Skip to content
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.

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

Add support for AvxVnni instructions under Experimental. #51998

Merged
merged 12 commits into from
Jun 2, 2021
82 changes: 50 additions & 32 deletions src/coreclr/inc/corinfoinstructionset.h
Original file line number Diff line number Diff line change
Expand Up @@ -56,22 +56,24 @@ enum CORINFO_InstructionSet
InstructionSet_POPCNT=16,
InstructionSet_Vector128=17,
InstructionSet_Vector256=18,
InstructionSet_X86Base_X64=19,
InstructionSet_SSE_X64=20,
InstructionSet_SSE2_X64=21,
InstructionSet_SSE3_X64=22,
InstructionSet_SSSE3_X64=23,
InstructionSet_SSE41_X64=24,
InstructionSet_SSE42_X64=25,
InstructionSet_AVX_X64=26,
InstructionSet_AVX2_X64=27,
InstructionSet_AES_X64=28,
InstructionSet_BMI1_X64=29,
InstructionSet_BMI2_X64=30,
InstructionSet_FMA_X64=31,
InstructionSet_LZCNT_X64=32,
InstructionSet_PCLMULQDQ_X64=33,
InstructionSet_POPCNT_X64=34,
InstructionSet_AVXVNNI=19,
InstructionSet_X86Base_X64=20,
InstructionSet_SSE_X64=21,
InstructionSet_SSE2_X64=22,
InstructionSet_SSE3_X64=23,
InstructionSet_SSSE3_X64=24,
InstructionSet_SSE41_X64=25,
InstructionSet_SSE42_X64=26,
InstructionSet_AVX_X64=27,
InstructionSet_AVX2_X64=28,
InstructionSet_AES_X64=29,
InstructionSet_BMI1_X64=30,
InstructionSet_BMI2_X64=31,
InstructionSet_FMA_X64=32,
InstructionSet_LZCNT_X64=33,
InstructionSet_PCLMULQDQ_X64=34,
InstructionSet_POPCNT_X64=35,
InstructionSet_AVXVNNI_X64=36,
#endif // TARGET_AMD64
#ifdef TARGET_X86
InstructionSet_X86Base=1,
Expand All @@ -92,22 +94,24 @@ enum CORINFO_InstructionSet
InstructionSet_POPCNT=16,
InstructionSet_Vector128=17,
InstructionSet_Vector256=18,
InstructionSet_X86Base_X64=19,
InstructionSet_SSE_X64=20,
InstructionSet_SSE2_X64=21,
InstructionSet_SSE3_X64=22,
InstructionSet_SSSE3_X64=23,
InstructionSet_SSE41_X64=24,
InstructionSet_SSE42_X64=25,
InstructionSet_AVX_X64=26,
InstructionSet_AVX2_X64=27,
InstructionSet_AES_X64=28,
InstructionSet_BMI1_X64=29,
InstructionSet_BMI2_X64=30,
InstructionSet_FMA_X64=31,
InstructionSet_LZCNT_X64=32,
InstructionSet_PCLMULQDQ_X64=33,
InstructionSet_POPCNT_X64=34,
InstructionSet_AVXVNNI=19,
InstructionSet_X86Base_X64=20,
InstructionSet_SSE_X64=21,
InstructionSet_SSE2_X64=22,
InstructionSet_SSE3_X64=23,
InstructionSet_SSSE3_X64=24,
InstructionSet_SSE41_X64=25,
InstructionSet_SSE42_X64=26,
InstructionSet_AVX_X64=27,
InstructionSet_AVX2_X64=28,
InstructionSet_AES_X64=29,
InstructionSet_BMI1_X64=30,
InstructionSet_BMI2_X64=31,
InstructionSet_FMA_X64=32,
InstructionSet_LZCNT_X64=33,
InstructionSet_PCLMULQDQ_X64=34,
InstructionSet_POPCNT_X64=35,
InstructionSet_AVXVNNI_X64=36,
#endif // TARGET_X86

