-
Notifications
You must be signed in to change notification settings - Fork 1.7k
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
[AMD][Navi31] Convert WMMA dot op to LLVM
-Add WMMA conversion logic for dot operation -Fix helper hunctions for WMMA layout -Add lit test for WMMA dot operation conversion Signed-off-by: joviliast <iveselov.nn@gmail.com>
- Loading branch information
Showing
5 changed files
with
292 additions
and
20 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,19 @@ | ||
| // RUN: triton-opt %s --split-input-file --convert-triton-amdgpu-to-llvm | FileCheck %s | ||
|
|
||
| // CHECK-LABEL: wmma_dot | ||
| #blocked = #triton_gpu.blocked<{sizePerThread = [1, 8], threadsPerWarp = [4, 8], warpsPerCTA = [4, 1], order = [1, 0]}> | ||
| #mma = #triton_gpu.amd_wmma<{warpsPerCTA = [2, 2]}> | ||
| module attributes {"triton_gpu.compute-capability" = 90 : i32, "triton_gpu.num-ctas" = 1 : i32, "triton_gpu.num-warps" = 4 : i32, "triton_gpu.threads-per-warp" = 32 : i32} { | ||
| tt.func @wmma_dot(%arg0: tensor<16x16xf16, #triton_gpu.dot_op<{opIdx = 0, parent = #mma}>>, %arg1: tensor<16x16xf16, #triton_gpu.dot_op<{opIdx = 1, parent = #mma}>>, %arg2: tensor<16x16xf16, #mma>) { | ||
| // CHECK-COUNT-2: llvm.extractvalue %{{.*}} : !llvm.struct<(f16)> | ||
| // CHECK-COUNT-8: llvm.extractvalue %{{.*}} : !llvm.struct<(f16, f16, f16, f16, f16, f16, f16, f16)> | ||
| // CHECK: llvm.mlir.undef : vector<16xf16> | ||
| // CHECK-COUNT-8: llvm.insertelement {{.*}} : vector<16xf16> | ||
| // CHECK: rocdl.wmma.f16.16x16x16.f16 {{.*}} : (f16, f16, vector<16xf16>, i1) -> vector<16xf16> | ||
| %0 = tt.dot %arg0, %arg1, %arg2 {allowTF32 = false, maxNumImpreciseAcc = 0 : i32} : tensor<16x16xf16, #triton_gpu.dot_op<{opIdx = 0, parent = #mma}>> * tensor<16x16xf16, #triton_gpu.dot_op<{opIdx = 1, parent = #mma}>> -> tensor<16x16xf16, #mma> | ||
| // CHECK-COUNT-8: llvm.extractelement {{.*}} : vector<16xf16> | ||
| // CHECK: llvm.mlir.undef : !llvm.struct<(f16, f16, f16, f16, f16, f16, f16, f16)> | ||
| // CHECK-COUNT-8: llvm.insertvalue {{.*}} : !llvm.struct<(f16, f16, f16, f16, f16, f16, f16, f16)> | ||
| tt.return | ||
| } | ||
| } |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
245 changes: 245 additions & 0 deletions
245
third_party/amd/lib/TritonAMDGPUToLLVM/DotOpToLLVM/WMMA.cpp
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,245 @@ | ||
| /* | ||
| * Copyright (c) 2024, Advanced Micro Devices, Inc. All rights reserved. | ||
| * | ||
| * Permission is hereby granted, free of charge, to any person obtaining | ||
| * a copy of this software and associated documentation files | ||
| * (the "Software"), to deal in the Software without restriction, | ||
| * including without limitation the rights to use, copy, modify, merge, | ||
| * publish, distribute, sublicense, and/or sell copies of the Software, | ||
| * and to permit persons to whom the Software is furnished to do so, | ||
| * subject to the following conditions: | ||
| * | ||
| * The above copyright notice and this permission notice shall be | ||
| * included in all copies or substantial portions of the Software. | ||
