"
 +  ]
  
--  source_set("qs8-vlrelu_arm64") {
+-  source_set("qs8-vmul_arm64") {
 -    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -76890,9 +79017,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-neon-u32.c",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-select-u4.c",
+-      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-scalar-u4.c",
+-      "src/src/qs8-vmul/gen/qs8-vmul-minmax-rndnu-neon-ld64-u16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -76920,16 +79046,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vlrelu_arm64_standalone") {
+-  source_set("qs8-vmul_arm64_standalone") {
 -    cflags = []
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-neon-u32.c",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-andxor-u4.c",
--      "src/src/qs8-vlrelu/gen/qs8-vlrelu-scalar-select-u4.c",
+-      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-scalar-u4.c",
+-      "src/src/qs8-vmul/gen/qs8-vmul-minmax-rndnu-neon-ld64-u16.c",
 -    ]
-+source_set("s8-ibilinear_ppc64") {
++source_set("qu8-vmulc_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -76947,7 +79072,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-ibilinear/gen/s8-ibilinear-scalar-c1.c"
++    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -76958,7 +79083,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qs8-vmul_arm64") {
+-  source_set("qs8-vmulc_arm64") {
 -    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -76969,8 +79094,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-scalar-u4.c",
--      "src/src/qs8-vmul/gen/qs8-vmul-minmax-rndnu-neon-ld64-u16.c",
+-      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-scalar-u4.c",
+-      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-rndnu-neon-ld64-u16.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -76986,7 +79111,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("s8-ibilinear_ppc64_standalone") {
++source_set("qu8-vmulc_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -76994,17 +79119,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vmul_arm64_standalone") {
+-  source_set("qs8-vmulc_arm64_standalone") {
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-ibilinear/gen/s8-ibilinear-scalar-c1.c"
++    "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmul/gen/qs8-vmul-minmax-fp32-scalar-u4.c",
--      "src/src/qs8-vmul/gen/qs8-vmul-minmax-rndnu-neon-ld64-u16.c",
+-      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-scalar-u4.c",
+-      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-rndnu-neon-ld64-u16.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -77038,15 +79163,17 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qs8-vmulc_arm64") {
+-  source_set("qs8-vprelu_arm64") {
 -    cflags = []
--
++source_set("qu8-vprelu_ppc64") {
++  cflags = [
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-scalar-u4.c",
--      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-rndnu-neon-ld64-u16.c",
+-      "src/src/qs8-vprelu/gen/qs8-vprelu-scalar-u8.c",
 -    ]
--
++  ]
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -77057,25 +79184,135 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+source_set("s8-maxpool_ppc64") {
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qu8-vprelu/gen/qu8-vprelu-scalar-u8.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-vprelu_arm64_standalone") {
+-    cflags = []
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vprelu/gen/qs8-vprelu-scalar-u8.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("qu8-vprelu_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 -  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qu8-vprelu/gen/qu8-vprelu-scalar-u8.c"
++  ]
+ 
+-  source_set("qs8-vpreluc_arm64") {
+-    cflags = []
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vpreluc/gen/qs8-vpreluc-scalar-u8.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
 -  # This is a target that cannot depend on //base.
--  source_set("qs8-vmulc_arm64_standalone") {
+-  source_set("qs8-vpreluc_arm64_standalone") {
+-    cflags = []
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vpreluc/gen/qs8-vpreluc-scalar-u8.c",
+-    ]
+-
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
+-
+-    public_configs = [ ":xnnpack_config" ]
++source_set("qu8-vpreluc_ppc64") {
++  cflags = [
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  ]
+ 
+-  source_set("qs8-vrpreluc_arm64") {
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-scalar-u1.c"
++    "src/src/qu8-vpreluc/gen/qu8-vpreluc-scalar-u8.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-fp32-scalar-u4.c",
--      "src/src/qs8-vmulc/gen/qs8-vmulc-minmax-rndnu-neon-ld64-u16.c",
+-      "src/src/qs8-vrpreluc/gen/qs8-vrpreluc-scalar-u8.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -77089,7 +79326,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/cpuinfo",
 -      "//third_party/fp16",
 -      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
+-      "//third_party/pthreadpool",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -77099,9 +79336,43 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
+-  }
 +  public_configs = [ ":xnnpack_config" ]
 +}
  
+-  # This is a target that cannot depend on //base.
+-  source_set("qs8-vrpreluc_arm64_standalone") {
+-    cflags = []
++# This is a target that cannot depend on //base.
++source_set("qu8-vpreluc_ppc64_standalone") {
++  cflags = [
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qs8-vrpreluc/gen/qs8-vrpreluc-scalar-u8.c",
+-    ]
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qu8-vpreluc/gen/qu8-vpreluc-scalar-u8.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
@@ -77113,17 +79384,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
 -      "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-rndnu-neon-mul8.c",
 -      "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-rndnu-neon-mul8.c",
 -      "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
 -      "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("s8-maxpool_ppc64_standalone") {
-+  cflags = [
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -77135,14 +79407,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-scalar-u1.c"
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
 -  source_set("qu8-dwconv_arm64_standalone") {
@@ -77151,14 +79422,89 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-25p1c-minmax-fp32-scalar-imagic.c",
 -      "src/src/qu8-dwconv/gen/qu8-dwconv-25p2c-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qu8-dwconv/gen/qu8-dwconv-25p8c-minmax-rndnu-neon-mul8.c",
 -      "src/src/qu8-dwconv/gen/qu8-dwconv-9p16c-minmax-rndnu-neon-mul8.c",
 -      "src/src/qu8-dwconv/gen/qu8-dwconv-9p1c-minmax-fp32-scalar-fmagic.c",
--      "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-imagic.c",
 -      "src/src/qu8-dwconv/gen/qu8-dwconv-9p2c-minmax-fp32-scalar-lrintf.c",
 -    ]
++source_set("qu8-vrpreluc_ppc64") {
++  cflags = [
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qu8-vrpreluc/gen/qu8-vrpreluc-scalar-u8.c"
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  source_set("qu8-f32-vcvt_arm64") {
+-    cflags = []
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-neon-u32.c",
+-      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("qu8-vrpreluc_ppc64_standalone") {
++  cflags = [
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qu8-f32-vcvt_arm64_standalone") {
+-    cflags = []
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/qu8-vrpreluc/gen/qu8-vrpreluc-scalar-u8.c"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-neon-u32.c",
+-      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -77191,95 +79537,11 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-f32-vcvt_arm64") {
--    cflags = []
-+source_set("s8-rdminmax_ppc64") {
-+  cflags = [
- 
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-neon-u32.c",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u1.c",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c",
--    ]
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool",
--    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-scalar-c2.c",
-+    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-scalar-c2.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--  # This is a target that cannot depend on //base.
--  source_set("qu8-f32-vcvt_arm64_standalone") {
--    cflags = []
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-neon-u32.c",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u1.c",
--      "src/src/qu8-f32-vcvt/gen/qu8-f32-vcvt-scalar-u4.c",
--    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
- 
--    configs -= [ "//build/config/compiler:chromium_code" ]
--    configs += [ "//build/config/compiler:no_chromium_code" ]
--    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
--
--    deps = [
--      "//third_party/cpuinfo",
--      "//third_party/fp16",
--      "//third_party/fxdiv",
--      "//third_party/pthreadpool:pthreadpool_standalone",
--    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
-+# This is a target that cannot depend on //base.
-+source_set("s8-rdminmax_ppc64_standalone") {
-+  cflags = [
- 
--    if (!(is_android && use_order_profiling)) {
--      assert_no_deps = [ "//base" ]
--    }
--  }
-+  ]
- 
 -  source_set("qu8-gemm_arch=armv8.2-a+fp16+dotprod") {
 -    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-scalar-c2.c",
-+    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-scalar-c2.c"
-+  ]
- 
+-
 -    asmflags = cflags
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S",
@@ -77287,13 +79549,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75.S",
 -      "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu16-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -77304,21 +79560,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
+-
 -    public_configs = [ ":xnnpack_config" ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++source_set("reference_ppc64") {
++  cflags = [
  
 -  # This is a target that cannot depend on //base.
 -  source_set("qu8-gemm_arch=armv8.2-a+fp16+dotprod_standalone") {
 -    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
-+source_set("s8-rminmax_ppc64") {
-+  cflags = [
++  ]
  
 -    asmflags = cflags
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/reference/binary-elementwise.cc",
++    "src/src/reference/packing.cc",
++    "src/src/reference/unary-elementwise.cc"
 +  ]
  
 -    sources = [
@@ -77328,12 +79586,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-asm-aarch64-neon-mlal-lane-cortex-a75.S",
 -      "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu16-asm-aarch64-neon-mlal-lane-cortex-a53-prfm.S",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/s8-rminmax/gen/s8-rmax-scalar-u2-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rmin-scalar-u2-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rminmax-scalar-u2-acc2.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -77345,11 +79600,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -77357,6 +79607,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
@@ -77369,16 +79623,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/qu8-gemm/gen/qu8-gemm-1x16-minmax-rndnu-neon-mlal-lane.c",
 -      "src/src/qu8-gemm/gen/qu8-gemm-1x16-minmax-rndnu16-neon-mlal-lane.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-1x2-minmax-fp32-scalar-imagic.c",
 -      "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qu8-gemm/gen/qu8-gemm-1x8-minmax-rndnu-neon-mlal-lane.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-2x2-minmax-fp32-scalar-imagic.c",
 -      "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qu8-gemm/gen/qu8-gemm-3x8-minmax-rndnu-neon-mlal-lane.c",
 -      "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-neon-mlal-lane.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++# This is a target that cannot depend on //base.
++source_set("reference_ppc64_standalone") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -77390,12 +79643,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("s8-rminmax_ppc64_standalone") {
-+  cflags = [
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/reference/binary-elementwise.cc",
++    "src/src/reference/packing.cc",
++    "src/src/reference/unary-elementwise.cc"
 +  ]
  
