[llvm] [SPIRV] Addition of matrix multiply accumulate operands (PR #138665)
Aadesh Premkumar via llvm-commits
llvm-commits at lists.llvm.org
Tue May 6 04:54:23 PDT 2025
https://github.com/aadeshps-mcw updated https://github.com/llvm/llvm-project/pull/138665
>From 4ad6d5bab3a5a585b908bfbece7a10b05a09b979 Mon Sep 17 00:00:00 2001
From: Aadesh PremKumar <aadesh.premkumar at multicorewareinc.com>
Date: Tue, 6 May 2025 17:23:33 +0530
Subject: [PATCH] --Added matrix multiply accumulate operands for the extension
SPV_INTEL_subgroup_matrix_multiply_accumulate
---
.../Target/SPIRV/MCTargetDesc/SPIRVBaseInfo.h | 5 ++
.../SPIRV/MCTargetDesc/SPIRVInstPrinter.cpp | 31 ++++++++++
.../lib/Target/SPIRV/SPIRVSymbolicOperands.td | 38 ++++++++++++
...roup_matrix_multiply_accumulate_generic.ll | 58 +++++++++----------
4 files changed, 103 insertions(+), 29 deletions(-)
diff --git a/llvm/lib/Target/SPIRV/MCTargetDesc/SPIRVBaseInfo.h b/llvm/lib/Target/SPIRV/MCTargetDesc/SPIRVBaseInfo.h
index d009244a92259..3435e319fc816 100644
--- a/llvm/lib/Target/SPIRV/MCTargetDesc/SPIRVBaseInfo.h
+++ b/llvm/lib/Target/SPIRV/MCTargetDesc/SPIRVBaseInfo.h
@@ -57,6 +57,11 @@ namespace MemoryModel {
#include "SPIRVGenTables.inc"
} // namespace MemoryModel
+namespace MatrixMultiplyAccumulate {
+#define GET_MatrixMultiplyAccumulate_DECL
+#include "SPIRVGenTables.inc"
+} // namespace MatrixMultiplyAccumulate
+
namespace ExecutionMode {
#define GET_ExecutionMode_DECL
#include "SPIRVGenTables.inc"
diff --git a/llvm/lib/Target/SPIRV/MCTargetDesc/SPIRVInstPrinter.cpp b/llvm/lib/Target/SPIRV/MCTargetDesc/SPIRVInstPrinter.cpp
index b486132077e3b..4688b8dc7dcd0 100644
--- a/llvm/lib/Target/SPIRV/MCTargetDesc/SPIRVInstPrinter.cpp
+++ b/llvm/lib/Target/SPIRV/MCTargetDesc/SPIRVInstPrinter.cpp
@@ -242,6 +242,37 @@ void SPIRVInstPrinter::printInst(const MCInst *MI, uint64_t Address,
}
break;
}
+ case SPIRV::OpSubgroupMatrixMultiplyAccumulateINTEL: {
+ const unsigned NumOps = MI->getNumOperands();
+ if (NumFixedOps == NumOps)
+ break; // No extra operands, so no flags to process
+ OS << ' ';
+ // Extract the last operand only if it exists
+ if (NumOps > NumFixedOps) {
+ const unsigned Flags = MI->getOperand(NumOps - 1).getImm();
+ if (Flags == 0) {
+ printSymbolicOperand<
+ OperandCategory::MatrixMultiplyAccumulateOperandsOperand>(
+ MI, NumOps - 1, OS);
+ } else {
+ std::string Buffer;
+ for (unsigned Mask = 0x1;
+ Mask != SPIRV::MatrixMultiplyAccumulate::
+ MatrixBPackedBFloat16INTEL; // Replace with
+ // actual last flag
+ Mask <<= 1) {
+ if (Flags & Mask) {
+ if (!Buffer.empty())
+ Buffer += '|';
+ Buffer += getSymbolicOperandMnemonic(
+ OperandCategory::MatrixMultiplyAccumulateOperandsOperand, Mask);
+ }
+ }
+ OS << Buffer;
+ }
+ }
+ break;
+ }
default:
printRemainingVariableOps(MI, NumFixedOps, OS);
break;
diff --git a/llvm/lib/Target/SPIRV/SPIRVSymbolicOperands.td b/llvm/lib/Target/SPIRV/SPIRVSymbolicOperands.td
index cc32200a0a261..854368247b336 100644
--- a/llvm/lib/Target/SPIRV/SPIRVSymbolicOperands.td
+++ b/llvm/lib/Target/SPIRV/SPIRVSymbolicOperands.td
@@ -172,6 +172,7 @@ def KernelProfilingInfoOperand : OperandCategory;
def OpcodeOperand : OperandCategory;
def CooperativeMatrixLayoutOperand : OperandCategory;
def CooperativeMatrixOperandsOperand : OperandCategory;
+def MatrixMultiplyAccumulateOperandsOperand :OperandCategory;
//===----------------------------------------------------------------------===//
// Multiclass used to define Extesions enum values and at the same time
@@ -1746,3 +1747,40 @@ defm MatrixAAndBTF32ComponentsINTEL : CooperativeMatrixOperandsOperand<0x20, [SP
