Skip to content

Revert "[X86][AVX10.2] Support YMM rounding new instructions (#101825)" #132362

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 1 commit into from
Mar 21, 2025
Merged

Revert "[X86][AVX10.2] Support YMM rounding new instructions (#101825)" #132362

merged 1 commit into from
Mar 21, 2025

Conversation

phoebewang
Copy link
Contributor

This reverts commit 0dba538.

YMM rounding was removed from AVX10 whitepaper. Ref: https://cdrdv2.intel.com/v1/dl/getContent/784343

The MINMAX and SATURATING CONVERT instructions will be removed as a follow up.

@phoebewang
Revert "[X86][AVX10.2] Support YMM rounding new instructions (llvm#10…
4b1276d 
…1825)"

This reverts commit 0dba538.

YMM rounding was removed from AVX10 whitepaper. Ref: https://cdrdv2.intel.com/v1/dl/getContent/784343

The MINMAX and SATURATING CONVERT instructions will be removed as a follow up.
@llvmbot llvmbot added clang Clang issues not falling into any other category backend:X86 clang:frontend Language frontend issues, e.g. anything involving "Sema" clang:headers Headers provided by Clang, e.g. for intrinsics clang:codegen IR generation bugs: mangling, exceptions, etc. mc Machine (object) code llvm:ir labels Mar 21, 2025
@llvmbot
Copy link
Member

llvmbot commented Mar 21, 2025 •
edited
Loading

@llvm/pr-subscribers-mc
@llvm/pr-subscribers-clang-codegen

@llvm/pr-subscribers-backend-x86

Author: Phoebe Wang (phoebewang)

Changes

This reverts commit 0dba538.

YMM rounding was removed from AVX10 whitepaper. Ref: https://cdrdv2.intel.com/v1/dl/getContent/784343

The MINMAX and SATURATING CONVERT instructions will be removed as a follow up.

Patch is 799.63 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/132362.diff

16 Files Affected:

  • (modified) clang/include/clang/Basic/BuiltinsX86.td (-126)
  • (modified) clang/lib/CodeGen/CGBuiltin.cpp (-120)
  • (modified) clang/lib/Headers/avx10_2niintrin.h (-1666)
  • (modified) clang/lib/Sema/SemaX86.cpp (-139)
  • (modified) clang/test/CodeGen/X86/avx10_2ni-builtins.c (-2405)
  • (modified) llvm/include/llvm/IR/IntrinsicsX86.td (-436)
  • (modified) llvm/lib/Target/X86/Disassembler/X86Disassembler.cpp (+1-1)
  • (modified) llvm/lib/Target/X86/X86InstrAVX10.td (-288)
  • (modified) llvm/lib/Target/X86/X86InstrAVX512.td (+4-4)
  • (modified) llvm/lib/Target/X86/X86InstrFMA3Info.cpp (+3-6)
  • (modified) llvm/lib/Target/X86/X86IntrinsicsInfo.h (-183)
  • (modified) llvm/test/CodeGen/X86/avx10_2ni-intrinsics.ll (-4393)
  • (modified) llvm/test/MC/Disassembler/X86/avx10_2ni-32.txt (-1778)
  • (modified) llvm/test/MC/Disassembler/X86/avx10_2ni-64.txt (-1778)
  • (modified) llvm/test/MC/X86/avx10_2ni-32-intel.s (-1778)
  • (modified) llvm/test/MC/X86/avx10_2ni-64-att.s (-1778)
diff --git a/clang/include/clang/Basic/BuiltinsX86.td b/clang/include/clang/Basic/BuiltinsX86.td
index e2252620bab56..ea0d6df4a33c2 100644
--- a/clang/include/clang/Basic/BuiltinsX86.td
+++ b/clang/include/clang/Basic/BuiltinsX86.td
@@ -4555,132 +4555,6 @@ let Features = "avx10.2-512", Attributes = [NoThrow, Const, RequiredVectorWidth<
   def mpsadbw512 : X86Builtin<"_Vector<32, short>(_Vector<64, char>, _Vector<64, char>, _Constant char)">;
 }
 
