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.

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

[AMD] ElementwiseOpToLLVM: Do not convert types if they are equal #3091

Merged
merged 5 commits into from
Mar 7, 2024
Merged

[AMD] ElementwiseOpToLLVM: Do not convert types if they are equal #3091

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
9bc9a24
[AMD] ElementwiseOpToLLVM: Do not convert types if they are equal
joviliast Feb 7, 2024
affffb5
[AMD] Add mixed fma dot operation lit test
joviliast Mar 5, 2024
01a8a87
Merge branch 'main' into fma-matmul
joviliast Mar 6, 2024
7f96f22
Merge branch 'main' into fma-matmul
joviliast Mar 6, 2024
bfde33e
Merge branch 'main' into fma-matmul
joviliast Mar 7, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 11 additions & 0 deletions test/Conversion/amd/fp_to_fp.mlir
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,11 @@
// RUN: triton-opt %s --split-input-file --convert-triton-amdgpu-to-llvm | FileCheck %s

// CHECK-LABEL: f16_to_f32
#blocked = #triton_gpu.blocked<{sizePerThread = [1, 8], threadsPerWarp = [4, 8], warpsPerCTA = [4, 1], order = [1, 0]}>
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 @f16_to_f32(%arg0: tensor<8x8xf16, #triton_gpu.dot_op<{opIdx = 0, parent = #blocked}>>) {
// CHECK-COUNT-8: llvm.inline_asm asm_dialect {{.*}}v_cvt_f32_f16 {{.*}}: (f16) -> f32
%0 = tt.fp_to_fp %arg0 : tensor<8x8xf16, #triton_gpu.dot_op<{opIdx = 0, parent = #blocked}>> -> tensor<8x8xf32, #triton_gpu.dot_op<{opIdx = 0, parent = #blocked}>>
tt.return
}
}
10 changes: 8 additions & 2 deletions third_party/amd/lib/TritonAMDGPUToLLVM/ElementwiseOpToLLVM.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1230,7 +1230,6 @@ struct FpToFpOpConversion
bool isDstFP32 = dstElementType.isF32();
Type srcType = useFP16IntermediateSrc ? f16_ty : srcElementType;
Type dstType = isDstFP32 ? f16_ty : dstElementType;
auto cvtFunc = getConversionFunc(srcType, dstType);
SmallVector<Value> inVals;
for (unsigned i = 0; i < std::min(numElements, operands.size()); i++) {
inVals.push_back(operands[i][0]);
Expand All @@ -1239,7 +1238,14 @@ struct FpToFpOpConversion
for (Value &v : inVals)
v = convertFp32ToFp16NZ(loc, rewriter, v);
inVals.resize(numElements, undef(typeConverter->convertType(srcType)));
SmallVector<Value> outVals = cvtFunc(loc, rewriter, inVals);
SmallVector<Value> outVals;
if (srcType != dstType) {
auto cvtFunc = getConversionFunc(srcType, dstType);
outVals = cvtFunc(loc, rewriter, inVals);
} else {
outVals = inVals;
}

assert(outVals.size() == inVals.size());
outVals.resize(std::min(numElements, operands.size()));
if (isDstFP32)
Expand Down
Loading