 -  # This is a target that cannot depend on //base.
@@ -77406,20 +79662,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/qu8-gemm/gen/qu8-gemm-1x16-minmax-rndnu-neon-mlal-lane.c",
 -      "src/src/qu8-gemm/gen/qu8-gemm-1x16-minmax-rndnu16-neon-mlal-lane.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-1x2-minmax-fp32-scalar-imagic.c",
 -      "src/src/qu8-gemm/gen/qu8-gemm-1x4-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qu8-gemm/gen/qu8-gemm-1x8-minmax-rndnu-neon-mlal-lane.c",
--      "src/src/qu8-gemm/gen/qu8-gemm-2x2-minmax-fp32-scalar-imagic.c",
 -      "src/src/qu8-gemm/gen/qu8-gemm-3x4-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qu8-gemm/gen/qu8-gemm-3x8-minmax-rndnu-neon-mlal-lane.c",
 -      "src/src/qu8-gemm/gen/qu8-gemm-4x16-minmax-rndnu-neon-mlal-lane.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/s8-rminmax/gen/s8-rmax-scalar-u2-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rmin-scalar-u2-acc2.c",
-+    "src/src/s8-rminmax/gen/s8-rminmax-scalar-u2-acc2.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -77431,10 +79682,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
-+
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -77455,7 +79702,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("qu8-igemm_arch=armv8.2-a+fp16+dotprod") {
 -    cflags = [ "-march=armv8.2-a+fp16+dotprod" ]
-+source_set("s8-vclamp_ppc64") {
++source_set("s8-ibilinear_ppc64") {
 +  cflags = [
  
 -    asmflags = cflags
@@ -77470,7 +79717,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-vclamp/s8-vclamp-scalar-u4.c"
++    "src/src/s8-ibilinear/gen/s8-ibilinear-scalar-c1.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -77504,7 +79751,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    asmflags = cflags
 +# This is a target that cannot depend on //base.
-+source_set("s8-vclamp_ppc64_standalone") {
++source_set("s8-ibilinear_ppc64_standalone") {
 +  cflags = [
  
 -    sources = [
@@ -77528,7 +79775,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/s8-vclamp/s8-vclamp-scalar-u4.c"
++    "src/src/s8-ibilinear/gen/s8-ibilinear-scalar-c1.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -77548,10 +79795,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/qu8-igemm/gen/qu8-igemm-1x16-minmax-rndnu-neon-mlal-lane.c",
 -      "src/src/qu8-igemm/gen/qu8-igemm-1x16-minmax-rndnu16-neon-mlal-lane.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-1x2-minmax-fp32-scalar-imagic.c",
 -      "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qu8-igemm/gen/qu8-igemm-1x8-minmax-rndnu-neon-mlal-lane.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-2x2-minmax-fp32-scalar-imagic.c",
 -      "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qu8-igemm/gen/qu8-igemm-3x8-minmax-rndnu-neon-mlal-lane.c",
 -      "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-neon-mlal-lane.c",
@@ -77589,10 +79834,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/include/xnnpack.h",
 -      "src/src/qu8-igemm/gen/qu8-igemm-1x16-minmax-rndnu-neon-mlal-lane.c",
 -      "src/src/qu8-igemm/gen/qu8-igemm-1x16-minmax-rndnu16-neon-mlal-lane.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-1x2-minmax-fp32-scalar-imagic.c",
 -      "src/src/qu8-igemm/gen/qu8-igemm-1x4-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qu8-igemm/gen/qu8-igemm-1x8-minmax-rndnu-neon-mlal-lane.c",
--      "src/src/qu8-igemm/gen/qu8-igemm-2x2-minmax-fp32-scalar-imagic.c",
 -      "src/src/qu8-igemm/gen/qu8-igemm-3x4-minmax-fp32-scalar-lrintf.c",
 -      "src/src/qu8-igemm/gen/qu8-igemm-3x8-minmax-rndnu-neon-mlal-lane.c",
 -      "src/src/qu8-igemm/gen/qu8-igemm-4x16-minmax-rndnu-neon-mlal-lane.c",
@@ -77615,7 +79858,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+source_set("subgraph_ppc64") {
++source_set("s8-maxpool_ppc64") {
 +  cflags = [
  
 -  source_set("qu8-rdsum_arm64") {
@@ -77630,34 +79873,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/subgraph/argmax-pooling-2d.c",
-+    "src/src/subgraph/average-pooling-2d.c",
-+    "src/src/subgraph/batch-matrix-multiply.c",
-+    "src/src/subgraph/binary.c",
-+    "src/src/subgraph/concatenate.c",
-+    "src/src/subgraph/convolution-2d.c",
-+    "src/src/subgraph/copy.c",
-+    "src/src/subgraph/deconvolution-2d.c",
-+    "src/src/subgraph/deprecated.c",
-+    "src/src/subgraph/depth-to-space-2d.c",
-+    "src/src/subgraph/depthwise-convolution-2d.c",
-+    "src/src/subgraph/even-split.c",
-+    "src/src/subgraph/fully-connected-sparse.c",
-+    "src/src/subgraph/fully-connected.c",
-+    "src/src/subgraph/max-pooling-2d.c",
-+    "src/src/subgraph/pack-lh.c",
-+    "src/src/subgraph/reshape-helpers.c",
-+    "src/src/subgraph/rope.c",
-+    "src/src/subgraph/softmax.c",
-+    "src/src/subgraph/space-to-depth-2d.c",
-+    "src/src/subgraph/static-constant-pad.c",
-+    "src/src/subgraph/static-reduce.c",
-+    "src/src/subgraph/static-resize-bilinear-2d.c",
-+    "src/src/subgraph/static-slice.c",
-+    "src/src/subgraph/static-transpose.c",
-+    "src/src/subgraph/unary.c",
-+    "src/src/subgraph/unpooling-2d.c",
-+    "src/src/subgraph/validation.c"
++    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-scalar-u1.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -77707,7 +79923,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("subgraph_ppc64_standalone") {
++source_set("s8-maxpool_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -77719,34 +79935,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/subgraph/argmax-pooling-2d.c",
-+    "src/src/subgraph/average-pooling-2d.c",
-+    "src/src/subgraph/batch-matrix-multiply.c",
-+    "src/src/subgraph/binary.c",
-+    "src/src/subgraph/concatenate.c",
-+    "src/src/subgraph/convolution-2d.c",
-+    "src/src/subgraph/copy.c",
-+    "src/src/subgraph/deconvolution-2d.c",
-+    "src/src/subgraph/deprecated.c",
-+    "src/src/subgraph/depth-to-space-2d.c",
-+    "src/src/subgraph/depthwise-convolution-2d.c",
-+    "src/src/subgraph/even-split.c",
-+    "src/src/subgraph/fully-connected-sparse.c",
-+    "src/src/subgraph/fully-connected.c",
-+    "src/src/subgraph/max-pooling-2d.c",
-+    "src/src/subgraph/pack-lh.c",
-+    "src/src/subgraph/reshape-helpers.c",
-+    "src/src/subgraph/rope.c",
-+    "src/src/subgraph/softmax.c",
-+    "src/src/subgraph/space-to-depth-2d.c",
-+    "src/src/subgraph/static-constant-pad.c",
-+    "src/src/subgraph/static-reduce.c",
-+    "src/src/subgraph/static-resize-bilinear-2d.c",
-+    "src/src/subgraph/static-slice.c",
-+    "src/src/subgraph/static-transpose.c",
-+    "src/src/subgraph/unary.c",
-+    "src/src/subgraph/unpooling-2d.c",
-+    "src/src/subgraph/validation.c"
++    "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-scalar-u1.c"
 +  ]
  
 -  source_set("qu8-rsum_arm64") {
@@ -77788,7 +79977,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  # This is a target that cannot depend on //base.
 -  source_set("qu8-rsum_arm64_standalone") {
 -    cflags = []
-+source_set("tables_ppc64") {
++source_set("s8-rdminmax_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -77810,15 +79999,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/tables/exp2-k-over-2048.c",
-+    "src/src/tables/exp2-k-over-64.c",
-+    "src/src/tables/exp2minus-k-over-16.c",
-+    "src/src/tables/exp2minus-k-over-2048.c",
-+    "src/src/tables/exp2minus-k-over-32.c",
-+    "src/src/tables/exp2minus-k-over-4.c",
-+    "src/src/tables/exp2minus-k-over-64.c",
-+    "src/src/tables/exp2minus-k-over-8.c",
-+    "src/src/tables/vlog.c"
++    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-scalar-c2.c",
++    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-scalar-c2.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -77864,7 +80046,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("tables_ppc64_standalone") {
++source_set("s8-rdminmax_ppc64_standalone") {
 +  cflags = [
  