| * | ||
| * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, | ||
| * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF | ||
| * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. | ||
| * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY | ||
| * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, | ||
| * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE | ||
| * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. | ||
| */ | ||
|
|
||
| #include "../DotOpToLLVM.h" | ||
| #include "Utility.h" | ||
|
|
||
| #include "mlir/Dialect/LLVMIR/ROCDLDialect.h" | ||
|
|
||
| using namespace mlir; | ||
| using namespace mlir::triton; | ||
|
|
||
| namespace AMD { | ||
| namespace { | ||
|
|
||
| using ::AMD::TritonGPUToLLVMTypeConverter; | ||
| using ::mlir::triton::gpu::AMDWmmaEncodingAttr; | ||
| using ::mlir::triton::gpu::DotOperandEncodingAttr; | ||
| using ::mlir::triton::gpu::SharedEncodingAttr; | ||
|
|
||
| enum class WMMAInstrType : uint8_t { | ||
| // D = AB + C; | ||
| // typeof(D) == typeof(C) | ||
| // typeof(A) == typeof(B) | ||
| // typeof(D), typeof(A): | ||
| FP32_FP16, | ||
| FP32_BF16, | ||
| FP16_FP16, | ||
| BF16_BF16, | ||
| INT32_IU8, | ||
| INT32_IU4, | ||
| NOT_APPLICABLE, | ||
| }; | ||
|
|
||
| using ValueTable = std::map<std::pair<unsigned, unsigned>, Value>; | ||
|
|
||
| ValueTable | ||
| getValuesFromDotOperandLayoutStruct(ConversionPatternRewriter &rewriter, | ||
| TritonGPUToLLVMTypeConverter *typeConverter, | ||
| Value value, int n0, int n1, Type type, | ||
| Location loc) { | ||
| auto elems = typeConverter->unpackLLElements(loc, value, rewriter); | ||
| ValueTable vals; | ||
| for (int i = 0; i < n0; i++) { | ||
| for (int j = 0; j < n1; j++) { | ||
| vals[{i, j}] = elems[n1 * i + j]; | ||
| } | ||
| } | ||
| return vals; | ||
| } | ||
|
|
||
| static WMMAInstrType getWMMAInstrTypeFromDot(DotOp op) { | ||
| auto aOperandTy = op.getA().getType(); | ||
| auto aTensorTy = aOperandTy.cast<RankedTensorType>(); | ||
| auto aElemTy = aTensorTy.getElementType(); | ||
| auto bOperandTy = op.getB().getType(); | ||
| auto bTensorTy = bOperandTy.cast<RankedTensorType>(); | ||
| auto bElemTy = bTensorTy.getElementType(); | ||
| assert(aElemTy == bElemTy); | ||
| auto cOperandTy = op.getC().getType(); | ||
| auto cTensorTy = cOperandTy.cast<RankedTensorType>(); | ||
| auto cElemTy = cTensorTy.getElementType(); | ||
| auto dOperandTy = op.getD().getType(); | ||
| auto dTensorTy = dOperandTy.cast<RankedTensorType>(); | ||
| auto dElemTy = dTensorTy.getElementType(); | ||
| assert(cElemTy == dElemTy); | ||
|
|
||
| if (dElemTy.isF32() && aElemTy.isF16()) | ||
| return WMMAInstrType::FP32_FP16; | ||
| if (dElemTy.isF32() && aElemTy.isBF16()) | ||
| return WMMAInstrType::FP32_BF16; | ||
| if (dElemTy.isF16() && aElemTy.isF16()) | ||
| return WMMAInstrType::FP16_FP16; | ||
| if (dElemTy.isBF16() && aElemTy.isBF16()) | ||
| return WMMAInstrType::BF16_BF16; | ||
| if (dElemTy.isSignedInteger(32) && aElemTy.isUnsignedInteger(8)) | ||
| return WMMAInstrType::INT32_IU8; | ||
| if (dElemTy.isSignedInteger(32) && aElemTy.isUnsignedInteger(4)) | ||
