intel/llvm
1
0
Fork 0
mirror of https://github.com/intel/llvm.git synced 2026-02-04 20:00:11 +08:00
Code Issues Packages Projects Releases Wiki Activity

Add a function to get lowering patterns from GPU to NVVM.

Browse Source 
This enables combining the patterns with other patterns into larger lowerings.

PiperOrigin-RevId: 284979271
This commit is contained in:
Stephan Herhut
2019-12-11 07:13:54 -08:00
committed by A. Unique TensorFlower
parent bae8a7a724
commit b96f86daaf
2 changed files with 22 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
mlir/include/mlir/Conversion/GPUToNVVM/GPUToNVVMPass.h
View File

@@ -26,6 +26,10 @@ class OwningRewritePatternList;
class ModuleOp;
template <typename OpT> class OpPassBase;
/// Collect a set of patterns to convert from the GPU dialect to NVVM.
void populateGpuToNVVMConversionPatterns(LLVMTypeConverter &converter,
OwningRewritePatternList &patterns);
/// Creates a pass that lowers GPU dialect operations to NVVM counterparts.
std::unique_ptr<OpPassBase<ModuleOp>> createLowerGpuOpsToNVVMOpsPass();

31
mlir/lib/Conversion/GPUToNVVM/LowerGpuOpsToNVVMOps.cpp
View File

@@ -605,19 +605,7 @@ public:
OwningRewritePatternList patterns;
LLVMTypeConverter converter(m.getContext());
populateStdToLLVMConversionPatterns(converter, patterns);
populateWithGenerated(&getContext(), &patterns);
patterns.insert<
GPUIndexIntrinsicOpLowering<gpu::ThreadIdOp, NVVM::ThreadIdXOp,
NVVM::ThreadIdYOp, NVVM::ThreadIdZOp>,
GPUIndexIntrinsicOpLowering<gpu::BlockDimOp, NVVM::BlockDimXOp,
NVVM::BlockDimYOp, NVVM::BlockDimZOp>,
GPUIndexIntrinsicOpLowering<gpu::BlockIdOp, NVVM::BlockIdXOp,
NVVM::BlockIdYOp, NVVM::BlockIdZOp>,
GPUIndexIntrinsicOpLowering<gpu::GridDimOp, NVVM::GridDimXOp,
NVVM::GridDimYOp, NVVM::GridDimZOp>,
GPUAllReduceOpLowering, FuncOpLowering>(converter);
patterns.insert<OpToFuncCallLowering<ExpOp>>(converter, "__nv_expf",
"__nv_exp");
populateGpuToNVVMConversionPatterns(converter, patterns);
ConversionTarget target(getContext());
target.addIllegalDialect<gpu::GPUDialect>();
target.addIllegalOp<LLVM::ExpOp>();
@@ -633,6 +621,23 @@ public:
} // anonymous namespace
void mlir::populateGpuToNVVMConversionPatterns(
LLVMTypeConverter &converter, OwningRewritePatternList &patterns) {
populateWithGenerated(converter.getDialect()->getContext(), &patterns);
patterns
.insert<GPUIndexIntrinsicOpLowering<gpu::ThreadIdOp, NVVM::ThreadIdXOp,
NVVM::ThreadIdYOp, NVVM::ThreadIdZOp>,
GPUIndexIntrinsicOpLowering<gpu::BlockDimOp, NVVM::BlockDimXOp,
NVVM::BlockDimYOp, NVVM::BlockDimZOp>,
GPUIndexIntrinsicOpLowering<gpu::BlockIdOp, NVVM::BlockIdXOp,
NVVM::BlockIdYOp, NVVM::BlockIdZOp>,
GPUIndexIntrinsicOpLowering<gpu::GridDimOp, NVVM::GridDimXOp,
NVVM::GridDimYOp, NVVM::GridDimZOp>,
GPUAllReduceOpLowering, FuncOpLowering>(converter);
patterns.insert<OpToFuncCallLowering<ExpOp>>(converter, "__nv_expf",
"__nv_exp");
}
std::unique_ptr<OpPassBase<ModuleOp>> mlir::createLowerGpuOpsToNVVMOpsPass() {
return std::make_unique<LowerGpuOpsToNVVMOpsPass>();
}