-let Features = "avx10.2-256", Attributes = [NoThrow, RequiredVectorWidth<256>] in {
-  def vaddpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vaddph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vaddps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-  def vcmppd256_round_mask : X86Builtin<"unsigned char(_Vector<4, double>, _Vector<4, double>, _Constant int, unsigned char, _Constant int)">;
-  def vcmpph256_round_mask : X86Builtin<"unsigned short(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int, unsigned short, _Constant int)">;
-  def vcmpps256_round_mask : X86Builtin<"unsigned char(_Vector<8, float>, _Vector<8, float>, _Constant int, unsigned char, _Constant int)">;
-  def vcvtdq2ph256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<8, int>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtdq2ps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, int>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vcvtpd2dq256_round_mask : X86Builtin<"_Vector<4, int>(_Vector<4, double>, _Vector<4, int>, unsigned char, _Constant int)">;
-  def vcvtpd2ph256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<4, double>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtpd2ps256_round_mask : X86Builtin<"_Vector<4, float>(_Vector<4, double>, _Vector<4, float>, unsigned char, _Constant int)">;
-  def vcvtpd2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<4, double>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvtpd2udq256_round_mask : X86Builtin<"_Vector<4, unsigned int>(_Vector<4, double>, _Vector<4, unsigned int>, unsigned char, _Constant int)">;
-  def vcvtpd2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<4, double>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvtph2dq256_round_mask : X86Builtin<"_Vector<8, int>(_Vector<8, _Float16>, _Vector<8, int>, unsigned char, _Constant int)">;
-  def vcvtph2pd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<8, _Float16>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vcvtph2psx256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, _Float16>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vcvtph2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<8, _Float16>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvtph2udq256_round_mask : X86Builtin<"_Vector<8, unsigned int>(_Vector<8, _Float16>, _Vector<8, unsigned int>, unsigned char, _Constant int)">;
-  def vcvtph2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<8, _Float16>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvtph2uw256_round_mask : X86Builtin<"_Vector<16, unsigned short>(_Vector<16, _Float16>, _Vector<16, unsigned short>, unsigned short, _Constant int)">;
-  def vcvtph2w256_round_mask : X86Builtin<"_Vector<16, short>(_Vector<16, _Float16>, _Vector<16, short>, unsigned short, _Constant int)">;
-  def vcvtps2dq256_round_mask : X86Builtin<"_Vector<8, int>(_Vector<8, float>, _Vector<8, int>, unsigned char, _Constant int)">;
-  def vcvtps2pd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, float>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vcvtps2phx256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<8, float>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtps2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<4, float>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvtps2udq256_round_mask : X86Builtin<"_Vector<8, unsigned int>(_Vector<8, float>, _Vector<8, unsigned int>, unsigned char, _Constant int)">;
-  def vcvtps2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<4, float>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvtqq2pd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, long long int>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vcvtqq2ph256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<4, long long int>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtqq2ps256_round_mask : X86Builtin<"_Vector<4, float>(_Vector<4, long long int>, _Vector<4, float>, unsigned char, _Constant int)">;
-  def vcvttpd2dq256_round_mask : X86Builtin<"_Vector<4, int>(_Vector<4, double>, _Vector<4, int>, unsigned char, _Constant int)">;
-  def vcvttpd2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<4, double>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvttpd2udq256_round_mask : X86Builtin<"_Vector<4, unsigned int>(_Vector<4, double>, _Vector<4, unsigned int>, unsigned char, _Constant int)">;
-  def vcvttpd2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<4, double>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvttph2dq256_round_mask : X86Builtin<"_Vector<8, int>(_Vector<8, _Float16>, _Vector<8, int>, unsigned char, _Constant int)">;
-  def vcvttph2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<8, _Float16>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvttph2udq256_round_mask : X86Builtin<"_Vector<8, unsigned int>(_Vector<8, _Float16>, _Vector<8, unsigned int>, unsigned char, _Constant int)">;
-  def vcvttph2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<8, _Float16>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvttph2uw256_round_mask : X86Builtin<"_Vector<16, unsigned short>(_Vector<16, _Float16>, _Vector<16, unsigned short>, unsigned short, _Constant int)">;
-  def vcvttph2w256_round_mask : X86Builtin<"_Vector<16, short>(_Vector<16, _Float16>, _Vector<16, short>, unsigned short, _Constant int)">;
-  def vcvttps2dq256_round_mask : X86Builtin<"_Vector<8, int>(_Vector<8, float>, _Vector<8, int>, unsigned char, _Constant int)">;
-  def vcvttps2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<4, float>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvttps2udq256_round_mask : X86Builtin<"_Vector<8, unsigned int>(_Vector<8, float>, _Vector<8, unsigned int>, unsigned char, _Constant int)">;
-  def vcvttps2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<4, float>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvtudq2ph256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<8, unsigned int>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtudq2ps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, unsigned int>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vcvtuqq2pd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, unsigned long long int>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vcvtuqq2ph256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<4, unsigned long long int>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtuqq2ps256_round_mask : X86Builtin<"_Vector<4, float>(_Vector<4, unsigned long long int>, _Vector<4, float>, unsigned char, _Constant int)">;
-  def vcvtuw2ph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, unsigned short>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vcvtw2ph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, short>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vdivpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vdivph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vdivps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-  def vfcmaddcph256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfcmaddcph256_round_maskz : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfcmaddcph256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfcmulcph256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfixupimmpd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, long long int>, _Constant int, unsigned char, _Constant int)">;
-  def vfixupimmpd256_round_maskz : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, long long int>, _Constant int, unsigned char, _Constant int)">;
-  def vfixupimmps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, int>, _Constant int, unsigned char, _Constant int)">;
-  def vfixupimmps256_round_maskz : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, int>, _Constant int, unsigned char, _Constant int)">;
-  def vfmaddpd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddpd256_round_maskz : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddpd256_round_mask3 : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddph256_round_maskz : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddph256_round_mask3 : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddps256_round_maskz : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddps256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddcph256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddcph256_round_maskz : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddcph256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddsubpd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddsubpd256_round_maskz : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddsubpd256_round_mask3 : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddsubph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddsubph256_round_maskz : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddsubph256_round_mask3 : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddsubps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddsubps256_round_maskz : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddsubps256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmsubpd256_round_mask3 : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmsubph256_round_mask3 : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmsubps256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmsubaddpd256_round_mask3 : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmsubaddph256_round_mask3 : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmsubaddps256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmulcph256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vgetexppd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vgetexpph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vgetexpps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vgetmantpd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Constant int, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vgetmantph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Constant int, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vgetmantps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Constant int, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vmaxpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vmaxph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vmaxps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-  def vminpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vminph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vminps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-  def vmulpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vmulph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vmulps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-  def vrangepd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vrangeps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vreducepd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Constant int, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vreduceph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Constant int, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vreduceps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Constant int, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vrndscalepd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Constant int, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vrndscaleph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Constant int, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vrndscaleps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Constant int, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vscalefpd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vscalefph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vscalefps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vsqrtpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Constant int)">;
-  def vsqrtph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Constant int)">;
-  def vsqrtps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Constant int)">;
-  def vsubpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vsubph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vsubps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-}
-
 let Features = "avxvnniint16|avx10.2-256", Attributes = [NoThrow, RequiredVectorWidth<128>] in {
   def vpdpwsud128 : X86Builtin<"_Vector<4, int>(_Vector<4, int>, _Vector<4, int>, _Vector<4, int>)">;
 }
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 3536bdd91ebc6..6c30d5c7c85a5 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -14801,54 +14801,6 @@ static Value *EmitX86FMAExpr(CodeGenFunction &CGF, const CallExpr *E,
   case clang::X86::BI__builtin_ia32_vfmaddsubpd512_mask3:
     IID = Intrinsic::x86_avx512_vfmaddsub_pd_512;
     break;
-  case clang::X86::BI__builtin_ia32_vfmsubph256_round_mask3:
-    Subtract = true;
-    LLVM_FALLTHROUGH;
-  case clang::X86::BI__builtin_ia32_vfmaddph256_round_mask:
-  case clang::X86::BI__builtin_ia32_vfmaddph256_round_maskz:
-  case clang::X86::BI__builtin_ia32_vfmaddph256_round_mask3:
-    IID = Intrinsic::x86_avx10_vfmaddph256;
-    break;
-  case clang::X86::BI__builtin_ia32_vfmsubaddph256_round_mask3:
-    Subtract = true;
-    LLVM_FALLTHROUGH;
-  case clang::X86::BI__builtin_ia32_vfmaddsubph256_round_mask:
-  case clang::X86::BI__builtin_ia32_vfmaddsubph256_round_maskz:
-  case clang::X86::BI__builtin_ia32_vfmaddsubph256_round_mask3:
-    IID = Intrinsic::x86_avx10_vfmaddsubph256;
-    break;
-  case clang::X86::BI__builtin_ia32_vfmsubps256_round_mask3:
-    Subtract = true;
-    LLVM_FALLTHROUGH;
-  case clang::X86::BI__builtin_ia32_vfmaddps256_round_mask:
-  case clang::X86::BI__builtin_ia32_vfmaddps256_round_maskz:
-  case clang::X86::BI__builtin_ia32_vfmaddps256_round_mask3:
-    IID = Intrinsic::x86_avx10_vfmaddps256;
-    break;
-  case clang::X86::BI__builtin_ia32_vfmsubpd256_round_mask3:
-    Subtract = true;
-    LLVM_FALLTHROUGH;
-  case clang::X86::BI__builtin_ia32_vfmaddpd256_round_mask:
-  case clang::X86::BI__bu...
[truncated]