| return WMMAInstrType::INT32_IU4; | ||
|
|
||
| return WMMAInstrType::NOT_APPLICABLE; | ||
| } | ||
|
|
||
| Value generateWMMAOp(ConversionPatternRewriter &rewriter, Location loc, | ||
| WMMAInstrType wmmaType, Value valA, Value valB, | ||
| Value valC) { | ||
| auto resType = valC.getType(); | ||
| Value falseFlag = int_val(1, false); | ||
| switch (wmmaType) { | ||
| case WMMAInstrType::FP32_FP16: | ||
| return rewriter.create<ROCDL::wmma_f32_16x16x16_f16>( | ||
| loc, TypeRange{resType}, ValueRange{valA, valB, valC}); | ||
| case WMMAInstrType::FP32_BF16: | ||
| return rewriter.create<ROCDL::wmma_f32_16x16x16_bf16>( | ||
| loc, TypeRange{resType}, ValueRange{valA, valB, valC}); | ||
| case WMMAInstrType::FP16_FP16: | ||
| return rewriter.create<ROCDL::wmma_f16_16x16x16_f16>( | ||
| loc, TypeRange{resType}, ValueRange{valA, valB, valC, falseFlag}); | ||
| case WMMAInstrType::BF16_BF16: | ||
| return rewriter.create<ROCDL::wmma_bf16_16x16x16_bf16>( | ||
| loc, TypeRange{resType}, ValueRange{valA, valB, valC, falseFlag}); | ||
| case WMMAInstrType::INT32_IU8: | ||
| return rewriter.create<ROCDL::wmma_i32_16x16x16_iu8>( | ||
| loc, TypeRange{resType}, ValueRange{valA, valB, valC, falseFlag}); | ||
| case WMMAInstrType::INT32_IU4: | ||
| return rewriter.create<ROCDL::wmma_i32_16x16x16_iu4>( | ||
| loc, TypeRange{resType}, ValueRange{valA, valB, valC, falseFlag}); | ||
| default: | ||
| llvm::report_fatal_error("WMMA data type not supported"); | ||
| } | ||
| return Value(); | ||
| } | ||
|
|
||
| // Conduct the Dot conversion. | ||
| LogicalResult convertDot(DotOp op, DotOpAdaptor adaptor, | ||
| ConversionPatternRewriter &rewriter, | ||
| TritonGPUToLLVMTypeConverter *typeConverter) { | ||
| auto wmmaLayout = op.getResult() | ||
| .getType() | ||
| .cast<RankedTensorType>() | ||
| .getEncoding() | ||
| .cast<AMDWmmaEncodingAttr>(); | ||
| auto warpsPerCTA = wmmaLayout.getWarpsPerCTA(); | ||
| auto mnkDim = AMDWmmaEncodingAttr::getMNKDimPerWMMAInstr(); | ||
| auto wmmaInstrType = getWMMAInstrTypeFromDot(op); | ||
|
|
||
| auto loc = op.getLoc(); | ||
| Value a = op.getA(); | ||
| Value b = op.getB(); | ||
| Value d = op.getD(); | ||
| auto aTensorTy = a.getType().cast<RankedTensorType>(); | ||
| auto bTensorTy = b.getType().cast<RankedTensorType>(); | ||
| auto dTensorTy = d.getType().cast<RankedTensorType>(); | ||
| auto elemTy = aTensorTy.getElementType(); | ||
|
|
||
| auto aEncoding = aTensorTy.getEncoding().cast<DotOperandEncodingAttr>(); | ||
| auto bEncoding = bTensorTy.getEncoding().cast<DotOperandEncodingAttr>(); | ||
| int kWidth = aEncoding.getKWidth(); | ||
|
|
||
| auto repA = | ||
| wmmaLayout.getWMMARepForOperands(aTensorTy.getShape(), elemTy, kWidth, 0); | ||
| auto repB = | ||
| wmmaLayout.getWMMARepForOperands(bTensorTy.getShape(), elemTy, kWidth, 1); | ||
|
|
||
| assert(repA[1] == repB[0]); | ||
|
|
||
| Value loadedA = adaptor.getA(); | ||
| Value loadedB = adaptor.getB(); | ||
| Value loadedC = adaptor.getC(); | ||
| auto numRepM = repA[0]; | ||
| auto numRepN = repB[1]; | ||
| auto numRepK = repA[1]; | ||
|
|
||
| ValueTable ha = getValuesFromDotOperandLayoutStruct( | ||