defm MatrixAAndBBFloat16ComponentsINTEL : CooperativeMatrixOperandsOperand<0x40, [SPV_INTEL_joint_matrix], [CooperativeMatrixBFloat16ComponentTypeINTEL]>;
defm MatrixCBFloat16ComponentsINTEL : CooperativeMatrixOperandsOperand<0x80, [SPV_INTEL_joint_matrix], [CooperativeMatrixBFloat16ComponentTypeINTEL]>;
defm MatrixResultBFloat16ComponentsINTEL : CooperativeMatrixOperandsOperand<0x100, [SPV_INTEL_joint_matrix], [CooperativeMatrixBFloat16ComponentTypeINTEL]>;
+
+//===----------------------------------------------------------------------===//
+// Multiclass used to define Matrix Multiply Accumulate Operands enum values and at the same time
+// SymbolicOperand entries with string mnemonics and capabilities.
+//===----------------------------------------------------------------------===//
+def MatrixMultiplyAccumulate : GenericEnum, Operand<i32> {
+ let FilterClass = "MatrixMultiplyAccumulate";
+ let NameField = "Name";
+ let ValueField = "Value";
+ let PrintMethod = !strconcat("printSymbolicOperand<OperandCategory::", FilterClass, "Operand>");
+}
+
+class MatrixMultiplyAccumulate<string name, bits<32> value> {
+ string Name = name;
+ bits<32> Value = value;
+}
+
+multiclass MatrixMultiplyAccumulateOperandsOperand<bits<32> value, list<Extension> reqExtensions> {
+def : MatrixMultiplyAccumulate<NAME, value>;
+ defm : SymbolicOperandWithRequirements< MatrixMultiplyAccumulateOperandsOperand, value, NAME, 0, 0, reqExtensions, []>;
+}
+
+defm None : MatrixMultiplyAccumulateOperandsOperand<0x0, [SPV_INTEL_subgroup_matrix_multiply_accumulate]>;
+defm MatrixASignedComponentsINTEL : MatrixMultiplyAccumulateOperandsOperand<0x1, [SPV_INTEL_subgroup_matrix_multiply_accumulate]>;
+defm MatrixBSignedComponentsINTEL : MatrixMultiplyAccumulateOperandsOperand<0x2, [SPV_INTEL_subgroup_matrix_multiply_accumulate]>;
+defm MatrixCBFloat16INTEL : MatrixMultiplyAccumulateOperandsOperand<0x4, [SPV_INTEL_subgroup_matrix_multiply_accumulate]>;
+defm MatrixResultBFloat16INTEL : MatrixMultiplyAccumulateOperandsOperand<0x8, [SPV_INTEL_subgroup_matrix_multiply_accumulate]>;
+defm MatrixAPackedInt8INTEL : MatrixMultiplyAccumulateOperandsOperand<0x10, [SPV_INTEL_subgroup_matrix_multiply_accumulate]>;
+defm MatrixBPackedInt8INTEL : MatrixMultiplyAccumulateOperandsOperand<0x20, [SPV_INTEL_subgroup_matrix_multiply_accumulate]>;
+defm MatrixAPackedInt4INTEL : MatrixMultiplyAccumulateOperandsOperand<0x40, [SPV_INTEL_subgroup_matrix_multiply_accumulate]>;
+defm MatrixBPackedInt4INTEL : MatrixMultiplyAccumulateOperandsOperand<0x80, [SPV_INTEL_subgroup_matrix_multiply_accumulate]>;
+defm MatrixATF32INTEL : MatrixMultiplyAccumulateOperandsOperand<0x100, [SPV_INTEL_subgroup_matrix_multiply_accumulate]>;
+defm MatrixBTF32INTEL : MatrixMultiplyAccumulateOperandsOperand<0x200, [SPV_INTEL_subgroup_matrix_multiply_accumulate]>;
+defm MatrixAPackedFloat16INTEL : MatrixMultiplyAccumulateOperandsOperand<0x400, [SPV_INTEL_subgroup_matrix_multiply_accumulate]>;
+defm MatrixBPackedFloat16INTEL : MatrixMultiplyAccumulateOperandsOperand<0x800, [SPV_INTEL_subgroup_matrix_multiply_accumulate]>;
+defm MatrixAPackedBFloat16INTEL : MatrixMultiplyAccumulateOperandsOperand<0x1000, [SPV_INTEL_subgroup_matrix_multiply_accumulate]>;
+defm MatrixBPackedBFloat16INTEL : MatrixMultiplyAccumulateOperandsOperand<0x2000, [SPV_INTEL_subgroup_matrix_multiply_accumulate]>;
diff --git a/llvm/test/CodeGen/SPIRV/extensions/SPV_INTEL_subgroup_matrix_multiply_accumulate/subgroup_matrix_multiply_accumulate_generic.ll b/llvm/test/CodeGen/SPIRV/extensions/SPV_INTEL_subgroup_matrix_multiply_accumulate/subgroup_matrix_multiply_accumulate_generic.ll