@llvmbot
Copy link
Member

llvmbot commented Mar 21, 2025

@llvm/pr-subscribers-llvm-ir

Author: Phoebe Wang (phoebewang)

Changes

This reverts commit 0dba538.

YMM rounding was removed from AVX10 whitepaper. Ref: https://cdrdv2.intel.com/v1/dl/getContent/784343

The MINMAX and SATURATING CONVERT instructions will be removed as a follow up.

Patch is 799.63 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/132362.diff

16 Files Affected:

  • (modified) clang/include/clang/Basic/BuiltinsX86.td (-126)
  • (modified) clang/lib/CodeGen/CGBuiltin.cpp (-120)
  • (modified) clang/lib/Headers/avx10_2niintrin.h (-1666)
  • (modified) clang/lib/Sema/SemaX86.cpp (-139)
  • (modified) clang/test/CodeGen/X86/avx10_2ni-builtins.c (-2405)
  • (modified) llvm/include/llvm/IR/IntrinsicsX86.td (-436)
  • (modified) llvm/lib/Target/X86/Disassembler/X86Disassembler.cpp (+1-1)
  • (modified) llvm/lib/Target/X86/X86InstrAVX10.td (-288)
  • (modified) llvm/lib/Target/X86/X86InstrAVX512.td (+4-4)
  • (modified) llvm/lib/Target/X86/X86InstrFMA3Info.cpp (+3-6)
  • (modified) llvm/lib/Target/X86/X86IntrinsicsInfo.h (-183)
  • (modified) llvm/test/CodeGen/X86/avx10_2ni-intrinsics.ll (-4393)
  • (modified) llvm/test/MC/Disassembler/X86/avx10_2ni-32.txt (-1778)
  • (modified) llvm/test/MC/Disassembler/X86/avx10_2ni-64.txt (-1778)
  • (modified) llvm/test/MC/X86/avx10_2ni-32-intel.s (-1778)
  • (modified) llvm/test/MC/X86/avx10_2ni-64-att.s (-1778)
diff --git a/clang/include/clang/Basic/BuiltinsX86.td b/clang/include/clang/Basic/BuiltinsX86.td
index e2252620bab56..ea0d6df4a33c2 100644
--- a/clang/include/clang/Basic/BuiltinsX86.td
+++ b/clang/include/clang/Basic/BuiltinsX86.td
@@ -4555,132 +4555,6 @@ let Features = "avx10.2-512", Attributes = [NoThrow, Const, RequiredVectorWidth<
   def mpsadbw512 : X86Builtin<"_Vector<32, short>(_Vector<64, char>, _Vector<64, char>, _Constant char)">;
 }
 