 -  # This is a target that cannot depend on //base.
@@ -77892,15 +80074,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/tables/exp2-k-over-2048.c",
-+    "src/src/tables/exp2-k-over-64.c",
-+    "src/src/tables/exp2minus-k-over-16.c",
-+    "src/src/tables/exp2minus-k-over-2048.c",
-+    "src/src/tables/exp2minus-k-over-32.c",
-+    "src/src/tables/exp2minus-k-over-4.c",
-+    "src/src/tables/exp2minus-k-over-64.c",
-+    "src/src/tables/exp2minus-k-over-8.c",
-+    "src/src/tables/vlog.c"
++    "src/src/s8-rdminmax/gen/s8-rdmax-2p2x-scalar-c2.c",
++    "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-scalar-c2.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -77959,7 +80134,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u1.c",
 -      "src/src/qu8-vaddc/gen/qu8-vaddc-minmax-scalar-u4.c",
 -    ]
--
++source_set("s8-rminmax_ppc64") {
++  cflags = [
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -77970,29 +80147,116 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/s8-rminmax/gen/s8-rmax-scalar-u2-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rmin-scalar-u2-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rminmax-scalar-u2-acc2.c"
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  source_set("qu8-vcvt_arm64") {
+-    cflags = []
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-vcvt/gen/qu8-vcvt-neon-u32.c",
+-      "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
 -
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("s8-rminmax_ppc64_standalone") {
++  cflags = [
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qu8-vcvt_arm64_standalone") {
+-    cflags = []
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/s8-rminmax/gen/s8-rmax-scalar-u2-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rmin-scalar-u2-acc2.c",
++    "src/src/s8-rminmax/gen/s8-rminmax-scalar-u2-acc2.c"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-vcvt/gen/qu8-vcvt-neon-u32.c",
+-      "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
 -    public_configs = [ ":xnnpack_config" ]
--
++  public_configs = [ ":xnnpack_config" ]
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
--  }
-+source_set("u8-ibilinear_ppc64") {
-+  cflags = [
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
--  source_set("qu8-vcvt_arm64") {
+-  source_set("qu8-vlrelu_arm64") {
 -    cflags = []
-+  ]
++source_set("s8-vclamp_ppc64") {
++  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-neon-u32.c",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u1.c",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c",
+-      "src/src/qu8-vlrelu/gen/qu8-vlrelu-neon-u32.c",
+-      "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/u8-ibilinear/gen/u8-ibilinear-scalar-c1.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -78005,12 +80269,20 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/s8-vclamp/s8-vclamp-scalar-u4.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--    public_configs = [ ":xnnpack_config" ]
--  }
+-  # This is a target that cannot depend on //base.
+-  source_set("qu8-vlrelu_arm64_standalone") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -78018,15 +80290,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("qu8-vcvt_arm64_standalone") {
--    cflags = []
--
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-neon-u32.c",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u1.c",
--      "src/src/qu8-vcvt/gen/qu8-vcvt-scalar-u4.c",
+-      "src/src/qu8-vlrelu/gen/qu8-vlrelu-neon-u32.c",
+-      "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -78042,7 +80309,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("u8-ibilinear_ppc64_standalone") {
++source_set("s8-vclamp_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -78054,10 +80321,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-ibilinear/gen/u8-ibilinear-scalar-c1.c"
++    "src/src/s8-vclamp/s8-vclamp-scalar-u4.c"
 +  ]
  
--  source_set("qu8-vlrelu_arm64") {
+-  source_set("qu8-vmul_arm64") {
 -    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -78065,9 +80332,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-neon-u32.c",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-select-u4.c",
+-      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c",
+-      "src/src/qu8-vmul/gen/qu8-vmul-minmax-rndnu-neon-ld64-u16.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -78095,16 +80361,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vlrelu_arm64_standalone") {
+-  source_set("qu8-vmul_arm64_standalone") {
 -    cflags = []
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-neon-u32.c",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-andxor-u4.c",
--      "src/src/qu8-vlrelu/gen/qu8-vlrelu-scalar-select-u4.c",
+-      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c",
+-      "src/src/qu8-vmul/gen/qu8-vmul-minmax-rndnu-neon-ld64-u16.c",
 -    ]
-+source_set("u8-lut32norm_ppc64") {
++source_set("subgraph_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -78122,7 +80387,35 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-lut32norm/u8-lut32norm-scalar.c"
++    "src/src/subgraph/argmax-pooling-2d.c",
++    "src/src/subgraph/average-pooling-2d.c",
++    "src/src/subgraph/batch-matrix-multiply.c",
++    "src/src/subgraph/binary.c",
++    "src/src/subgraph/concatenate.c",
++    "src/src/subgraph/convolution-2d.c",
++    "src/src/subgraph/copy.c",
++    "src/src/subgraph/deconvolution-2d.c",
++    "src/src/subgraph/deprecated.c",
++    "src/src/subgraph/depth-to-space-2d.c",
++    "src/src/subgraph/depthwise-convolution-2d.c",
++    "src/src/subgraph/even-split.c",
++    "src/src/subgraph/fully-connected-sparse.c",
++    "src/src/subgraph/fully-connected.c",
++    "src/src/subgraph/max-pooling-2d.c",
++    "src/src/subgraph/pack-lh.c",
++    "src/src/subgraph/reshape-helpers.c",
++    "src/src/subgraph/rope.c",
++    "src/src/subgraph/softmax.c",
++    "src/src/subgraph/space-to-depth-2d.c",
++    "src/src/subgraph/static-constant-pad.c",
++    "src/src/subgraph/static-reduce.c",
++    "src/src/subgraph/static-resize-bilinear-2d.c",
++    "src/src/subgraph/static-slice.c",
++    "src/src/subgraph/static-transpose.c",
++    "src/src/subgraph/subgraph-utils.c",
++    "src/src/subgraph/unary.c",
++    "src/src/subgraph/unpooling-2d.c",
++    "src/src/subgraph/validation.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -78133,7 +80426,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  source_set("qu8-vmul_arm64") {
+-  source_set("qu8-vmulc_arm64") {
 -    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -78144,8 +80437,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c",
--      "src/src/qu8-vmul/gen/qu8-vmul-minmax-rndnu-neon-ld64-u16.c",
+-      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c",
+-      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-rndnu-neon-ld64-u16.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -78161,7 +80454,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("u8-lut32norm_ppc64_standalone") {
++source_set("subgraph_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -78169,17 +80462,45 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vmul_arm64_standalone") {
+-  source_set("qu8-vmulc_arm64_standalone") {
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-lut32norm/u8-lut32norm-scalar.c"
++    "src/src/subgraph/argmax-pooling-2d.c",
++    "src/src/subgraph/average-pooling-2d.c",
++    "src/src/subgraph/batch-matrix-multiply.c",
++    "src/src/subgraph/binary.c",
++    "src/src/subgraph/concatenate.c",
++    "src/src/subgraph/convolution-2d.c",
++    "src/src/subgraph/copy.c",
++    "src/src/subgraph/deconvolution-2d.c",
++    "src/src/subgraph/deprecated.c",
++    "src/src/subgraph/depth-to-space-2d.c",
++    "src/src/subgraph/depthwise-convolution-2d.c",
++    "src/src/subgraph/even-split.c",
++    "src/src/subgraph/fully-connected-sparse.c",
++    "src/src/subgraph/fully-connected.c",
++    "src/src/subgraph/max-pooling-2d.c",
++    "src/src/subgraph/pack-lh.c",
++    "src/src/subgraph/reshape-helpers.c",
++    "src/src/subgraph/rope.c",
++    "src/src/subgraph/softmax.c",
++    "src/src/subgraph/space-to-depth-2d.c",
++    "src/src/subgraph/static-constant-pad.c",
++    "src/src/subgraph/static-reduce.c",
++    "src/src/subgraph/static-resize-bilinear-2d.c",
++    "src/src/subgraph/static-slice.c",
++    "src/src/subgraph/static-transpose.c",
++    "src/src/subgraph/subgraph-utils.c",
++    "src/src/subgraph/unary.c",
++    "src/src/subgraph/unpooling-2d.c",
++    "src/src/subgraph/validation.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmul/gen/qu8-vmul-minmax-fp32-scalar-u4.c",
--      "src/src/qu8-vmul/gen/qu8-vmul-minmax-rndnu-neon-ld64-u16.c",
+-      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c",
+-      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-rndnu-neon-ld64-u16.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -78213,15 +80534,14 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("qu8-vmulc_arm64") {
+-  source_set("qu8-vprelu_arm64") {
 -    cflags = []
-+source_set("u8-maxpool_ppc64") {
++source_set("tables_ppc64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c",
--      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-rndnu-neon-ld64-u16.c",
+-      "src/src/qu8-vprelu/gen/qu8-vprelu-scalar-u8.c",
 -    ]
 +  ]
  
@@ -78237,7 +80557,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-scalar-u1.c"
++    "src/src/tables/exp2-k-over-2048.c",
++    "src/src/tables/exp2-k-over-64.c",
++    "src/src/tables/exp2minus-k-over-16.c",
++    "src/src/tables/exp2minus-k-over-2048.c",
++    "src/src/tables/exp2minus-k-over-32.c",
++    "src/src/tables/exp2minus-k-over-4.c",
++    "src/src/tables/exp2minus-k-over-64.c",
++    "src/src/tables/exp2minus-k-over-8.c",
++    "src/src/tables/vlog.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -78247,7 +80575,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("qu8-vmulc_arm64_standalone") {
+-  source_set("qu8-vprelu_arm64_standalone") {
 -    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -78258,8 +80586,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-fp32-scalar-u4.c",
--      "src/src/qu8-vmulc/gen/qu8-vmulc-minmax-rndnu-neon-ld64-u16.c",
+-      "src/src/qu8-vprelu/gen/qu8-vprelu-scalar-u8.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -78275,7 +80602,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("u8-maxpool_ppc64_standalone") {
++source_set("tables_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -78287,10 +80614,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-scalar-u1.c"
++    "src/src/tables/exp2-k-over-2048.c",
++    "src/src/tables/exp2-k-over-64.c",
++    "src/src/tables/exp2minus-k-over-16.c",
++    "src/src/tables/exp2minus-k-over-2048.c",
++    "src/src/tables/exp2minus-k-over-32.c",
++    "src/src/tables/exp2minus-k-over-4.c",
++    "src/src/tables/exp2minus-k-over-64.c",
++    "src/src/tables/exp2minus-k-over-8.c",
++    "src/src/tables/vlog.c"
 +  ]
  