index 0cd6992936eeb..b945b8795baa0 100644
--- a/llvm/test/CodeGen/SPIRV/extensions/SPV_INTEL_subgroup_matrix_multiply_accumulate/subgroup_matrix_multiply_accumulate_generic.ll
+++ b/llvm/test/CodeGen/SPIRV/extensions/SPV_INTEL_subgroup_matrix_multiply_accumulate/subgroup_matrix_multiply_accumulate_generic.ll
@@ -131,35 +131,35 @@
; CHECK: %[[#hM2:]] = OpFunctionParameter %[[#Vec2HalfTy]]
; CHECK: %[[#hM4:]] = OpFunctionParameter %[[#Vec4HalfTy]]
; CHECK: %[[#hM8:]] = OpFunctionParameter %[[#Vec8HalfTy]]
-; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Int32Ty]] %[[#Const42]] %[[#iM]] %[[#iM8]] %[[#iM]] 10
-; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec2Int32Ty]] %[[#Const42]] %[[#iM2]] %[[#iM8]] %[[#iM2]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec4Int32Ty]] %[[#Const42]] %[[#iM4]] %[[#iM8]] %[[#iM4]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec8Int32Ty]] %[[#Const42]] %[[#iM8]] %[[#iM8]] %[[#iM8]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#FloatTy]] %[[#Const42]] %[[#iM]] %[[#iM8]] %[[#fM]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec2FloatTy]] %[[#Const42]] %[[#iM2]] %[[#iM8]] %[[#fM2]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec4FloatTy]] %[[#Const42]] %[[#iM4]] %[[#iM8]] %[[#fM4]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec8FloatTy]] %[[#Const42]] %[[#iM8]] %[[#iM8]] %[[#fM8]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Int32Ty]] %[[#Const42]] %[[#sM]] %[[#iM8]] %[[#iM]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec2Int32Ty]] %[[#Const42]] %[[#sM2]] %[[#iM8]] %[[#iM2]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec4Int32Ty]] %[[#Const42]] %[[#sM4]] %[[#iM8]] %[[#iM4]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec8Int32Ty]] %[[#Const42]] %[[#sM8]] %[[#iM8]] %[[#iM8]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#FloatTy]] %[[#Const42]] %[[#sM]] %[[#iM8]] %[[#fM]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec2FloatTy]] %[[#Const42]] %[[#sM2]] %[[#iM8]] %[[#fM2]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec4FloatTy]] %[[#Const42]] %[[#sM4]] %[[#iM8]] %[[#fM4]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec8FloatTy]] %[[#Const42]] %[[#sM8]] %[[#iM8]] %[[#fM8]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#HalfTy]] %[[#Const42]] %[[#sM]] %[[#iM8]] %[[#hM]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec2HalfTy]] %[[#Const42]] %[[#sM2]] %[[#iM8]] %[[#hM2]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec4HalfTy]] %[[#Const42]] %[[#sM4]] %[[#iM8]] %[[#hM4]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec8HalfTy]] %[[#Const42]] %[[#sM8]] %[[#iM8]] %[[#hM8]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Int16Ty]] %[[#Const42]] %[[#sM]] %[[#iM8]] %[[#sM]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec2Int16Ty]] %[[#Const42]] %[[#sM2]] %[[#iM8]] %[[#sM2]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec4Int16Ty]] %[[#Const42]] %[[#sM4]] %[[#iM8]] %[[#sM4]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec8Int16Ty]] %[[#Const42]] %[[#sM8]] %[[#iM8]] %[[#sM8]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#FloatTy]] %[[#Const42]] %[[#fM]] %[[#fM8]] %[[#fM]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec2FloatTy]] %[[#Const42]] %[[#fM2]] %[[#fM8]] %[[#fM2]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec4FloatTy]] %[[#Const42]] %[[#fM4]] %[[#fM8]] %[[#fM4]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec8FloatTy]] %[[#Const42]] %[[#fM8]] %[[#fM8]] %[[#fM8]] 10