-let Features = "avx10.2-256", Attributes = [NoThrow, RequiredVectorWidth<256>] in {
-  def vaddpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vaddph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vaddps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-  def vcmppd256_round_mask : X86Builtin<"unsigned char(_Vector<4, double>, _Vector<4, double>, _Constant int, unsigned char, _Constant int)">;
-  def vcmpph256_round_mask : X86Builtin<"unsigned short(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int, unsigned short, _Constant int)">;
-  def vcmpps256_round_mask : X86Builtin<"unsigned char(_Vector<8, float>, _Vector<8, float>, _Constant int, unsigned char, _Constant int)">;
-  def vcvtdq2ph256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<8, int>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtdq2ps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, int>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vcvtpd2dq256_round_mask : X86Builtin<"_Vector<4, int>(_Vector<4, double>, _Vector<4, int>, unsigned char, _Constant int)">;
-  def vcvtpd2ph256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<4, double>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtpd2ps256_round_mask : X86Builtin<"_Vector<4, float>(_Vector<4, double>, _Vector<4, float>, unsigned char, _Constant int)">;
-  def vcvtpd2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<4, double>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvtpd2udq256_round_mask : X86Builtin<"_Vector<4, unsigned int>(_Vector<4, double>, _Vector<4, unsigned int>, unsigned char, _Constant int)">;
-  def vcvtpd2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<4, double>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvtph2dq256_round_mask : X86Builtin<"_Vector<8, int>(_Vector<8, _Float16>, _Vector<8, int>, unsigned char, _Constant int)">;
-  def vcvtph2pd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<8, _Float16>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vcvtph2psx256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, _Float16>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vcvtph2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<8, _Float16>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvtph2udq256_round_mask : X86Builtin<"_Vector<8, unsigned int>(_Vector<8, _Float16>, _Vector<8, unsigned int>, unsigned char, _Constant int)">;
-  def vcvtph2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<8, _Float16>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvtph2uw256_round_mask : X86Builtin<"_Vector<16, unsigned short>(_Vector<16, _Float16>, _Vector<16, unsigned short>, unsigned short, _Constant int)">;
-  def vcvtph2w256_round_mask : X86Builtin<"_Vector<16, short>(_Vector<16, _Float16>, _Vector<16, short>, unsigned short, _Constant int)">;
-  def vcvtps2dq256_round_mask : X86Builtin<"_Vector<8, int>(_Vector<8, float>, _Vector<8, int>, unsigned char, _Constant int)">;
-  def vcvtps2pd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, float>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vcvtps2phx256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<8, float>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtps2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<4, float>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvtps2udq256_round_mask : X86Builtin<"_Vector<8, unsigned int>(_Vector<8, float>, _Vector<8, unsigned int>, unsigned char, _Constant int)">;
-  def vcvtps2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<4, float>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvtqq2pd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, long long int>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vcvtqq2ph256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<4, long long int>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtqq2ps256_round_mask : X86Builtin<"_Vector<4, float>(_Vector<4, long long int>, _Vector<4, float>, unsigned char, _Constant int)">;
-  def vcvttpd2dq256_round_mask : X86Builtin<"_Vector<4, int>(_Vector<4, double>, _Vector<4, int>, unsigned char, _Constant int)">;
-  def vcvttpd2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<4, double>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvttpd2udq256_round_mask : X86Builtin<"_Vector<4, unsigned int>(_Vector<4, double>, _Vector<4, unsigned int>, unsigned char, _Constant int)">;
-  def vcvttpd2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<4, double>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvttph2dq256_round_mask : X86Builtin<"_Vector<8, int>(_Vector<8, _Float16>, _Vector<8, int>, unsigned char, _Constant int)">;
-  def vcvttph2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<8, _Float16>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvttph2udq256_round_mask : X86Builtin<"_Vector<8, unsigned int>(_Vector<8, _Float16>, _Vector<8, unsigned int>, unsigned char, _Constant int)">;
-  def vcvttph2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<8, _Float16>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvttph2uw256_round_mask : X86Builtin<"_Vector<16, unsigned short>(_Vector<16, _Float16>, _Vector<16, unsigned short>, unsigned short, _Constant int)">;
-  def vcvttph2w256_round_mask : X86Builtin<"_Vector<16, short>(_Vector<16, _Float16>, _Vector<16, short>, unsigned short, _Constant int)">;
-  def vcvttps2dq256_round_mask : X86Builtin<"_Vector<8, int>(_Vector<8, float>, _Vector<8, int>, unsigned char, _Constant int)">;
-  def vcvttps2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<4, float>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvttps2udq256_round_mask : X86Builtin<"_Vector<8, unsigned int>(_Vector<8, float>, _Vector<8, unsigned int>, unsigned char, _Constant int)">;
-  def vcvttps2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<4, float>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvtudq2ph256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<8, unsigned int>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtudq2ps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, unsigned int>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vcvtuqq2pd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, unsigned long long int>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vcvtuqq2ph256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<4, unsigned long long int>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtuqq2ps256_round_mask : X86Builtin<"_Vector<4, float>(_Vector<4, unsigned long long int>, _Vector<4, float>, unsigned char, _Constant int)">;
-  def vcvtuw2ph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, unsigned short>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vcvtw2ph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, short>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vdivpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vdivph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vdivps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-  def vfcmaddcph256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfcmaddcph256_round_maskz : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfcmaddcph256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfcmulcph256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfixupimmpd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, long long int>, _Constant int, unsigned char, _Constant int)">;
-  def vfixupimmpd256_round_maskz : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, long long int>, _Constant int, unsigned char, _Constant int)">;
-  def vfixupimmps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, int>, _Constant int, unsigned char, _Constant int)">;
-  def vfixupimmps256_round_maskz : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, int>, _Constant int, unsigned char, _Constant int)">;
-  def vfmaddpd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddpd256_round_maskz : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddpd256_round_mask3 : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddph256_round_maskz : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddph256_round_mask3 : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddps256_round_maskz : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddps256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddcph256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddcph256_round_maskz : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddcph256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddsubpd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddsubpd256_round_maskz : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddsubpd256_round_mask3 : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddsubph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddsubph256_round_maskz : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddsubph256_round_mask3 : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddsubps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddsubps256_round_maskz : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddsubps256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmsubpd256_round_mask3 : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmsubph256_round_mask3 : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmsubps256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmsubaddpd256_round_mask3 : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmsubaddph256_round_mask3 : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmsubaddps256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmulcph256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vgetexppd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vgetexpph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vgetexpps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vgetmantpd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Constant int, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vgetmantph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Constant int, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vgetmantps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Constant int, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vmaxpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vmaxph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vmaxps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-  def vminpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vminph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vminps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-  def vmulpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vmulph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vmulps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-  def vrangepd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vrangeps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vreducepd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Constant int, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vreduceph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Constant int, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vreduceps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Constant int, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vrndscalepd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Constant int, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vrndscaleph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Constant int, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vrndscaleps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Constant int, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vscalefpd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vscalefph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vscalefps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vsqrtpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Constant int)">;
-  def vsqrtph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Constant int)">;
-  def vsqrtps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Constant int)">;
-  def vsubpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vsubph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vsubps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-}
-
 let Features = "avxvnniint16|avx10.2-256", Attributes = [NoThrow, RequiredVectorWidth<128>] in {
   def vpdpwsud128 : X86Builtin<"_Vector<4, int>(_Vector<4, int>, _Vector<4, int>, _Vector<4, int>)">;
 }
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 3536bdd91ebc6..6c30d5c7c85a5 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -14801,54 +14801,6 @@ static Value *EmitX86FMAExpr(CodeGenFunction &CGF, const CallExpr *E,
   case clang::X86::BI__builtin_ia32_vfmaddsubpd512_mask3:
     IID = Intrinsic::x86_avx512_vfmaddsub_pd_512;
     break;
-  case clang::X86::BI__builtin_ia32_vfmsubph256_round_mask3:
-    Subtract = true;
-    LLVM_FALLTHROUGH;
-  case clang::X86::BI__builtin_ia32_vfmaddph256_round_mask:
-  case clang::X86::BI__builtin_ia32_vfmaddph256_round_maskz:
-  case clang::X86::BI__builtin_ia32_vfmaddph256_round_mask3:
-    IID = Intrinsic::x86_avx10_vfmaddph256;
-    break;
-  case clang::X86::BI__builtin_ia32_vfmsubaddph256_round_mask3:
-    Subtract = true;
-    LLVM_FALLTHROUGH;
-  case clang::X86::BI__builtin_ia32_vfmaddsubph256_round_mask:
-  case clang::X86::BI__builtin_ia32_vfmaddsubph256_round_maskz:
-  case clang::X86::BI__builtin_ia32_vfmaddsubph256_round_mask3:
-    IID = Intrinsic::x86_avx10_vfmaddsubph256;
-    break;
-  case clang::X86::BI__builtin_ia32_vfmsubps256_round_mask3:
-    Subtract = true;
-    LLVM_FALLTHROUGH;
-  case clang::X86::BI__builtin_ia32_vfmaddps256_round_mask:
-  case clang::X86::BI__builtin_ia32_vfmaddps256_round_maskz:
-  case clang::X86::BI__builtin_ia32_vfmaddps256_round_mask3:
-    IID = Intrinsic::x86_avx10_vfmaddps256;
-    break;
-  case clang::X86::BI__builtin_ia32_vfmsubpd256_round_mask3:
-    Subtract = true;
-    LLVM_FALLTHROUGH;
-  case clang::X86::BI__builtin_ia32_vfmaddpd256_round_mask:
-  case clang::X86::BI__bu...
[truncated]