| rewriter, typeConverter, loadedA, numRepM, numRepK, | ||
| aTensorTy.getElementType(), loc); | ||
| ValueTable hb = getValuesFromDotOperandLayoutStruct( | ||
| rewriter, typeConverter, loadedB, numRepN, numRepK, | ||
| aTensorTy.getElementType(), loc); | ||
| auto dstElemTy = dTensorTy.getElementType(); | ||
| auto fc = typeConverter->unpackLLElements(loc, loadedC, rewriter); | ||
|
|
||
| unsigned warpSize = triton::gpu::getWarpSize(wmmaLayout); | ||
| // TODO get rid of magic numbers | ||
| unsigned vgprElemWidth = 32; | ||
| unsigned paddedOutputElemSize = | ||
| vgprElemWidth / dstElemTy.getIntOrFloatBitWidth(); | ||
| // compute number of output elements that each thread holds for one WMMA | ||
| // instruction. | ||
| auto elemsPerVec = mnkDim[0] * mnkDim[1] * paddedOutputElemSize / warpSize; | ||
| auto dElemsToStorePerThread = mnkDim[0] * mnkDim[1] / warpSize; | ||
| auto vecTy = vec_ty(dstElemTy, elemsPerVec); | ||
| for (int m = 0; m < numRepM; ++m) { | ||
| for (int n = 0; n < numRepN; ++n) { | ||
| Value acc = undef(vecTy); | ||
| for (unsigned v = 0; v < dElemsToStorePerThread; ++v) { | ||
| acc = insert_element(vecTy, acc, | ||
| fc[m * numRepN * dElemsToStorePerThread + | ||
| n * dElemsToStorePerThread + v], | ||
| i32_val(v * paddedOutputElemSize)); | ||
| } | ||
| for (size_t k = 0; k < numRepK; k++) { | ||
| acc = generateWMMAOp(rewriter, loc, wmmaInstrType, ha[{m, k}], | ||
| hb[{n, k}], acc); | ||
| } | ||
| for (unsigned v = 0; v < dElemsToStorePerThread; ++v) { | ||
| fc[m * numRepN * dElemsToStorePerThread + n * dElemsToStorePerThread + | ||
| v] = | ||
| extract_element(dstElemTy, acc, i32_val(v * paddedOutputElemSize)); | ||
| } | ||
| } | ||
| } | ||
|
|
||
| // replace with new packed result | ||
| Type structTy = LLVM::LLVMStructType::getLiteral( | ||
| wmmaLayout.getContext(), SmallVector<Type>(fc.size(), dstElemTy)); | ||
| Value res = typeConverter->packLLElements(loc, fc, rewriter, structTy); | ||
| rewriter.replaceOp(op, res); | ||
| return success(); | ||
| } | ||
|
|
||
| } // namespace | ||
|
|
||
| LogicalResult convertWMMA(triton::DotOp op, triton::DotOp::Adaptor adaptor, | ||
| TritonGPUToLLVMTypeConverter *typeConverter, | ||
| ConversionPatternRewriter &rewriter) { | ||
| auto rankedTType = [](Value tensor) { | ||
| return tensor.getType().cast<RankedTensorType>(); | ||
| }; | ||
|
|
||
| assert(rankedTType(op.getA()).getEncoding().isa<DotOperandEncodingAttr>() && | ||
| rankedTType(op.getB()).getEncoding().isa<DotOperandEncodingAttr>() && | ||
| "Both $a and %b should be DotOperand layout."); | ||
|
|
||
| auto cTensorTy = rankedTType(op.getC()); | ||
| auto dTensorTy = rankedTType(op.getD()); | ||
| assert(cTensorTy.getEncoding().isa<AMDWmmaEncodingAttr>() && | ||
| "Currently, we only support $c with a wmma layout."); | ||
|
|
||
| assert(cTensorTy.getShape()[0] == dTensorTy.getShape()[0] && | ||
| cTensorTy.getShape()[1] == dTensorTy.getShape()[1] && | ||
| "DotOp's $c operand should pass the same number of values as $d"); | ||
|
|
||
| return convertDot(op, adaptor, rewriter, typeConverter); | ||
| } | ||
| } // namespace AMD |