};
Expand Down Expand Up @@ -205,6 +209,8 @@ struct CORINFO_InstructionSetFlags
AddInstructionSet(InstructionSet_PCLMULQDQ_X64);
if (HasInstructionSet(InstructionSet_POPCNT))
AddInstructionSet(InstructionSet_POPCNT_X64);
if (HasInstructionSet(InstructionSet_AVXVNNI))
AddInstructionSet(InstructionSet_AVXVNNI_X64);
#endif // TARGET_AMD64
#ifdef TARGET_X86
#endif // TARGET_X86
Expand Down Expand Up @@ -342,6 +348,10 @@ inline CORINFO_InstructionSetFlags EnsureInstructionSetFlagsAreValid(CORINFO_Ins
resultflags.RemoveInstructionSet(InstructionSet_POPCNT);
if (resultflags.HasInstructionSet(InstructionSet_POPCNT_X64) && !resultflags.HasInstructionSet(InstructionSet_POPCNT))
resultflags.RemoveInstructionSet(InstructionSet_POPCNT_X64);
if (resultflags.HasInstructionSet(InstructionSet_AVXVNNI) && !resultflags.HasInstructionSet(InstructionSet_AVXVNNI_X64))
resultflags.RemoveInstructionSet(InstructionSet_AVXVNNI);
if (resultflags.HasInstructionSet(InstructionSet_AVXVNNI_X64) && !resultflags.HasInstructionSet(InstructionSet_AVXVNNI))
resultflags.RemoveInstructionSet(InstructionSet_AVXVNNI_X64);
if (resultflags.HasInstructionSet(InstructionSet_SSE) && !resultflags.HasInstructionSet(InstructionSet_X86Base))
resultflags.RemoveInstructionSet(InstructionSet_SSE);
if (resultflags.HasInstructionSet(InstructionSet_SSE2) && !resultflags.HasInstructionSet(InstructionSet_SSE))
Expand Down Expand Up @@ -530,6 +540,10 @@ inline const char *InstructionSetToString(CORINFO_InstructionSet instructionSet)
return "Vector128";
case InstructionSet_Vector256 :
return "Vector256";
case InstructionSet_AVXVNNI :
return "AVXVNNI";
case InstructionSet_AVXVNNI_X64 :
return "AVXVNNI_X64";
#endif // TARGET_AMD64
#ifdef TARGET_X86
case InstructionSet_X86Base :
Expand Down Expand Up @@ -568,6 +582,8 @@ inline const char *InstructionSetToString(CORINFO_InstructionSet instructionSet)
return "Vector128";
case InstructionSet_Vector256 :
return "Vector256";
case InstructionSet_AVXVNNI :
return "AVXVNNI";
#endif // TARGET_X86

default:
Expand Down Expand Up @@ -615,6 +631,7 @@ inline CORINFO_InstructionSet InstructionSetFromR2RInstructionSet(ReadyToRunInst
case READYTORUN_INSTRUCTION_Lzcnt: return InstructionSet_LZCNT;
case READYTORUN_INSTRUCTION_Pclmulqdq: return InstructionSet_PCLMULQDQ;
case READYTORUN_INSTRUCTION_Popcnt: return InstructionSet_POPCNT;
case READYTORUN_INSTRUCTION_AvxVnni: return InstructionSet_AVXVNNI;
#endif // TARGET_AMD64
#ifdef TARGET_X86
case READYTORUN_INSTRUCTION_X86Base: return InstructionSet_X86Base;
Expand All @@ -633,6 +650,7 @@ inline CORINFO_InstructionSet InstructionSetFromR2RInstructionSet(ReadyToRunInst
case READYTORUN_INSTRUCTION_Lzcnt: return InstructionSet_LZCNT;
case READYTORUN_INSTRUCTION_Pclmulqdq: return InstructionSet_PCLMULQDQ;
case READYTORUN_INSTRUCTION_Popcnt: return InstructionSet_POPCNT;
case READYTORUN_INSTRUCTION_AvxVnni: return InstructionSet_AVXVNNI;
#endif // TARGET_X86

default:
Expand Down
1 change: 1 addition & 0 deletions src/coreclr/inc/readytoruninstructionset.h
Original file line number Diff line number Diff line change
Expand Up @@ -33,6 +33,7 @@ enum ReadyToRunInstructionSet
READYTORUN_INSTRUCTION_X86Base=22,
READYTORUN_INSTRUCTION_Dp=23,
READYTORUN_INSTRUCTION_Rdm=24,
READYTORUN_INSTRUCTION_AvxVnni=25,