@llvmbot
Copy link
Member

llvmbot commented Mar 21, 2025

@llvm/pr-subscribers-clang

Author: Phoebe Wang (phoebewang)

Changes

This reverts commit 0dba538.

YMM rounding was removed from AVX10 whitepaper. Ref: https://cdrdv2.intel.com/v1/dl/getContent/784343

The MINMAX and SATURATING CONVERT instructions will be removed as a follow up.

Patch is 799.63 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/132362.diff

16 Files Affected:

  • (modified) clang/include/clang/Basic/BuiltinsX86.td (-126)
  • (modified) clang/lib/CodeGen/CGBuiltin.cpp (-120)
  • (modified) clang/lib/Headers/avx10_2niintrin.h (-1666)
  • (modified) clang/lib/Sema/SemaX86.cpp (-139)
  • (modified) clang/test/CodeGen/X86/avx10_2ni-builtins.c (-2405)
  • (modified) llvm/include/llvm/IR/IntrinsicsX86.td (-436)
  • (modified) llvm/lib/Target/X86/Disassembler/X86Disassembler.cpp (+1-1)
  • (modified) llvm/lib/Target/X86/X86InstrAVX10.td (-288)
  • (modified) llvm/lib/Target/X86/X86InstrAVX512.td (+4-4)
  • (modified) llvm/lib/Target/X86/X86InstrFMA3Info.cpp (+3-6)
  • (modified) llvm/lib/Target/X86/X86IntrinsicsInfo.h (-183)
  • (modified) llvm/test/CodeGen/X86/avx10_2ni-intrinsics.ll (-4393)
  • (modified) llvm/test/MC/Disassembler/X86/avx10_2ni-32.txt (-1778)
  • (modified) llvm/test/MC/Disassembler/X86/avx10_2ni-64.txt (-1778)
  • (modified) llvm/test/MC/X86/avx10_2ni-32-intel.s (-1778)
  • (modified) llvm/test/MC/X86/avx10_2ni-64-att.s (-1778)
diff --git a/clang/include/clang/Basic/BuiltinsX86.td b/clang/include/clang/Basic/BuiltinsX86.td
index e2252620bab56..ea0d6df4a33c2 100644
--- a/clang/include/clang/Basic/BuiltinsX86.td
+++ b/clang/include/clang/Basic/BuiltinsX86.td
@@ -4555,132 +4555,6 @@ let Features = "avx10.2-512", Attributes = [NoThrow, Const, RequiredVectorWidth<
   def mpsadbw512 : X86Builtin<"_Vector<32, short>(_Vector<64, char>, _Vector<64, char>, _Constant char)">;
 }
 