--  source_set("reference_arm64") {
+-  source_set("qu8-vpreluc_arm64") {
 -    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -78298,9 +80633,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/reference/binary-elementwise.cc",
--      "src/src/reference/packing.cc",
--      "src/src/reference/unary-elementwise.cc",
+-      "src/src/qu8-vpreluc/gen/qu8-vpreluc-scalar-u8.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -78328,16 +80661,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("reference_arm64_standalone") {
+-  source_set("qu8-vpreluc_arm64_standalone") {
 -    cflags = []
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/reference/binary-elementwise.cc",
--      "src/src/reference/packing.cc",
--      "src/src/reference/unary-elementwise.cc",
+-      "src/src/qu8-vpreluc/gen/qu8-vpreluc-scalar-u8.c",
 -    ]
-+source_set("u8-rdminmax_ppc64") {
++source_set("u8-ibilinear_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -78351,22 +80682,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-scalar-c2.c",
-+    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-scalar-c2.c"
++    "src/src/u8-ibilinear/gen/u8-ibilinear-scalar-c1.c"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  source_set("qu8-vrpreluc_arm64") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -78374,19 +80706,85 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  source_set("s8-ibilinear_arm64") {
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-vrpreluc/gen/qu8-vrpreluc-scalar-u8.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("u8-ibilinear_ppc64_standalone") {
++  cflags = [
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("qu8-vrpreluc_arm64_standalone") {
 -    cflags = []
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-ibilinear/gen/u8-ibilinear-scalar-c1.c"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/qu8-vrpreluc/gen/qu8-vrpreluc-scalar-u8.c",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
 +}
  
+-  source_set("reference_arm64") {
+-    cflags = []
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/s8-ibilinear/gen/s8-ibilinear-neon-c16.c",
--      "src/src/s8-ibilinear/gen/s8-ibilinear-neon-c8.c",
--      "src/src/s8-ibilinear/gen/s8-ibilinear-scalar-c1.c",
+-      "src/src/reference/binary-elementwise.cc",
+-      "src/src/reference/packing.cc",
+-      "src/src/reference/unary-elementwise.cc",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("u8-rdminmax_ppc64_standalone") {
++source_set("u8-lut32norm_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -78405,19 +80803,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-scalar-c2.c",
-+    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-scalar-c2.c"
++    "src/src/u8-lut32norm/u8-lut32norm-scalar.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("s8-ibilinear_arm64_standalone") {
+-  source_set("reference_arm64_standalone") {
 -    cflags = []
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/s8-ibilinear/gen/s8-ibilinear-neon-c16.c",
--      "src/src/s8-ibilinear/gen/s8-ibilinear-neon-c8.c",
--      "src/src/s8-ibilinear/gen/s8-ibilinear-scalar-c1.c",
+-      "src/src/reference/binary-elementwise.cc",
+-      "src/src/reference/packing.cc",
+-      "src/src/reference/unary-elementwise.cc",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -78437,30 +80834,34 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++    "//third_party/pthreadpool",
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
++}
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
+-  }
++# This is a target that cannot depend on //base.
++source_set("u8-lut32norm_ppc64_standalone") {
++  cflags = [
  
--  source_set("s8-maxpool_arm64") {
+-  source_set("s8-ibilinear_arm64") {
 -    cflags = []
-+source_set("u8-rminmax_ppc64") {
-+  cflags = [
++  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-neon-u16.c",
--      "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-scalar-u1.c",
+-      "src/src/s8-ibilinear/gen/s8-ibilinear-neon-c16.c",
+-      "src/src/s8-ibilinear/gen/s8-ibilinear-neon-c8.c",
+-      "src/src/s8-ibilinear/gen/s8-ibilinear-scalar-c1.c",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-lut32norm/u8-lut32norm-scalar.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -78473,22 +80874,82 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/u8-rminmax/gen/u8-rmax-scalar-u2-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rmin-scalar-u2-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rminmax-scalar-u2-acc2.c"
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
+-  # This is a target that cannot depend on //base.
+-  source_set("s8-ibilinear_arm64_standalone") {
+-    cflags = []
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/s8-ibilinear/gen/s8-ibilinear-neon-c16.c",
+-      "src/src/s8-ibilinear/gen/s8-ibilinear-neon-c8.c",
+-      "src/src/s8-ibilinear/gen/s8-ibilinear-scalar-c1.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
++  }
++}
+ 
 -    public_configs = [ ":xnnpack_config" ]
++source_set("u8-maxpool_ppc64") {
++  cflags = [
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
 -  }
++  ]
+ 
+-  source_set("s8-maxpool_arm64") {
+-    cflags = []
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-scalar-u1.c"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-neon-u16.c",
+-      "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-scalar-u1.c",
+-    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
--  # This is a target that cannot depend on //base.
--  source_set("s8-maxpool_arm64_standalone") {
--    cflags = []
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -78496,13 +80957,24 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("s8-maxpool_arm64_standalone") {
+-    cflags = []
++# This is a target that cannot depend on //base.
++source_set("u8-maxpool_ppc64_standalone") {
++  cflags = [
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-neon-u16.c",
 -      "src/src/s8-maxpool/gen/s8-maxpool-9p-minmax-scalar-u1.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -78514,12 +80986,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("u8-rminmax_ppc64_standalone") {
-+  cflags = [
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-maxpool/gen/u8-maxpool-9p-minmax-scalar-u1.c"
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
@@ -78536,11 +81011,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-neon-c32.c",
 -      "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-scalar-c2.c",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/u8-rminmax/gen/u8-rmax-scalar-u2-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rmin-scalar-u2-acc2.c",
-+    "src/src/u8-rminmax/gen/u8-rminmax-scalar-u2-acc2.c"
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -78553,18 +81028,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
 -  source_set("s8-rdminmax_arm64_standalone") {
@@ -78577,8 +81047,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-neon-c32.c",
 -      "src/src/s8-rdminmax/gen/s8-rdmin-2p2x-scalar-c2.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -78589,14 +81058,11 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-+  }
-+}
++source_set("u8-rdminmax_ppc64") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
-+source_set("u8-vclamp_ppc64") {
-+  cflags = [
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
@@ -78615,6 +81081,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/s8-rminmax/gen/s8-rminmax-neon-u32-acc2.c",
 -      "src/src/s8-rminmax/gen/s8-rminmax-scalar-u2-acc2.c",
 -    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-scalar-c2.c",
++    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-scalar-c2.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -78627,17 +81097,19 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/u8-vclamp/u8-vclamp-scalar-u4.c"
-+  ]
- 
--    public_configs = [ ":xnnpack_config" ]
--  }
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
 -  # This is a target that cannot depend on //base.
 -  source_set("s8-rminmax_arm64_standalone") {
 -    cflags = []
@@ -78651,12 +81123,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/s8-rminmax/gen/s8-rminmax-neon-u32-acc2.c",
 -      "src/src/s8-rminmax/gen/s8-rminmax-scalar-u2-acc2.c",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -78668,35 +81136,40 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
-+source_set("u8-vclamp_ppc64_standalone") {
++source_set("u8-rdminmax_ppc64_standalone") {
 +  cflags = [
  
+-    public_configs = [ ":xnnpack_config" ]
++  ]
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/u8-rdminmax/gen/u8-rdmax-2p2x-scalar-c2.c",
++    "src/src/u8-rdminmax/gen/u8-rdmin-2p2x-scalar-c2.c"
 +  ]
  
 -  source_set("s8-vclamp_arm64") {
 -    cflags = []
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/u8-vclamp/u8-vclamp-scalar-u4.c"
-+  ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
 -      "src/src/s8-vclamp/s8-vclamp-neon-u64.c",
 -      "src/src/s8-vclamp/s8-vclamp-scalar-u4.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -78708,13 +81181,6 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
-+  ]
-+
 +  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -78726,7 +81192,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  # This is a target that cannot depend on //base.
 -  source_set("s8-vclamp_arm64_standalone") {
 -    cflags = []
-+source_set("x16-transposec_ppc64") {
++source_set("u8-rminmax_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -78748,7 +81214,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c"
++    "src/src/u8-rminmax/gen/u8-rmax-scalar-u2-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rmin-scalar-u2-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rminmax-scalar-u2-acc2.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -78791,6 +81259,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/subgraph/static-resize-bilinear-2d.c",
 -      "src/src/subgraph/static-slice.c",
 -      "src/src/subgraph/static-transpose.c",
+-      "src/src/subgraph/subgraph-utils.c",
 -      "src/src/subgraph/unary.c",
 -      "src/src/subgraph/unpooling-2d.c",
 -      "src/src/subgraph/validation.c",
@@ -78818,7 +81287,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("x16-transposec_ppc64_standalone") {
++source_set("u8-rminmax_ppc64_standalone") {
 +  cflags = [
  