};

Expand Down
5 changes: 5 additions & 0 deletions src/coreclr/jit/compiler.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -2415,6 +2415,11 @@ void Compiler::compSetProcessor()
instructionSetFlags.RemoveInstructionSet(InstructionSet_AVX2);
}

if (!JitConfig.EnableAVXVNNI())
{
instructionSetFlags.RemoveInstructionSet(InstructionSet_AVXVNNI);
}

if (!JitConfig.EnableLZCNT())
{
instructionSetFlags.RemoveInstructionSet(InstructionSet_LZCNT);
Expand Down
18 changes: 16 additions & 2 deletions src/coreclr/jit/emitxarch.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -44,6 +44,11 @@ bool IsFMAInstruction(instruction ins)
return (ins >= INS_FIRST_FMA_INSTRUCTION) && (ins <= INS_LAST_FMA_INSTRUCTION);
}

bool IsAVXVNNIInstruction(instruction ins)
{
return (ins >= INS_FIRST_AVXVNNI_INSTRUCTION) && (ins <= INS_LAST_AVXVNNI_INSTRUCTION);
}

bool IsBMIInstruction(instruction ins)
{
return (ins >= INS_FIRST_BMI_INSTRUCTION) && (ins <= INS_LAST_BMI_INSTRUCTION);
Expand Down Expand Up @@ -6158,7 +6163,7 @@ void emitter::emitIns_SIMD_R_R_S_I(
void emitter::emitIns_SIMD_R_R_R_A(
instruction ins, emitAttr attr, regNumber targetReg, regNumber op1Reg, regNumber op2Reg, GenTreeIndir* indir)
{
assert(IsFMAInstruction(ins));
assert(IsFMAInstruction(ins) || IsAVXVNNIInstruction(ins));
assert(UseVEXEncoding());

if (op1Reg != targetReg)
Expand Down Expand Up @@ -6267,6 +6272,11 @@ void emitter::emitIns_SIMD_R_R_R_R(

emitIns_R_R_R(ins, attr, targetReg, op2Reg, op3Reg);
}
else if (IsAVXVNNIInstruction(ins))
{
assert(UseVEXEncoding());
emitIns_R_R_R(ins, attr, targetReg, op2Reg, op3Reg);
tannergooding marked this conversation as resolved.
Show resolved Hide resolved
}
else if (UseVEXEncoding())
{
assert(isAvxBlendv(ins) || isSse41Blendv(ins));
Expand Down Expand Up @@ -6328,7 +6338,7 @@ void emitter::emitIns_SIMD_R_R_R_R(
void emitter::emitIns_SIMD_R_R_R_S(
instruction ins, emitAttr attr, regNumber targetReg, regNumber op1Reg, regNumber op2Reg, int varx, int offs)
{
assert(IsFMAInstruction(ins));
assert(IsFMAInstruction(ins) || IsAVXVNNIInstruction(ins));
assert(UseVEXEncoding());

if (op1Reg != targetReg)
Expand Down Expand Up @@ -15391,6 +15401,10 @@ emitter::insExecutionCharacteristics emitter::getInsExecutionCharacteristics(ins
case INS_vfnmsub132ss:
case INS_vfnmsub213ss:
case INS_vfnmsub231ss:
case INS_vpdpbusd: //will be populated when the HW becomes publicly available
Copy link
Member

Choose a reason for hiding this comment

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

Can we get a small issue tracking this to ensure it doesn't get lost/forgotten?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Do you want me to open an issue now, or at the time when the code could be merged?