-; CHECK: OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec4FloatTy]] %[[#Const42]] %[[#sM4]] %[[#iM8]] %[[#fM4]]
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Int32Ty]] %[[#Const42]] %[[#iM]] %[[#iM8]] %[[#iM]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec2Int32Ty]] %[[#Const42]] %[[#iM2]] %[[#iM8]] %[[#iM2]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec4Int32Ty]] %[[#Const42]] %[[#iM4]] %[[#iM8]] %[[#iM4]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec8Int32Ty]] %[[#Const42]] %[[#iM8]] %[[#iM8]] %[[#iM8]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#FloatTy]] %[[#Const42]] %[[#iM]] %[[#iM8]] %[[#fM]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec2FloatTy]] %[[#Const42]] %[[#iM2]] %[[#iM8]] %[[#fM2]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec4FloatTy]] %[[#Const42]] %[[#iM4]] %[[#iM8]] %[[#fM4]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec8FloatTy]] %[[#Const42]] %[[#iM8]] %[[#iM8]] %[[#fM8]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Int32Ty]] %[[#Const42]] %[[#sM]] %[[#iM8]] %[[#iM]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec2Int32Ty]] %[[#Const42]] %[[#sM2]] %[[#iM8]] %[[#iM2]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec4Int32Ty]] %[[#Const42]] %[[#sM4]] %[[#iM8]] %[[#iM4]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec8Int32Ty]] %[[#Const42]] %[[#sM8]] %[[#iM8]] %[[#iM8]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#FloatTy]] %[[#Const42]] %[[#sM]] %[[#iM8]] %[[#fM]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec2FloatTy]] %[[#Const42]] %[[#sM2]] %[[#iM8]] %[[#fM2]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec4FloatTy]] %[[#Const42]] %[[#sM4]] %[[#iM8]] %[[#fM4]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec8FloatTy]] %[[#Const42]] %[[#sM8]] %[[#iM8]] %[[#fM8]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#HalfTy]] %[[#Const42]] %[[#sM]] %[[#iM8]] %[[#hM]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec2HalfTy]] %[[#Const42]] %[[#sM2]] %[[#iM8]] %[[#hM2]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec4HalfTy]] %[[#Const42]] %[[#sM4]] %[[#iM8]] %[[#hM4]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec8HalfTy]] %[[#Const42]] %[[#sM8]] %[[#iM8]] %[[#hM8]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Int16Ty]] %[[#Const42]] %[[#sM]] %[[#iM8]] %[[#sM]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec2Int16Ty]] %[[#Const42]] %[[#sM2]] %[[#iM8]] %[[#sM2]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec4Int16Ty]] %[[#Const42]] %[[#sM4]] %[[#iM8]] %[[#sM4]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec8Int16Ty]] %[[#Const42]] %[[#sM8]] %[[#iM8]] %[[#sM8]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#FloatTy]] %[[#Const42]] %[[#fM]] %[[#fM8]] %[[#fM]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec2FloatTy]] %[[#Const42]] %[[#fM2]] %[[#fM8]] %[[#fM2]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec4FloatTy]] %[[#Const42]] %[[#fM4]] %[[#fM8]] %[[#fM4]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec8FloatTy]] %[[#Const42]] %[[#fM8]] %[[#fM8]] %[[#fM8]] MatrixBSignedComponentsINTEL|MatrixResultBFloat16INTEL
+; CHECK: %[[#]] = OpSubgroupMatrixMultiplyAccumulateINTEL %[[#Vec4FloatTy]] %[[#Const42]] %[[#sM4]] %[[#iM8]] %[[#fM4]]
define spir_func void @foo(i32 %iM, <2 x i32> %iM2, <4 x i32> %iM4, <8 x i32> %iM8,
i16 signext %sM, <2 x i16> %sM2, <4 x i16> %sM4, <8 x i16> %sM8,
More information about the llvm-commits
mailing list