e1da62910e
- Define a gpu.printf op, which can be lowered to any GPU printf() support (which is present in CUDA, HIP, and OpenCL). This op only supports constant format strings and scalar arguments - Define the lowering of gpu.pirntf to a call to printf() (which is what is required for AMD GPUs when using OpenCL) as well as to the hostcall interface present in the AMD Open Compute device library, which is the interface present when kernels are running under HIP. - Add a "runtime" enum that allows specifying which of the possible runtimes a ROCDL kernel will be executed under or that the runtime is unknown. This enum controls how gpu.printf is lowered This change does not enable lowering for Nvidia GPUs, but such a lowering should be possible in principle. And: [MLIR][AMDGPU] Always set amdgpu-implicitarg-num-bytes=56 on kernels This is something that Clang always sets on both OpenCL and HIP kernels, and failing to include it causes mysterious crashes with printf() support. In addition, revert the max-flat-work-group-size to (1, 256) to avoid triggering bugs in the AMDGPU backend. Reviewed By: mehdi_amini Differential Revision: https://reviews.llvm.org/D110448
176 lines
8.2 KiB
C++
176 lines
8.2 KiB
C++
//===- LowerGpuOpsToROCDLOps.cpp - MLIR GPU to ROCDL lowering passes ------===//
|
|
//
|
|
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
|
|
// See https://llvm.org/LICENSE.txt for license information.
|
|
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
//
|
|
// This file implements a pass to generate ROCDLIR operations for higher-level
|
|
// GPU operations.
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
#include "mlir/Conversion/GPUToROCDL/GPUToROCDLPass.h"
|
|
|
|
#include "mlir/Conversion/ArithmeticToLLVM/ArithmeticToLLVM.h"
|
|
#include "mlir/Conversion/LLVMCommon/ConversionTarget.h"
|
|
#include "mlir/Conversion/LLVMCommon/LoweringOptions.h"
|
|
#include "mlir/Conversion/LLVMCommon/TypeConverter.h"
|
|
#include "mlir/Conversion/MemRefToLLVM/MemRefToLLVM.h"
|
|
#include "mlir/Conversion/StandardToLLVM/ConvertStandardToLLVM.h"
|
|
#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h"
|
|
#include "mlir/Conversion/VectorToROCDL/VectorToROCDL.h"
|
|
#include "mlir/Dialect/GPU/GPUDialect.h"
|
|
#include "mlir/Dialect/GPU/Passes.h"
|
|
#include "mlir/Dialect/LLVMIR/ROCDLDialect.h"
|
|
#include "mlir/Dialect/Math/IR/Math.h"
|
|
#include "mlir/Dialect/Vector/VectorOps.h"
|
|
#include "mlir/Pass/Pass.h"
|
|
#include "mlir/Transforms/DialectConversion.h"
|
|
#include "mlir/Transforms/GreedyPatternRewriteDriver.h"
|
|
#include "llvm/Support/FormatVariadic.h"
|
|
|
|
#include "../GPUCommon/GPUOpsLowering.h"
|
|
#include "../GPUCommon/IndexIntrinsicsOpLowering.h"
|
|
#include "../GPUCommon/OpToFuncCallLowering.h"
|
|
#include "../PassDetail.h"
|
|
|
|
using namespace mlir;
|
|
|
|
namespace {
|
|
|
|
/// Import the GPU Ops to ROCDL Patterns.
|
|
#include "GPUToROCDL.cpp.inc"
|
|
|
|
// A pass that replaces all occurrences of GPU device operations with their
|
|
// corresponding ROCDL equivalent.
|
|
//
|
|
// This pass only handles device code and is not meant to be run on GPU host
|
|
// code.
|
|
struct LowerGpuOpsToROCDLOpsPass
|
|
: public ConvertGpuOpsToROCDLOpsBase<LowerGpuOpsToROCDLOpsPass> {
|
|
LowerGpuOpsToROCDLOpsPass() = default;
|
|
LowerGpuOpsToROCDLOpsPass(unsigned indexBitwidth, gpu::amd::Runtime runtime) {
|
|
this->indexBitwidth = indexBitwidth;
|
|
this->runtime = runtime;
|
|
}
|
|
|
|
void runOnOperation() override {
|
|
gpu::GPUModuleOp m = getOperation();
|
|
|
|
/// Customize the bitwidth used for the device side index computations.