Copy link
Member

Choose a reason for hiding this comment

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

Either is fine, just as long as we are tracking updating it once the official numbers become available.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

@tannergooding, #52121 is tracking this.

Copy link
Member

Choose a reason for hiding this comment

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

Thanks!

case INS_vpdpwssd: //will be populated when the HW becomes publicly available
case INS_vpdpbusds: //will be populated when the HW becomes publicly available
case INS_vpdpwssds: //will be populated when the HW becomes publicly available
// uops.info
result.insThroughput = PERFSCORE_THROUGHPUT_2X;
result.insLatency += PERFSCORE_LATENCY_4C;
Expand Down
15 changes: 15 additions & 0 deletions src/coreclr/jit/hwintrinsiccodegenxarch.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -333,7 +333,22 @@ void CodeGen::genHWIntrinsic(GenTreeHWIntrinsic* node)
genHWIntrinsic_R_R_RM_R(node, ins);
break;
}
case NI_AVXVNNI_MultiplyWideningAndAdd:
case NI_AVXVNNI_MultiplyWideningAndAddSaturate:
{
assert(targetReg != REG_NA);
Copy link
Contributor

Choose a reason for hiding this comment

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

These assertions are not really needed here - we would check the same invariants at the beginning of genHWIntrinsic_R_R_R_RM

assert(op1Reg != REG_NA);
assert(op2Reg != REG_NA);

if (targetReg != op1Reg)
{
node->SetRegNum(op1Reg);
targetReg = op1Reg;
Copy link
Member

Choose a reason for hiding this comment

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

We aren't doing this for any other HWIntrinsic instructions, instead we generate a movaps as required to put op1Reg in targetReg.

I'd expect register allocation to have already done everything at this point, but I'm not an expert here. @kunalspathak ?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

@kunalspathak Does this change looks good to you or could you give me some suggestion on how should we handle it? Thanks.

Copy link
Member

Choose a reason for hiding this comment

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

I agree with @tannergooding . While I have limited knowledge, but I can see that we generate movaps if targetReg != op1Reg going forward in genHWIntrinsic_R_R_R_RM(). Is there a situation where that doesn't happen and we have to do targetReg = opt1Reg explicitely?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

I double checked this part of the code and agree that we don't need to do this. I will update the code.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

While I was checking this part, I encountered some issue when I set COMPlus_jitDump=RunBasicScenario_Load. The error message I saw is as the following (same error also found in other API such as the ones in Fma_vector256):

Assert failure(PID 5076 [0x000013d4], Thread: 26664 [0x6828]): Assertion failed 'type < CORINFO_TYPE_COUNT' in 'JIT.HardwareIntrinsics.X86.SimpleTernaryOpTest__MultiplyAddDouble:RunBasicScenario_Load():this' during 'Morph - Global' (IL size 137)

File: ...\runtime\src\coreclr\jit\ee_il_dll.hpp Line: 273
Image: ...\runtime\artifacts\bin\coreclr\windows.x64.Debug\corerun.exe

@tannergooding, do you know what is happening here?

Copy link
Member

Choose a reason for hiding this comment

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

I'd need to see the full stack trace, but I suspect this would be fixed by merging with (or rebasing onto) the latest dotnet/main.

Copy link
Member

Choose a reason for hiding this comment

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

For reference, there were a few fixes that went in (#52016, 0e12823, and #51627) resolving asserts that had cropped up related to adding native integer support to the HWIntrinsics code.