 -  # This is a target that cannot depend on //base.
@@ -78852,6 +81321,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/subgraph/static-resize-bilinear-2d.c",
 -      "src/src/subgraph/static-slice.c",
 -      "src/src/subgraph/static-transpose.c",
+-      "src/src/subgraph/subgraph-utils.c",
 -      "src/src/subgraph/unary.c",
 -      "src/src/subgraph/unpooling-2d.c",
 -      "src/src/subgraph/validation.c",
@@ -78870,7 +81340,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c"
++    "src/src/u8-rminmax/gen/u8-rmax-scalar-u2-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rmin-scalar-u2-acc2.c",
++    "src/src/u8-rminmax/gen/u8-rminmax-scalar-u2-acc2.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -78939,7 +81411,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/tables/exp2minus-k-over-8.c",
 -      "src/src/tables/vlog.c",
 -    ]
-+source_set("x16-x32-packw_ppc64") {
++source_set("u8-vclamp_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -78957,8 +81429,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-gio-scalar.c",
-+    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-goi-scalar.c"
++    "src/src/u8-vclamp/u8-vclamp-scalar-u4.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -78998,7 +81469,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x16-x32-packw_ppc64_standalone") {
++source_set("u8-vclamp_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -79017,8 +81488,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-gio-scalar.c",
-+    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-goi-scalar.c"
++    "src/src/u8-vclamp/u8-vclamp-scalar-u4.c"
 +  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -79034,15 +81504,15 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
--    public_configs = [ ":xnnpack_config" ]
++
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
-+
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -79055,7 +81525,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -  source_set("u8-lut32norm_arm64") {
 -    cflags = []
-+source_set("x24-transposec_ppc64") {
++source_set("x16-transposec_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -79076,7 +81546,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x24-transposec/gen/x24-transposec-1x2-scalar.c"
++    "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -79113,7 +81583,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x24-transposec_ppc64_standalone") {
++source_set("x16-transposec_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -79125,7 +81595,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  }
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x24-transposec/gen/x24-transposec-1x2-scalar.c"
++    "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c"
 +  ]
  
 -  source_set("u8-maxpool_arm64") {
@@ -79167,7 +81637,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -  # This is a target that cannot depend on //base.
 -  source_set("u8-maxpool_arm64_standalone") {
 -    cflags = []
-+source_set("x32-packw_ppc64") {
++source_set("x16-x32-packw_ppc64") {
 +  cflags = [
  
 -    sources = [
@@ -79189,12 +81659,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x32-packw/gen/x32-packw-x2-gemm-gio-scalar.c",
-+    "src/src/x32-packw/gen/x32-packw-x2-gemm-goi-scalar-float-u4.c",
-+    "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-scalar-int-u2.c",
-+    "src/src/x32-packw/gen/x32-packw-x4-gemm-gio-scalar.c",
-+    "src/src/x32-packw/gen/x32-packw-x4-gemm-goi-scalar-float-u4.c",
-+    "src/src/x32-packw/gen/x32-packw-x64-gemm-goi-scalar-int-u2.c"
++    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-gio-scalar.c",
++    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-goi-scalar.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -79240,7 +81706,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 -  }
 +# This is a target that cannot depend on //base.
-+source_set("x32-packw_ppc64_standalone") {
++source_set("x16-x32-packw_ppc64_standalone") {
 +  cflags = [
  
 -  # This is a target that cannot depend on //base.
@@ -79268,12 +81734,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x32-packw/gen/x32-packw-x2-gemm-gio-scalar.c",
-+    "src/src/x32-packw/gen/x32-packw-x2-gemm-goi-scalar-float-u4.c",
-+    "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-scalar-int-u2.c",
-+    "src/src/x32-packw/gen/x32-packw-x4-gemm-gio-scalar.c",
-+    "src/src/x32-packw/gen/x32-packw-x4-gemm-goi-scalar-float-u4.c",
-+    "src/src/x32-packw/gen/x32-packw-x64-gemm-goi-scalar-int-u2.c"
++    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-gio-scalar.c",
++    "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-goi-scalar.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -79336,7 +81798,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "src/src/u8-rminmax/gen/u8-rminmax-neon-u32-acc2.c",
 -      "src/src/u8-rminmax/gen/u8-rminmax-scalar-u2-acc2.c",
 -    ]
-+source_set("x32-transposec_ppc64") {
++source_set("x24-transposec_ppc64") {
 +  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
@@ -79354,7 +81816,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x32-transposec/gen/x32-transposec-2x4-scalar-int.c"
++    "src/src/x24-transposec/gen/x24-transposec-1x2-scalar.c"
 +  ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -79393,7 +81855,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/pthreadpool",
 -    ]
 +# This is a target that cannot depend on //base.
-+source_set("x32-transposec_ppc64_standalone") {
++source_set("x24-transposec_ppc64_standalone") {
 +  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
@@ -79405,7 +81867,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 -    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
-+    "src/src/x32-transposec/gen/x32-transposec-2x4-scalar-int.c"
++    "src/src/x24-transposec/gen/x24-transposec-1x2-scalar.c"
 +  ]
  
 -    sources = [
@@ -79445,15 +81907,254 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("x16-packw_arm64") {
+-  source_set("x16-pack-lh_arch=armv8.2-a+sve+sve2") {
+-    cflags = [ "-march=armv8.2-a+sve+sve2" ]
++source_set("x32-packw_ppc64") {
++  cflags = [
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/x16-pack-lh/x16-packlh-neonsme2.c",
+-    ]
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-packw/gen/x32-packw-x2-gemm-gio-scalar.c",
++    "src/src/x32-packw/gen/x32-packw-x2-gemm-goi-scalar-float-u4.c",
++    "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-scalar-int-u2.c",
++    "src/src/x32-packw/gen/x32-packw-x4-gemm-gio-scalar.c",
++    "src/src/x32-packw/gen/x32-packw-x4-gemm-goi-scalar-float-u4.c",
++    "src/src/x32-packw/gen/x32-packw-x64-gemm-goi-scalar-int-u2.c"
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("x16-pack-lh_arch=armv8.2-a+sve+sve2_standalone") {
+-    cflags = [ "-march=armv8.2-a+sve+sve2" ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/x16-pack-lh/x16-packlh-neonsme2.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("x32-packw_ppc64_standalone") {
++  cflags = [
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-packw/gen/x32-packw-x2-gemm-gio-scalar.c",
++    "src/src/x32-packw/gen/x32-packw-x2-gemm-goi-scalar-float-u4.c",
++    "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-scalar-int-u2.c",
++    "src/src/x32-packw/gen/x32-packw-x4-gemm-gio-scalar.c",
++    "src/src/x32-packw/gen/x32-packw-x4-gemm-goi-scalar-float-u4.c",
++    "src/src/x32-packw/gen/x32-packw-x64-gemm-goi-scalar-int-u2.c"
++  ]
+ 
+-  source_set("x16-packw_arm64") {
+-    cflags = []
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/x16-packw/gen/x16-packw-x16-gemm-goi-neon-ld4lane-u8-prfm.c",
+-      "src/src/x16-packw/gen/x16-packw-x8-gemm-goi-neon-ld4lane-u8-prfm.c",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("x16-packw_arm64_standalone") {
+-    cflags = []
+-
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/x16-packw/gen/x16-packw-x16-gemm-goi-neon-ld4lane-u8-prfm.c",
+-      "src/src/x16-packw/gen/x16-packw-x8-gemm-goi-neon-ld4lane-u8-prfm.c",
+-    ]
++source_set("x32-transposec_ppc64") {
++  cflags = [
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-transposec/gen/x32-transposec-2x4-scalar-int.c"
++  ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
+-  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  source_set("x16-transposec_arm64") {
+-    cflags = []
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c",
+-      "src/src/x16-transposec/gen/x16-transposec-8x8-reuse-dec-zip-neon.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool",
+-    ]
++# This is a target that cannot depend on //base.
++source_set("x32-transposec_ppc64_standalone") {
++  cflags = [
+ 
+-    public_configs = [ ":xnnpack_config" ]
+-  }
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("x16-transposec_arm64_standalone") {
+-    cflags = []
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x32-transposec/gen/x32-transposec-2x4-scalar-int.c"
++  ]
+ 
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c",
+-      "src/src/x16-transposec/gen/x16-transposec-8x8-reuse-dec-zip-neon.c",
+-    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    configs -= [ "//build/config/compiler:chromium_code" ]
+-    configs += [ "//build/config/compiler:no_chromium_code" ]
+-    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+-
+-    deps = [
+-      "//third_party/cpuinfo",
+-      "//third_party/fp16",
+-      "//third_party/fxdiv",
+-      "//third_party/pthreadpool:pthreadpool_standalone",
+-    ]
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
++  public_configs = [ ":xnnpack_config" ]
+ 
+-    if (!(is_android && use_order_profiling)) {
+-      assert_no_deps = [ "//base" ]
+-    }
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
+ 
+-  source_set("x16-x32-packw_arm64") {
 -    cflags = []
 +source_set("x32-unpool_ppc64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x16-packw/gen/x16-packw-x16-gemm-goi-neon-ld4lane-u8-prfm.c",
--      "src/src/x16-packw/gen/x16-packw-x8-gemm-goi-neon-ld4lane-u8-prfm.c",
+-      "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-gio-scalar.c",
+-      "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-goi-scalar.c",
 -    ]
 +  ]
  
@@ -79479,7 +82180,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("x16-packw_arm64_standalone") {
+-  source_set("x16-x32-packw_arm64_standalone") {
 -    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -79490,8 +82191,8 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x16-packw/gen/x16-packw-x16-gemm-goi-neon-ld4lane-u8-prfm.c",
--      "src/src/x16-packw/gen/x16-packw-x8-gemm-goi-neon-ld4lane-u8-prfm.c",
+-      "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-gio-scalar.c",
+-      "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-goi-scalar.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}
@@ -79522,7 +82223,7 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/x32-unpool/x32-unpool-scalar.c"
 +  ]
  