|
|
LowerToLLVMOptions options(
|
|
m.getContext(),
|
|
DataLayout(cast<DataLayoutOpInterface>(m.getOperation())));
|
|
options.emitCWrappers = true;
|
|
if (indexBitwidth != kDeriveIndexBitwidthFromDataLayout)
|
|
options.overrideIndexBitwidth(indexBitwidth);
|
|
LLVMTypeConverter converter(m.getContext(), options);
|
|
|
|
RewritePatternSet patterns(m.getContext());
|
|
RewritePatternSet llvmPatterns(m.getContext());
|
|
|
|
populateGpuRewritePatterns(patterns);
|
|
(void)applyPatternsAndFoldGreedily(m, std::move(patterns));
|
|
|
|
mlir::arith::populateArithmeticToLLVMConversionPatterns(converter,
|
|
llvmPatterns);
|
|
populateVectorToLLVMConversionPatterns(converter, llvmPatterns);
|
|
populateVectorToROCDLConversionPatterns(converter, llvmPatterns);
|
|
populateStdToLLVMConversionPatterns(converter, llvmPatterns);
|
|
populateMemRefToLLVMConversionPatterns(converter, llvmPatterns);
|
|
populateGpuToROCDLConversionPatterns(converter, llvmPatterns, runtime);
|
|
LLVMConversionTarget target(getContext());
|
|
configureGpuToROCDLConversionLegality(target);
|
|
if (failed(applyPartialConversion(m, target, std::move(llvmPatterns))))
|
|
signalPassFailure();
|
|
}
|
|
};
|
|
|
|
} // namespace
|
|
|
|
void mlir::configureGpuToROCDLConversionLegality(ConversionTarget &target) {
|
|
target.addIllegalOp<FuncOp>();
|
|
target.addLegalDialect<::mlir::LLVM::LLVMDialect>();
|
|
target.addLegalDialect<ROCDL::ROCDLDialect>();
|
|
target.addIllegalDialect<gpu::GPUDialect>();
|
|
target.addIllegalOp<LLVM::CosOp, LLVM::ExpOp, LLVM::Exp2Op, LLVM::FAbsOp,
|
|
LLVM::FCeilOp, LLVM::FFloorOp, LLVM::LogOp, LLVM::Log10Op,
|
|
LLVM::Log2Op, LLVM::PowOp, LLVM::SinOp, LLVM::SqrtOp>();
|
|
|
|
// TODO: Remove once we support replacing non-root ops.
|
|
target.addLegalOp<gpu::YieldOp, gpu::GPUModuleOp, gpu::ModuleEndOp>();
|
|
}
|
|
|
|
void mlir::populateGpuToROCDLConversionPatterns(
|
|
LLVMTypeConverter &converter, RewritePatternSet &patterns,
|
|
mlir::gpu::amd::Runtime runtime) {
|
|
using mlir::gpu::amd::Runtime;
|
|
|
|
populateWithGenerated(patterns);
|
|
patterns
|
|
.add<GPUIndexIntrinsicOpLowering<gpu::ThreadIdOp, ROCDL::ThreadIdXOp,
|
|
ROCDL::ThreadIdYOp, ROCDL::ThreadIdZOp>,
|
|
GPUIndexIntrinsicOpLowering<gpu::BlockDimOp, ROCDL::BlockDimXOp,
|
|
ROCDL::BlockDimYOp, ROCDL::BlockDimZOp>,
|
|
GPUIndexIntrinsicOpLowering<gpu::BlockIdOp, ROCDL::BlockIdXOp,
|
|
ROCDL::BlockIdYOp, ROCDL::BlockIdZOp>,
|
|
GPUIndexIntrinsicOpLowering<gpu::GridDimOp, ROCDL::GridDimXOp,
|
|
ROCDL::GridDimYOp, ROCDL::GridDimZOp>,
|
|
GPUReturnOpLowering>(converter);
|
|
patterns.add<GPUFuncOpLowering>(
|