-let Features = "avx10.2-256", Attributes = [NoThrow, RequiredVectorWidth<256>] in {
-  def vaddpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vaddph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vaddps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-  def vcmppd256_round_mask : X86Builtin<"unsigned char(_Vector<4, double>, _Vector<4, double>, _Constant int, unsigned char, _Constant int)">;
-  def vcmpph256_round_mask : X86Builtin<"unsigned short(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int, unsigned short, _Constant int)">;
-  def vcmpps256_round_mask : X86Builtin<"unsigned char(_Vector<8, float>, _Vector<8, float>, _Constant int, unsigned char, _Constant int)">;
-  def vcvtdq2ph256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<8, int>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtdq2ps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, int>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vcvtpd2dq256_round_mask : X86Builtin<"_Vector<4, int>(_Vector<4, double>, _Vector<4, int>, unsigned char, _Constant int)">;
-  def vcvtpd2ph256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<4, double>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtpd2ps256_round_mask : X86Builtin<"_Vector<4, float>(_Vector<4, double>, _Vector<4, float>, unsigned char, _Constant int)">;
-  def vcvtpd2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<4, double>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvtpd2udq256_round_mask : X86Builtin<"_Vector<4, unsigned int>(_Vector<4, double>, _Vector<4, unsigned int>, unsigned char, _Constant int)">;
-  def vcvtpd2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<4, double>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvtph2dq256_round_mask : X86Builtin<"_Vector<8, int>(_Vector<8, _Float16>, _Vector<8, int>, unsigned char, _Constant int)">;
-  def vcvtph2pd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<8, _Float16>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vcvtph2psx256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, _Float16>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vcvtph2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<8, _Float16>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvtph2udq256_round_mask : X86Builtin<"_Vector<8, unsigned int>(_Vector<8, _Float16>, _Vector<8, unsigned int>, unsigned char, _Constant int)">;
-  def vcvtph2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<8, _Float16>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvtph2uw256_round_mask : X86Builtin<"_Vector<16, unsigned short>(_Vector<16, _Float16>, _Vector<16, unsigned short>, unsigned short, _Constant int)">;
-  def vcvtph2w256_round_mask : X86Builtin<"_Vector<16, short>(_Vector<16, _Float16>, _Vector<16, short>, unsigned short, _Constant int)">;
-  def vcvtps2dq256_round_mask : X86Builtin<"_Vector<8, int>(_Vector<8, float>, _Vector<8, int>, unsigned char, _Constant int)">;
-  def vcvtps2pd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, float>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vcvtps2phx256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<8, float>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtps2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<4, float>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvtps2udq256_round_mask : X86Builtin<"_Vector<8, unsigned int>(_Vector<8, float>, _Vector<8, unsigned int>, unsigned char, _Constant int)">;
-  def vcvtps2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<4, float>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvtqq2pd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, long long int>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vcvtqq2ph256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<4, long long int>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtqq2ps256_round_mask : X86Builtin<"_Vector<4, float>(_Vector<4, long long int>, _Vector<4, float>, unsigned char, _Constant int)">;
-  def vcvttpd2dq256_round_mask : X86Builtin<"_Vector<4, int>(_Vector<4, double>, _Vector<4, int>, unsigned char, _Constant int)">;
-  def vcvttpd2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<4, double>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvttpd2udq256_round_mask : X86Builtin<"_Vector<4, unsigned int>(_Vector<4, double>, _Vector<4, unsigned int>, unsigned char, _Constant int)">;
-  def vcvttpd2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<4, double>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvttph2dq256_round_mask : X86Builtin<"_Vector<8, int>(_Vector<8, _Float16>, _Vector<8, int>, unsigned char, _Constant int)">;
-  def vcvttph2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<8, _Float16>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvttph2udq256_round_mask : X86Builtin<"_Vector<8, unsigned int>(_Vector<8, _Float16>, _Vector<8, unsigned int>, unsigned char, _Constant int)">;
-  def vcvttph2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<8, _Float16>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvttph2uw256_round_mask : X86Builtin<"_Vector<16, unsigned short>(_Vector<16, _Float16>, _Vector<16, unsigned short>, unsigned short, _Constant int)">;
-  def vcvttph2w256_round_mask : X86Builtin<"_Vector<16, short>(_Vector<16, _Float16>, _Vector<16, short>, unsigned short, _Constant int)">;
-  def vcvttps2dq256_round_mask : X86Builtin<"_Vector<8, int>(_Vector<8, float>, _Vector<8, int>, unsigned char, _Constant int)">;
-  def vcvttps2qq256_round_mask : X86Builtin<"_Vector<4, long long int>(_Vector<4, float>, _Vector<4, long long int>, unsigned char, _Constant int)">;
-  def vcvttps2udq256_round_mask : X86Builtin<"_Vector<8, unsigned int>(_Vector<8, float>, _Vector<8, unsigned int>, unsigned char, _Constant int)">;
-  def vcvttps2uqq256_round_mask : X86Builtin<"_Vector<4, unsigned long long int>(_Vector<4, float>, _Vector<4, unsigned long long int>, unsigned char, _Constant int)">;
-  def vcvtudq2ph256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<8, unsigned int>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtudq2ps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, unsigned int>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vcvtuqq2pd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, unsigned long long int>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vcvtuqq2ph256_round_mask : X86Builtin<"_Vector<8, _Float16>(_Vector<4, unsigned long long int>, _Vector<8, _Float16>, unsigned char, _Constant int)">;
-  def vcvtuqq2ps256_round_mask : X86Builtin<"_Vector<4, float>(_Vector<4, unsigned long long int>, _Vector<4, float>, unsigned char, _Constant int)">;
-  def vcvtuw2ph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, unsigned short>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vcvtw2ph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, short>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vdivpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vdivph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vdivps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-  def vfcmaddcph256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfcmaddcph256_round_maskz : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfcmaddcph256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfcmulcph256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfixupimmpd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, long long int>, _Constant int, unsigned char, _Constant int)">;