--  source_set("x16-transposec_arm64") {
+-  source_set("x24-transposec_arm64") {
 -    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -79530,8 +82231,9 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c",
--      "src/src/x16-transposec/gen/x16-transposec-8x8-reuse-dec-zip-neon.c",
+-      "src/src/x24-transposec/gen/x24-transposec-1x2-scalar.c",
+-      "src/src/x24-transposec/x24-transposec-2x2-neon-tbl64.c",
+-      "src/src/x24-transposec/x24-transposec-4x4-aarch64-neon-tbl128.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -79559,17 +82261,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("x16-transposec_arm64_standalone") {
+-  source_set("x24-transposec_arm64_standalone") {
 -    cflags = []
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x16-transposec/gen/x16-transposec-2x4-scalar-int.c",
--      "src/src/x16-transposec/gen/x16-transposec-8x8-reuse-dec-zip-neon.c",
+-      "src/src/x24-transposec/gen/x24-transposec-1x2-scalar.c",
+-      "src/src/x24-transposec/x24-transposec-2x2-neon-tbl64.c",
+-      "src/src/x24-transposec/x24-transposec-4x4-aarch64-neon-tbl128.c",
 -    ]
-+source_set("x64-transposec_ppc64") {
-+  cflags = [
- 
+-
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -79580,38 +82281,31 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  ]
- 
+-
 -    public_configs = [ ":xnnpack_config" ]
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c"
-+  ]
++source_set("x64-transposec_ppc64") {
++  cflags = [
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  ]
  
--  source_set("x16-x32-packw_arm64") {
--    cflags = []
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
+-  source_set("x32-pack-lh_arch=armv8.2-a+sve+sve2") {
+-    cflags = [ "-march=armv8.2-a+sve+sve2" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c"
 +  ]
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-gio-scalar.c",
--      "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-goi-scalar.c",
+-      "src/src/x32-pack-lh/x32-packlh-neonsme2.c",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -79623,30 +82317,30 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("x64-transposec_ppc64_standalone") {
-+  cflags = [
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  ]
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -  # This is a target that cannot depend on //base.
--  source_set("x16-x32-packw_arm64_standalone") {
--    cflags = []
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c"
-+  ]
+-  source_set("x32-pack-lh_arch=armv8.2-a+sve+sve2_standalone") {
+-    cflags = [ "-march=armv8.2-a+sve+sve2" ]
++# This is a target that cannot depend on //base.
++source_set("x64-transposec_ppc64_standalone") {
++  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-gio-scalar.c",
--      "src/src/x16-x32-packw/gen/x16-x32-packw-x32c2-gemm-goi-scalar.c",
+-      "src/src/x32-pack-lh/x32-packlh-neonsme2.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -79658,35 +82352,42 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool:pthreadpool_standalone",
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c"
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
-+  if (!(is_android && use_order_profiling)) {
-+    assert_no_deps = [ "//base" ]
-   }
-+}
- 
--  source_set("x24-transposec_arm64") {
+-  }
+-
+-  source_set("x32-packw_arm64") {
 -    cflags = []
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x24-transposec/gen/x24-transposec-1x2-scalar.c",
--      "src/src/x24-transposec/x24-transposec-2x2-neon-tbl64.c",
--      "src/src/x24-transposec/x24-transposec-4x4-aarch64-neon-tbl128.c",
+-      "src/src/x32-packw/gen/x32-packw-x2-gemm-gio-scalar.c",
+-      "src/src/x32-packw/gen/x32-packw-x2-gemm-goi-neon-ld2lane-u2-prfm.c",
+-      "src/src/x32-packw/gen/x32-packw-x2-gemm-goi-scalar-float-u4.c",
+-      "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-scalar-int-u2.c",
+-      "src/src/x32-packw/gen/x32-packw-x4-gemm-gio-scalar.c",
+-      "src/src/x32-packw/gen/x32-packw-x4-gemm-goi-scalar-float-u4.c",
+-      "src/src/x32-packw/gen/x32-packw-x64-gemm-goi-scalar-int-u2.c",
+-      "src/src/x32-packw/gen/x32-packw-x8-gemm-goi-neon-ld4lane-u4-prfm.c",
+-      "src/src/x32-packw/gen/x32-packw-x8s4-gemm-goi-neon-ld4lane-u4-prfm.c",
 -    ]
-+source_set("x8-lut_ppc64") {
-+  cflags = [
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool:pthreadpool_standalone",
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -79698,28 +82399,32 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  ]
++  public_configs = [ ":xnnpack_config" ]
  
 -    public_configs = [ ":xnnpack_config" ]
--  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x8-lut/gen/x8-lut-scalar-u4.c"
-+  ]
++  if (!(is_android && use_order_profiling)) {
++    assert_no_deps = [ "//base" ]
+   }
++}
  
 -  # This is a target that cannot depend on //base.
--  source_set("x24-transposec_arm64_standalone") {
+-  source_set("x32-packw_arm64_standalone") {
 -    cflags = []
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x24-transposec/gen/x24-transposec-1x2-scalar.c",
--      "src/src/x24-transposec/x24-transposec-2x2-neon-tbl64.c",
--      "src/src/x24-transposec/x24-transposec-4x4-aarch64-neon-tbl128.c",
+-      "src/src/x32-packw/gen/x32-packw-x2-gemm-gio-scalar.c",
+-      "src/src/x32-packw/gen/x32-packw-x2-gemm-goi-neon-ld2lane-u2-prfm.c",
+-      "src/src/x32-packw/gen/x32-packw-x2-gemm-goi-scalar-float-u4.c",
+-      "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-scalar-int-u2.c",
+-      "src/src/x32-packw/gen/x32-packw-x4-gemm-gio-scalar.c",
+-      "src/src/x32-packw/gen/x32-packw-x4-gemm-goi-scalar-float-u4.c",
+-      "src/src/x32-packw/gen/x32-packw-x64-gemm-goi-scalar-int-u2.c",
+-      "src/src/x32-packw/gen/x32-packw-x8-gemm-goi-neon-ld4lane-u4-prfm.c",
+-      "src/src/x32-packw/gen/x32-packw-x8s4-gemm-goi-neon-ld4lane-u4-prfm.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++source_set("x8-lut_ppc64") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -79731,40 +82436,39 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  deps = [
-+    "//third_party/cpuinfo",
-+    "//third_party/fp16",
-+    "//third_party/fxdiv",
-+    "//third_party/pthreadpool",
 +  ]
  
 -    public_configs = [ ":xnnpack_config" ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-lut/gen/x8-lut-scalar-u4.c"
++  ]
  
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
--
--  source_set("x32-packw_arm64") {
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  source_set("x32-transposec_arm64") {
 -    cflags = []
--
++  deps = [
++    "//third_party/cpuinfo",
++    "//third_party/fp16",
++    "//third_party/fxdiv",
++    "//third_party/pthreadpool",
++  ]
+ 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x32-packw/gen/x32-packw-x2-gemm-gio-scalar.c",
--      "src/src/x32-packw/gen/x32-packw-x2-gemm-goi-neon-ld2lane-u2-prfm.c",
--      "src/src/x32-packw/gen/x32-packw-x2-gemm-goi-scalar-float-u4.c",
--      "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-scalar-int-u2.c",
--      "src/src/x32-packw/gen/x32-packw-x4-gemm-gio-scalar.c",
--      "src/src/x32-packw/gen/x32-packw-x4-gemm-goi-scalar-float-u4.c",
--      "src/src/x32-packw/gen/x32-packw-x64-gemm-goi-scalar-int-u2.c",
--      "src/src/x32-packw/gen/x32-packw-x8-gemm-goi-neon-ld4lane-u4-prfm.c",
--      "src/src/x32-packw/gen/x32-packw-x8s4-gemm-goi-neon-ld4lane-u4-prfm.c",
+-      "src/src/x32-transposec/gen/x32-transposec-2x4-scalar-int.c",
+-      "src/src/x32-transposec/gen/x32-transposec-4x4-reuse-dec-zip-neon.c",
+-      "src/src/x32-transposec/x32-transposec-4x4-aarch64-neon-tbl128.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("x8-lut_ppc64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -79776,34 +82480,28 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  ]
++# This is a target that cannot depend on //base.
++source_set("x8-lut_ppc64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
-+  sources = [
-+    "src/include/xnnpack.h",
-+    "src/src/x8-lut/gen/x8-lut-scalar-u4.c"
 +  ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("x32-packw_arm64_standalone") {
+-  source_set("x32-transposec_arm64_standalone") {
 -    cflags = []
 -
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x32-packw/gen/x32-packw-x2-gemm-gio-scalar.c",
--      "src/src/x32-packw/gen/x32-packw-x2-gemm-goi-neon-ld2lane-u2-prfm.c",
--      "src/src/x32-packw/gen/x32-packw-x2-gemm-goi-scalar-float-u4.c",
--      "src/src/x32-packw/gen/x32-packw-x32-gemm-goi-scalar-int-u2.c",
--      "src/src/x32-packw/gen/x32-packw-x4-gemm-gio-scalar.c",
--      "src/src/x32-packw/gen/x32-packw-x4-gemm-goi-scalar-float-u4.c",
--      "src/src/x32-packw/gen/x32-packw-x64-gemm-goi-scalar-int-u2.c",
--      "src/src/x32-packw/gen/x32-packw-x8-gemm-goi-neon-ld4lane-u4-prfm.c",
--      "src/src/x32-packw/gen/x32-packw-x8s4-gemm-goi-neon-ld4lane-u4-prfm.c",
+-      "src/src/x32-transposec/gen/x32-transposec-2x4-scalar-int.c",
+-      "src/src/x32-transposec/gen/x32-transposec-4x4-reuse-dec-zip-neon.c",
+-      "src/src/x32-transposec/x32-transposec-4x4-aarch64-neon-tbl128.c",
 -    ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
++  sources = [
++    "src/include/xnnpack.h",
++    "src/src/x8-lut/gen/x8-lut-scalar-u4.c"
++  ]
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -79815,14 +82513,18 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
 +    "//third_party/fxdiv",
 +    "//third_party/pthreadpool:pthreadpool_standalone",
 +  ]
- 
--    public_configs = [ ":xnnpack_config" ]
++
 +  public_configs = [ ":xnnpack_config" ]
  