|
converter, /*allocaAddrSpace=*/5,
|
|
StringAttr::get(&converter.getContext(),
|
|
ROCDL::ROCDLDialect::getKernelFuncAttrName()));
|
|
if (Runtime::HIP == runtime) {
|
|
patterns.add<GPUPrintfOpToHIPLowering>(converter);
|
|
} else if (Runtime::OpenCL == runtime) {
|
|
// Use address space = 4 to match the OpenCL definition of printf()
|
|
patterns.add<GPUPrintfOpToLLVMCallLowering>(converter, /*addressSpace=*/4);
|
|
}
|
|
|
|
patterns.add<OpToFuncCallLowering<math::AbsOp>>(converter, "__ocml_fabs_f32",
|
|
"__ocml_fabs_f64");
|
|
patterns.add<OpToFuncCallLowering<math::AtanOp>>(converter, "__ocml_atan_f32",
|
|
"__ocml_atan_f64");
|
|
patterns.add<OpToFuncCallLowering<math::Atan2Op>>(
|
|
converter, "__ocml_atan2_f32", "__ocml_atan2_f64");
|
|
patterns.add<OpToFuncCallLowering<math::CeilOp>>(converter, "__ocml_ceil_f32",
|
|
"__ocml_ceil_f64");
|
|
patterns.add<OpToFuncCallLowering<math::CosOp>>(converter, "__ocml_cos_f32",
|
|
"__ocml_cos_f64");
|
|
patterns.add<OpToFuncCallLowering<math::ExpOp>>(converter, "__ocml_exp_f32",
|
|
"__ocml_exp_f64");
|
|
patterns.add<OpToFuncCallLowering<math::Exp2Op>>(converter, "__ocml_exp2_f32",
|
|
"__ocml_exp2_f64");
|
|
patterns.add<OpToFuncCallLowering<math::ExpM1Op>>(
|
|
converter, "__ocml_expm1_f32", "__ocml_expm1_f64");
|
|
patterns.add<OpToFuncCallLowering<math::FloorOp>>(
|
|
converter, "__ocml_floor_f32", "__ocml_floor_f64");
|
|
patterns.add<OpToFuncCallLowering<math::LogOp>>(converter, "__ocml_log_f32",
|
|
"__ocml_log_f64");
|
|
patterns.add<OpToFuncCallLowering<math::Log10Op>>(
|
|
converter, "__ocml_log10_f32", "__ocml_log10_f64");
|
|
patterns.add<OpToFuncCallLowering<math::Log1pOp>>(
|
|
converter, "__ocml_log1p_f32", "__ocml_log1p_f64");
|
|
patterns.add<OpToFuncCallLowering<math::Log2Op>>(converter, "__ocml_log2_f32",
|
|
"__ocml_log2_f64");
|
|
patterns.add<OpToFuncCallLowering<math::PowFOp>>(converter, "__ocml_pow_f32",
|
|
"__ocml_pow_f64");
|
|
patterns.add<OpToFuncCallLowering<math::RsqrtOp>>(
|
|
converter, "__ocml_rsqrt_f32", "__ocml_rsqrt_f64");
|
|
patterns.add<OpToFuncCallLowering<math::SinOp>>(converter, "__ocml_sin_f32",
|
|
"__ocml_sin_f64");
|
|
patterns.add<OpToFuncCallLowering<math::SqrtOp>>(converter, "__ocml_sqrt_f32",
|
|
"__ocml_sqrt_f64");
|
|
patterns.add<OpToFuncCallLowering<math::TanhOp>>(converter, "__ocml_tanh_f32",
|
|
"__ocml_tanh_f64");
|
|
}
|
|
|
|
std::unique_ptr<OperationPass<gpu::GPUModuleOp>>
|
|
mlir::createLowerGpuOpsToROCDLOpsPass(unsigned indexBitwidth,
|
|
gpu::amd::Runtime runtime) {
|
|
return std::make_unique<LowerGpuOpsToROCDLOpsPass>(indexBitwidth, runtime);
|
|
}
|