-  def vfixupimmpd256_round_maskz : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, long long int>, _Constant int, unsigned char, _Constant int)">;
-  def vfixupimmps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, int>, _Constant int, unsigned char, _Constant int)">;
-  def vfixupimmps256_round_maskz : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, int>, _Constant int, unsigned char, _Constant int)">;
-  def vfmaddpd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddpd256_round_maskz : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddpd256_round_mask3 : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddph256_round_maskz : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddph256_round_mask3 : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddps256_round_maskz : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddps256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddcph256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddcph256_round_maskz : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddcph256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddsubpd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddsubpd256_round_maskz : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddsubpd256_round_mask3 : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmaddsubph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddsubph256_round_maskz : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddsubph256_round_mask3 : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmaddsubps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddsubps256_round_maskz : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmaddsubps256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmsubpd256_round_mask3 : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmsubph256_round_mask3 : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmsubps256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmsubaddpd256_round_mask3 : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vfmsubaddph256_round_mask3 : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vfmsubaddps256_round_mask3 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vfmulcph256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vgetexppd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vgetexpph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vgetexpps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vgetmantpd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Constant int, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vgetmantph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Constant int, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vgetmantps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Constant int, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vmaxpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vmaxph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vmaxps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-  def vminpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vminph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vminps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-  def vmulpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vmulph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vmulps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-  def vrangepd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vrangeps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vreducepd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Constant int, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vreduceph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Constant int, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vreduceps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Constant int, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vrndscalepd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Constant int, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vrndscaleph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Constant int, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vrndscaleps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Constant int, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vscalefpd256_round_mask : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Vector<4, double>, unsigned char, _Constant int)">;
-  def vscalefph256_round_mask : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Vector<16, _Float16>, unsigned short, _Constant int)">;
-  def vscalefps256_round_mask : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Vector<8, float>, unsigned char, _Constant int)">;
-  def vsqrtpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Constant int)">;
-  def vsqrtph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Constant int)">;
-  def vsqrtps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Constant int)">;
-  def vsubpd256_round : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>, _Constant int)">;
-  def vsubph256_round : X86Builtin<"_Vector<16, _Float16>(_Vector<16, _Float16>, _Vector<16, _Float16>, _Constant int)">;
-  def vsubps256_round : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>, _Constant int)">;
-}
-
 let Features = "avxvnniint16|avx10.2-256", Attributes = [NoThrow, RequiredVectorWidth<128>] in {
   def vpdpwsud128 : X86Builtin<"_Vector<4, int>(_Vector<4, int>, _Vector<4, int>, _Vector<4, int>)">;
 }
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 3536bdd91ebc6..6c30d5c7c85a5 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -14801,54 +14801,6 @@ static Value *EmitX86FMAExpr(CodeGenFunction &CGF, const CallExpr *E,
   case clang::X86::BI__builtin_ia32_vfmaddsubpd512_mask3:
     IID = Intrinsic::x86_avx512_vfmaddsub_pd_512;
     break;
-  case clang::X86::BI__builtin_ia32_vfmsubph256_round_mask3:
-    Subtract = true;
-    LLVM_FALLTHROUGH;
-  case clang::X86::BI__builtin_ia32_vfmaddph256_round_mask:
-  case clang::X86::BI__builtin_ia32_vfmaddph256_round_maskz:
-  case clang::X86::BI__builtin_ia32_vfmaddph256_round_mask3:
-    IID = Intrinsic::x86_avx10_vfmaddph256;
-    break;
-  case clang::X86::BI__builtin_ia32_vfmsubaddph256_round_mask3:
-    Subtract = true;
-    LLVM_FALLTHROUGH;
-  case clang::X86::BI__builtin_ia32_vfmaddsubph256_round_mask:
-  case clang::X86::BI__builtin_ia32_vfmaddsubph256_round_maskz:
-  case clang::X86::BI__builtin_ia32_vfmaddsubph256_round_mask3:
-    IID = Intrinsic::x86_avx10_vfmaddsubph256;
-    break;
-  case clang::X86::BI__builtin_ia32_vfmsubps256_round_mask3:
-    Subtract = true;
-    LLVM_FALLTHROUGH;
-  case clang::X86::BI__builtin_ia32_vfmaddps256_round_mask:
-  case clang::X86::BI__builtin_ia32_vfmaddps256_round_maskz:
-  case clang::X86::BI__builtin_ia32_vfmaddps256_round_mask3:
-    IID = Intrinsic::x86_avx10_vfmaddps256;
-    break;
-  case clang::X86::BI__builtin_ia32_vfmsubpd256_round_mask3:
-    Subtract = true;
-    LLVM_FALLTHROUGH;
-  case clang::X86::BI__builtin_ia32_vfmaddpd256_round_mask:
-  case clang::X86::BI__bu...
[truncated]