 -    if (!(is_android && use_order_profiling)) {
@@ -79833,16 +82535,15 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("x32-transposec_arm64") {
+-  source_set("x32-unpool_arm64") {
 -    cflags = []
 +source_set("x8-packq_ppc64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x32-transposec/gen/x32-transposec-2x4-scalar-int.c",
--      "src/src/x32-transposec/gen/x32-transposec-4x4-reuse-dec-zip-neon.c",
--      "src/src/x32-transposec/x32-transposec-4x4-aarch64-neon-tbl128.c",
+-      "src/src/x32-unpool/x32-unpool-neon.c",
+-      "src/src/x32-unpool/x32-unpool-scalar.c",
 -    ]
 +  ]
  
@@ -79868,15 +82569,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("x32-transposec_arm64_standalone") {
+-  source_set("x32-unpool_arm64_standalone") {
 -    cflags = []
--
--    sources = [
--      "src/include/xnnpack.h",
--      "src/src/x32-transposec/gen/x32-transposec-2x4-scalar-int.c",
--      "src/src/x32-transposec/gen/x32-transposec-4x4-reuse-dec-zip-neon.c",
--      "src/src/x32-transposec/x32-transposec-4x4-aarch64-neon-tbl128.c",
--    ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -79884,6 +82578,14 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
+-    sources = [
+-      "src/include/xnnpack.h",
+-      "src/src/x32-unpool/x32-unpool-neon.c",
+-      "src/src/x32-unpool/x32-unpool-scalar.c",
+-    ]
++  public_configs = [ ":xnnpack_config" ]
++}
+ 
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
 -    configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
@@ -79894,26 +82596,23 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
--    public_configs = [ ":xnnpack_config" ]
 +# This is a target that cannot depend on //base.
 +source_set("x8-packq_ppc64_standalone") {
 +  cflags = [
  
+-    public_configs = [ ":xnnpack_config" ]
++  ]
+ 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
-+  ]
-+
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/x8-packq/x8-packq-scalar-f32qp8-u1.c"
 +  ]
  
--  source_set("x32-unpool_arm64") {
+-  source_set("x64-transposec_arm64") {
 -    cflags = []
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -79921,8 +82620,9 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x32-unpool/x32-unpool-neon.c",
--      "src/src/x32-unpool/x32-unpool-scalar.c",
+-      "src/src/x64-transposec/gen/x64-transposec-2x2-multi-dec-zip-neon.c",
+-      "src/src/x64-transposec/gen/x64-transposec-2x2-reuse-dec-zip-neon.c",
+-      "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c",
 -    ]
 +  deps = [
 +    "//third_party/cpuinfo",
@@ -79950,17 +82650,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 +}
  
 -  # This is a target that cannot depend on //base.
--  source_set("x32-unpool_arm64_standalone") {
+-  source_set("x64-transposec_arm64_standalone") {
 -    cflags = []
-+source_set("x8-packw_ppc64") {
-+  cflags = [
- 
+-
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x32-unpool/x32-unpool-neon.c",
--      "src/src/x32-unpool/x32-unpool-scalar.c",
+-      "src/src/x64-transposec/gen/x64-transposec-2x2-multi-dec-zip-neon.c",
+-      "src/src/x64-transposec/gen/x64-transposec-2x2-reuse-dec-zip-neon.c",
+-      "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c",
 -    ]
-+  ]
++source_set("x8-packw_ppc64") {
++  cflags = [
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -79972,6 +82672,9 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool:pthreadpool_standalone",
 -    ]
++  ]
+ 
+-    public_configs = [ ":xnnpack_config" ]
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/x8-packw/gen/x8-packw-x16-gemm-goi-scalar-u2.c",
@@ -79980,15 +82683,16 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/x8-packw/gen/x8-packw-x8-gemm-goi-scalar-u2.c"
 +  ]
  
--    public_configs = [ ":xnnpack_config" ]
-+  configs -= [ "//build/config/compiler:chromium_code" ]
-+  configs += [ "//build/config/compiler:no_chromium_code" ]
-+  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
- 
 -    if (!(is_android && use_order_profiling)) {
 -      assert_no_deps = [ "//base" ]
 -    }
 -  }
++  configs -= [ "//build/config/compiler:chromium_code" ]
++  configs += [ "//build/config/compiler:no_chromium_code" ]
++  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
+ 
+-  source_set("x8-lut_arm64") {
+-    cflags = []
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -79996,20 +82700,13 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "//third_party/pthreadpool",
 +  ]
  
--  source_set("x64-transposec_arm64") {
--    cflags = []
-+  public_configs = [ ":xnnpack_config" ]
-+}
- 
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x64-transposec/gen/x64-transposec-2x2-multi-dec-zip-neon.c",
--      "src/src/x64-transposec/gen/x64-transposec-2x2-reuse-dec-zip-neon.c",
--      "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c",
+-      "src/src/x8-lut/gen/x8-lut-aarch64-neon-tbx128x4-u64.c",
+-      "src/src/x8-lut/gen/x8-lut-scalar-u4.c",
 -    ]
-+# This is a target that cannot depend on //base.
-+source_set("x8-packw_ppc64_standalone") {
-+  cflags = [
++  public_configs = [ ":xnnpack_config" ]
++}
  
 -    configs -= [ "//build/config/compiler:chromium_code" ]
 -    configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -80021,10 +82718,17 @@ File regenerated by running 'python3 generate_build_gn.py'
 -      "//third_party/fxdiv",
 -      "//third_party/pthreadpool",
 -    ]
-+  ]
++# This is a target that cannot depend on //base.
++source_set("x8-packw_ppc64_standalone") {
++  cflags = [
  
 -    public_configs = [ ":xnnpack_config" ]
 -  }
++  ]
+ 
+-  # This is a target that cannot depend on //base.
+-  source_set("x8-lut_arm64_standalone") {
+-    cflags = []
 +  sources = [
 +    "src/include/xnnpack.h",
 +    "src/src/x8-packw/gen/x8-packw-x16-gemm-goi-scalar-u2.c",
@@ -80033,15 +82737,10 @@ File regenerated by running 'python3 generate_build_gn.py'
 +    "src/src/x8-packw/gen/x8-packw-x8-gemm-goi-scalar-u2.c"
 +  ]
  
--  # This is a target that cannot depend on //base.
--  source_set("x64-transposec_arm64_standalone") {
--    cflags = []
--
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x64-transposec/gen/x64-transposec-2x2-multi-dec-zip-neon.c",
--      "src/src/x64-transposec/gen/x64-transposec-2x2-reuse-dec-zip-neon.c",
--      "src/src/x64-transposec/gen/x64-transposec-4x2-scalar-int.c",
+-      "src/src/x8-lut/gen/x8-lut-aarch64-neon-tbx128x4-u64.c",
+-      "src/src/x8-lut/gen/x8-lut-scalar-u4.c",
 -    ]
 +  configs -= [ "//build/config/compiler:chromium_code" ]
 +  configs += [ "//build/config/compiler:no_chromium_code" ]
@@ -80075,15 +82774,14 @@ File regenerated by running 'python3 generate_build_gn.py'
    }
 +}
  
--  source_set("x8-lut_arm64") {
--    cflags = []
+-  source_set("x8-pack-lh_arch=armv8.2-a+sve+sve2") {
+-    cflags = [ "-march=armv8.2-a+sve+sve2" ]
 +source_set("x8-transposec_ppc64") {
 +  cflags = [
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x8-lut/gen/x8-lut-aarch64-neon-tbx128x4-u64.c",
--      "src/src/x8-lut/gen/x8-lut-scalar-u4.c",
+-      "src/src/x8-pack-lh/x8--packlh-neonsme2.c",
 -    ]
 +  ]
  
@@ -80109,8 +82807,8 @@ File regenerated by running 'python3 generate_build_gn.py'
 +  configs += [ "//build/config/sanitizers:cfi_icall_generalize_pointers" ]
  
 -  # This is a target that cannot depend on //base.
--  source_set("x8-lut_arm64_standalone") {
--    cflags = []
+-  source_set("x8-pack-lh_arch=armv8.2-a+sve+sve2_standalone") {
+-    cflags = [ "-march=armv8.2-a+sve+sve2" ]
 +  deps = [
 +    "//third_party/cpuinfo",
 +    "//third_party/fp16",
@@ -80120,8 +82818,7 @@ File regenerated by running 'python3 generate_build_gn.py'
  
 -    sources = [
 -      "src/include/xnnpack.h",
--      "src/src/x8-lut/gen/x8-lut-aarch64-neon-tbx128x4-u64.c",
--      "src/src/x8-lut/gen/x8-lut-scalar-u4.c",
+-      "src/src/x8-pack-lh/x8--packlh-neonsme2.c",
 -    ]
 +  public_configs = [ ":xnnpack_config" ]
 +}

diff --git a/chromium-137-pdfium-png_decoder-build-error.patch b/chromium-137-pdfium-png_decoder-build-error.patch
deleted file mode 100644
index f45c757..0000000
--- a/chromium-137-pdfium-png_decoder-build-error.patch
+++ /dev/null
@@ -1,14 +0,0 @@
-diff -up chromium-137.0.7151.15/third_party/pdfium/core/fxcodec/png/png_decoder.cpp.me chromium-137.0.7151.15/third_party/pdfium/core/fxcodec/png/png_decoder.cpp
---- chromium-137.0.7151.15/third_party/pdfium/core/fxcodec/png/png_decoder.cpp.me	2025-05-10 23:12:58.189339791 +0200
-+++ chromium-137.0.7151.15/third_party/pdfium/core/fxcodec/png/png_decoder.cpp	2025-05-10 23:13:54.931224351 +0200
-@@ -52,8 +52,8 @@ void _png_load_bmp_attribute(png_structp
-                              CFX_DIBAttribute* pAttribute) {
-   if (pAttribute) {
- #if defined(PNG_pHYs_SUPPORTED)
--    pAttribute->xdpi_ = png_get_x_pixels_per_meter(png_ptr, info_ptr);
--    pAttribute->ydpi_ = png_get_y_pixels_per_meter(png_ptr, info_ptr);
-+    pAttribute->x_dpi_ = png_get_x_pixels_per_meter(png_ptr, info_ptr);
-+    pAttribute->y_dpi_ = png_get_y_pixels_per_meter(png_ptr, info_ptr);
-     png_uint_32 res_x, res_y;
-     int unit_type;
-     png_get_pHYs(png_ptr, info_ptr, &res_x, &res_y, &unit_type);