}

genHWIntrinsic_R_R_R_RM(ins, simdSize, targetReg, op1Reg, op2Reg, op3);
break;
}
default:
{
unreached();
Expand Down
8 changes: 7 additions & 1 deletion src/coreclr/jit/hwintrinsiclistxarch.h
Original file line number Diff line number Diff line change
Expand Up @@ -567,7 +567,13 @@ HARDWARE_INTRINSIC(AVX2, SubtractSaturate,
HARDWARE_INTRINSIC(AVX2, UnpackHigh, 32, 2, {INS_punpckhbw, INS_punpckhbw, INS_punpckhwd, INS_punpckhwd, INS_punpckhdq, INS_punpckhdq, INS_punpckhqdq, INS_punpckhqdq, INS_invalid, INS_invalid}, HW_Category_SimpleSIMD, HW_Flag_NoFlag)
HARDWARE_INTRINSIC(AVX2, UnpackLow, 32, 2, {INS_punpcklbw, INS_punpcklbw, INS_punpcklwd, INS_punpcklwd, INS_punpckldq, INS_punpckldq, INS_punpcklqdq, INS_punpcklqdq, INS_invalid, INS_invalid}, HW_Category_SimpleSIMD, HW_Flag_NoFlag)
HARDWARE_INTRINSIC(AVX2, Xor, 32, 2, {INS_pxor, INS_pxor, INS_pxor, INS_pxor, INS_pxor, INS_pxor, INS_pxor, INS_pxor, INS_invalid, INS_invalid}, HW_Category_SimpleSIMD, HW_Flag_Commutative)

// ***************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************
// ISA Function name SIMD size NumArg Instructions Category Flags
// {TYP_BYTE, TYP_UBYTE, TYP_SHORT, TYP_USHORT, TYP_INT, TYP_UINT, TYP_LONG, TYP_ULONG, TYP_FLOAT, TYP_DOUBLE}
// ***************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************
// AVXVNNI Intrinsics
HARDWARE_INTRINSIC(AVXVNNI, MultiplyWideningAndAdd, -1, 3, {INS_invalid, INS_vpdpbusd, INS_vpdpwssd, INS_invalid, INS_invalid, INS_invalid, INS_invalid, INS_invalid, INS_invalid, INS_invalid}, HW_Category_SimpleSIMD, HW_Flag_NoFloatingPointUsed|HW_Flag_BaseTypeFromSecondArg)
HARDWARE_INTRINSIC(AVXVNNI, MultiplyWideningAndAddSaturate, -1, 3, {INS_invalid, INS_vpdpbusds, INS_vpdpwssds, INS_invalid, INS_invalid, INS_invalid, INS_invalid, INS_invalid, INS_invalid, INS_invalid}, HW_Category_SimpleSIMD, HW_Flag_NoFloatingPointUsed|HW_Flag_BaseTypeFromSecondArg)
// ***************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************
// ISA Function name SIMD size NumArg Instructions Category Flags
// {TYP_BYTE, TYP_UBYTE, TYP_SHORT, TYP_USHORT, TYP_INT, TYP_UINT, TYP_LONG, TYP_ULONG, TYP_FLOAT, TYP_DOUBLE}
Expand Down
8 changes: 8 additions & 0 deletions src/coreclr/jit/hwintrinsicxarch.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -36,6 +36,8 @@ static CORINFO_InstructionSet X64VersionOfIsa(CORINFO_InstructionSet isa)
return InstructionSet_AVX_X64;
case InstructionSet_AVX2:
return InstructionSet_AVX2_X64;
case InstructionSet_AVXVNNI:
return InstructionSet_AVXVNNI_X64;
case InstructionSet_AES:
return InstructionSet_AES_X64;
case InstructionSet_BMI1:
Expand Down Expand Up @@ -80,6 +82,10 @@ static CORINFO_InstructionSet lookupInstructionSet(const char* className)
{
return InstructionSet_AVX2;
}
if (strcmp(className, "AvxVnni") == 0)
{
return InstructionSet_AVXVNNI;
}
}
else if (className[0] == 'S')
{
Expand Down Expand Up @@ -348,6 +354,8 @@ bool HWIntrinsicInfo::isFullyImplementedIsa(CORINFO_InstructionSet isa)
case InstructionSet_AVX_X64:
case InstructionSet_AVX2:
case InstructionSet_AVX2_X64:
case InstructionSet_AVXVNNI:
case InstructionSet_AVXVNNI_X64:
case InstructionSet_BMI1:
case InstructionSet_BMI1_X64:
case InstructionSet_BMI2:
Expand Down
7 changes: 7 additions & 0 deletions src/coreclr/jit/instrsxarch.h
Original file line number Diff line number Diff line change
Expand Up @@ -583,6 +583,13 @@ INST3(vfnmsub213ss, "fnmsub213ss", IUM_WR, BAD_CODE, BAD_CODE,
INST3(vfnmsub231ss, "fnmsub231ss", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xBF), INS_Flags_IsDstDstSrcAVXInstruction) //
INST3(LAST_FMA_INSTRUCTION, "LAST_FMA_INSTRUCTION", IUM_WR, BAD_CODE, BAD_CODE, BAD_CODE, INS_FLAGS_None)

INST3(FIRST_AVXVNNI_INSTRUCTION, "FIRST_AVXVNNI_INSTRUCTION", IUM_WR, BAD_CODE, BAD_CODE, BAD_CODE, INS_FLAGS_None)
INST3(vpdpbusd, "pdpbusd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x50), INS_Flags_IsDstDstSrcAVXInstruction) // Multiply and Add Unsigned and Signed Bytes
INST3(vpdpwssd, "pdpwssd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x52), INS_Flags_IsDstDstSrcAVXInstruction) // Multiply and Add Signed Word Integers
INST3(vpdpbusds, "pdpbusds", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x51), INS_Flags_IsDstDstSrcAVXInstruction) // Multiply and Add Unsigned and Signed Bytes with Saturation
INST3(vpdpwssds, "pdpwssds", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x53), INS_Flags_IsDstDstSrcAVXInstruction) // Multiply and Add Signed Word Integers with Saturation
INST3(LAST_AVXVNNI_INSTRUCTION, "LAST_AVXVNNI_INSTRUCTION", IUM_WR, BAD_CODE, BAD_CODE, BAD_CODE, INS_FLAGS_None)

// BMI1
INST3(FIRST_BMI_INSTRUCTION, "FIRST_BMI_INSTRUCTION", IUM_WR, BAD_CODE, BAD_CODE, BAD_CODE, INS_FLAGS_None)
INST3(andn, "andn", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xF2), INS_Flags_IsDstDstSrcAVXInstruction) // Logical AND NOT
Expand Down
1 change: 1 addition & 0 deletions src/coreclr/jit/jitconfigvalues.h
Original file line number Diff line number Diff line change
Expand Up @@ -273,6 +273,7 @@ CONFIG_INTEGER(EnableSSE41, W("EnableSSE41"), 1) // Enable SSE41
CONFIG_INTEGER(EnableSSE42, W("EnableSSE42"), 1) // Enable SSE42
CONFIG_INTEGER(EnableAVX, W("EnableAVX"), 1) // Enable AVX
CONFIG_INTEGER(EnableAVX2, W("EnableAVX2"), 1) // Enable AVX2
CONFIG_INTEGER(EnableAVXVNNI, W("EnableAVXVNNI"), 1) // Enable AVXVNNI
CONFIG_INTEGER(EnableFMA, W("EnableFMA"), 1) // Enable FMA
CONFIG_INTEGER(EnableAES, W("EnableAES"), 1) // Enable AES
CONFIG_INTEGER(EnableBMI1, W("EnableBMI1"), 1) // Enable BMI1
Expand Down
14 changes: 13 additions & 1 deletion src/coreclr/jit/lowerxarch.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -5609,7 +5609,19 @@ void Lowering::ContainCheckHWIntrinsic(GenTreeHWIntrinsic* node)
}
break;
}

case NI_AVXVNNI_MultiplyWideningAndAdd:
case NI_AVXVNNI_MultiplyWideningAndAddSaturate:
{
if (IsContainableHWIntrinsicOp(node, op3, &supportsRegOptional))
{
MakeSrcContained(node, op3);
}
else if (supportsRegOptional)
{
op3->SetRegOptional();
}
break;
}
case NI_BMI2_MultiplyNoFlags:
case NI_BMI2_X64_MultiplyNoFlags:
{
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -36,6 +36,7 @@ public enum ReadyToRunInstructionSet
X86Base=22,
Dp=23,
Rdm=24,
AvxVnni=25,

}
}
Loading