@github-actions GitHub Actions
Copy link

⚠️ C/C++ code formatter, clang-format found issues in your code. ⚠️

You can test this locally with the following command: 
git-clang-format --diff be258a2c2de0417716800fd07ef91572fbf20fce 4b1276dc87a2d5888e14db01011716f3db1e93b6 --extensions h,c,cpp -- clang/lib/CodeGen/CGBuiltin.cpp clang/lib/Headers/avx10_2niintrin.h clang/lib/Sema/SemaX86.cpp clang/test/CodeGen/X86/avx10_2ni-builtins.c llvm/lib/Target/X86/Disassembler/X86Disassembler.cpp llvm/lib/Target/X86/X86InstrFMA3Info.cpp llvm/lib/Target/X86/X86IntrinsicsInfo.h
View the diff from clang-format here. 
diff --git a/llvm/lib/Target/X86/X86InstrFMA3Info.cpp b/llvm/lib/Target/X86/X86InstrFMA3Info.cpp
index 9947fc24b5..f25954d037 100644
--- a/llvm/lib/Target/X86/X86InstrFMA3Info.cpp
+++ b/llvm/lib/Target/X86/X86InstrFMA3Info.cpp
@@ -103,9 +103,9 @@ static const X86InstrFMA3Group Groups[] = {
   FMA3GROUP_PACKED_AVX512_WIDTHS(Name, PH, Suf, Attrs) \
   FMA3GROUP_PACKED_AVX512_WIDTHS(Name, PS, Suf, Attrs)
 
-#define FMA3GROUP_PACKED_AVX512_ROUND(Name, Suf, Attrs) \
-  FMA3GROUP_MASKED(Name, PDZ##Suf, Attrs) \
-  FMA3GROUP_MASKED(Name, PHZ##Suf, Attrs) \
+#define FMA3GROUP_PACKED_AVX512_ROUND(Name, Suf, Attrs)                        \
+  FMA3GROUP_MASKED(Name, PDZ##Suf, Attrs)                                      \
+  FMA3GROUP_MASKED(Name, PHZ##Suf, Attrs)                                      \
   FMA3GROUP_MASKED(Name, PSZ##Suf, Attrs)
 
 #define FMA3GROUP_SCALAR_AVX512_ROUND(Name, Suf, Attrs) \

RKSimon
RKSimon approved these changes Mar 21, 2025
View reviewed changes
Copy link
Collaborator

@RKSimon RKSimon left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM - do we have to remove avx10-256 as well (and just use prefer vector width) or will that remain in some form?

@phoebewang
Copy link
Contributor Author

LGTM - do we have to remove avx10-256 as well (and just use prefer vector width) or will that remain in some form?

Yes, we will remove avx10-256 and evex512 finally.

@phoebewang phoebewang merged commit 09feaa9 into llvm:main Mar 21, 2025
18 of 19 checks passed
jthackray added a commit to jthackray/llvm-project that referenced this pull request Mar 21, 2025
@jthackray
fixup! Merge in changes from llvm#132362
42645ef
@jthackray jthackray mentioned this pull request Mar 21, 2025
Merged
@phoebewang phoebewang deleted the AVX10 branch March 22, 2025 11:30
phoebewang added a commit to phoebewang/llvm-project that referenced this pull request Mar 27, 2025
@phoebewang
Revert "[X86][AVX10.2] Support YMM rounding new instructions (llvm#10…
3cad497 
…1825)" (llvm#132362)

This reverts commit 0dba538.

YMM rounding was removed from AVX10 whitepaper. Ref:
https://cdrdv2.intel.com/v1/dl/getContent/784343

The MINMAX and SATURATING CONVERT instructions will be removed as a
follow up.
@phoebewang phoebewang mentioned this pull request Mar 27, 2025
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@RKSimon RKSimon RKSimon approved these changes

@e-kud e-kud Awaiting requested review from e-kud

@KanRobert KanRobert Awaiting requested review from KanRobert

@FreddyLeaf FreddyLeaf Awaiting requested review from FreddyLeaf

Assignees
No one assigned
Labels
backend:X86 clang:codegen IR generation bugs: mangling, exceptions, etc. clang:frontend Language frontend issues, e.g. anything involving "Sema" clang:headers Headers provided by Clang, e.g. for intrinsics clang Clang issues not falling into any other category llvm:ir mc Machine (object) code
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@phoebewang @llvmbot @RKSimon