diff --git a/chromium-137-simdutf-7.x-build-error.patch b/chromium-137-simdutf-7.x-build-error.patch
deleted file mode 100644
index 5e33d54..0000000
--- a/chromium-137-simdutf-7.x-build-error.patch
+++ /dev/null
@@ -1,21 +0,0 @@
-diff -up chromium-137.0.7151.40/v8/src/builtins/builtins-typed-array.cc.simdutf-7.x chromium-137.0.7151.40/v8/src/builtins/builtins-typed-array.cc
---- chromium-137.0.7151.40/v8/src/builtins/builtins-typed-array.cc.simdutf-7.x	2025-05-21 00:14:25.000000000 +0200
-+++ chromium-137.0.7151.40/v8/src/builtins/builtins-typed-array.cc	2025-05-23 23:31:59.961068931 +0200
-@@ -493,7 +493,7 @@ Maybe<simdutf::result> ArrayBufferFromBa
-   std::unique_ptr<char[]> output = std::make_unique<char[]>(output_length);
-   simdutf::result simd_result = simdutf::base64_to_binary_safe(
-       reinterpret_cast<const T>(input_vector), input_length, output.get(),
--      output_length, alphabet, last_chunk_handling);
-+      output_length, alphabet, last_chunk_handling, false);
- 
-   {
-     AllowGarbageCollection gc;
-@@ -526,7 +526,7 @@ simdutf::result ArrayBufferSetFromBase64
-   simdutf::result simd_result = simdutf::base64_to_binary_safe(
-       reinterpret_cast<const T>(input_vector), input_length,
-       reinterpret_cast<char*>(typed_array->DataPtr()), output_length, alphabet,
--      last_chunk_handling);
-+      last_chunk_handling, false);
- 
-   return simd_result;
- }

diff --git a/chromium-137-simdutf-build-error.patch b/chromium-137-simdutf-build-error.patch
deleted file mode 100644
index 50441b7..0000000
--- a/chromium-137-simdutf-build-error.patch
+++ /dev/null
@@ -1,12 +0,0 @@
-diff -up chromium-137.0.7151.40/v8/src/builtins/builtins-typed-array.cc.me chromium-137.0.7151.40/v8/src/builtins/builtins-typed-array.cc
---- chromium-137.0.7151.40/v8/src/builtins/builtins-typed-array.cc.me	2025-05-21 00:14:25.000000000 +0200
-+++ chromium-137.0.7151.40/v8/src/builtins/builtins-typed-array.cc	2025-05-23 23:31:59.961068931 +0200
-@@ -820,7 +820,7 @@ BUILTIN(Uint8ArrayPrototypeToBase64) {
- 
-     size_t simd_result_size;
-     if (uint8array->buffer()->is_shared()) {
--      simd_result_size = simdutf::atomic_binary_to_base64(
-+      simd_result_size = simdutf::binary_to_base64(
-           std::bit_cast<const char*>(uint8array->GetBuffer()->backing_store()),
-           length, reinterpret_cast<char*>(output->GetChars(no_gc)), alphabet);
-     } else {

diff --git a/chromium-138-checkversion-nodejs.patch b/chromium-138-checkversion-nodejs.patch
new file mode 100644
index 0000000..db389a7
--- /dev/null
+++ b/chromium-138-checkversion-nodejs.patch
@@ -0,0 +1,23 @@
+diff -up chromium-138.0.7204.15/third_party/node/node.gni.me chromium-138.0.7204.15/third_party/node/node.gni
+--- chromium-138.0.7204.15/third_party/node/node.gni.me	2025-06-03 23:20:09.000000000 +0200
++++ chromium-138.0.7204.15/third_party/node/node.gni	2025-06-09 09:59:21.788545159 +0200
+@@ -44,7 +44,6 @@ template("node") {
+       if (!defined(deps)) {
+         deps = []
+       }
+-      deps += [ "//third_party/node:check_version" ]
+     }
+   }
+ }
+diff -up chromium-138.0.7204.15/third_party/node/BUILD.gn.me chromium-138.0.7204.15/third_party/node/BUILD.gn
+diff -up chromium-138.0.7204.15/third_party/protobuf/proto_library.gni.me chromium-138.0.7204.15/third_party/protobuf/proto_library.gni
+--- chromium-138.0.7204.15/third_party/protobuf/proto_library.gni.me	2025-06-09 11:33:08.279645775 +0200
++++ chromium-138.0.7204.15/third_party/protobuf/proto_library.gni	2025-06-09 12:10:07.040045476 +0200
+@@ -562,7 +562,6 @@ template("proto_library") {
+                   _protoc_gen_ts_path,
+                   "//tools/protoc_wrapper/protoc-gen-ts_proto.py",
+                 ] + _protoc_gen_ts_runtime_deps
+-      deps += [ "//third_party/node:check_version" ]
+     }
+ 
+     if (_generate_with_plugin) {

diff --git a/chromium.spec b/chromium.spec
index 3e00d74..63953e2 100644
--- a/chromium.spec
+++ b/chromium.spec
@@ -201,7 +201,6 @@
 %endif
 %if 0%{?fedora}
 %global bundlecrc32c 0
-%global bundlesimdutf 0
 %endif
 %if 0%{?fedora} || 0%{?rhel} > 9
 %global bundlelibopenjpeg2 0
@@ -245,7 +244,7 @@
 %endif
 
 Name:	chromium
-Version: 137.0.7151.119
+Version: 138.0.7204.49
 Release: 1%{?dist}
 Summary: A WebKit (Blink) powered web browser that Google doesn't want you to use
 Url: http://www.chromium.org/Home
@@ -276,12 +275,7 @@ Patch90: chromium-121-system-libxml.patch
 Patch91: chromium-108-system-opus.patch
 
 # patch for Failed NodeJS version check
-Patch92: chromium-136-checkversion-nodejs.patch
-
-# Fix FTBFS
-Patch93: chromium-137-pdfium-png_decoder-build-error.patch
-Patch94: chromium-137-simdutf-build-error.patch
-Patch95: chromium-137-simdutf-7.x-build-error.patch
+Patch92: chromium-138-checkversion-nodejs.patch
 
 # system ffmpeg
 # need for old ffmpeg 5.x on epel9
@@ -801,7 +795,7 @@ Provides: bundled(boringssl)
 Provides: bundled(brotli) = 222564a95d9ab58865a096b8d9f7324ea5f2e03e
 %endif
 %if %{bundlesimdutf} 
-Provides: bundled(simdutf) = 6.4.0
+Provides: bundled(simdutf) = 7.0.0
 %endif
 Provides: bundled(bspatch) = 465265d0d473d107b76e74d969199eaf2cdc8750
 Provides: bundled(colorama) = 0.4.6
@@ -981,13 +975,6 @@ Qt6 UI for chromium.
 %endif
 
 %patch -P92 -p1 -b .nodejs-checkversion
-%patch -P93 -p1 -b .pdfium-png_decoder-build-error
-%if ! %{bundlesimdutf}
-%patch -P94 -p1 -b .simdutf-build-error
-%if 0%{?fedora} > 42
-%patch -P95 -p1 -b .simdutf-7.x-build-error
-%endif
-%endif
 
 %if ! %{bundleffmpegfree}
 %if 0%{?rhel} == 9
@@ -1752,6 +1739,12 @@ fi
 %endif
 
 %changelog
+* Tue Jun 24 2025 Than Ngo <than@redhat.com> - 138.0.7204.49-1
+- Update to 138.0.7204.49
+  * CVE-2025-6555: Use after free in Animation
+  * CVE-2025-6556: Insufficient policy enforcement in Loader
+  * CVE-2025-6557: Insufficient data validation in DevTools
+
 * Wed Jun 18 2025 Than Ngo <than@redhat.com> - 137.0.7151.119-1
 - Update to 137.0.7151.119
   * CVE-2025-6191: Integer overflow in V8

diff --git a/sources b/sources
index 07b5732..7f1da8b 100644
--- a/sources
+++ b/sources
@@ -1,2 +1,2 @@
 SHA512 (node-v22.14.0-stripped.tar.gz) = affddb541009c6d378049c2d7845b981335d9e7aa387efcc472e5efb621c345cd58ca69778a394f7e4ebbb5e4b1a115a389838ef1b6458ef5e98f2071b166e15
-SHA512 (chromium-137.0.7151.119-clean.tar.xz) = f30a176b8ff99b07b6c7743e3c4c72b2c176eb59ee60d827cd595ee473a42a0c2fafd02fb4c4f5d7332730df477564b2fbe35142a552422fabd7c1dc54a458ba
+SHA512 (chromium-138.0.7204.49-clean.tar.xz) = f96049dcb77dca50d005b22528064fb7a33673b70f2ccc1466acda9f34b41d88dc028e48ecfe0296a576d0535d0470fbc2ff064e57121e9a3e9d967ee5010a38

^ permalink raw reply related	[flat|nested] only message in thread

only message in thread, other threads:[~2026-08-07 16:07 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-08-07 16:07 [rpms/chromium] epel9-next: - Update to 138.0.7204.49 Than Ngo

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox