d45de8003a
Add support in `-convert-gpu-to-llvm-spv` to convert `gpu.func` to `llvm.func` operations. - `spir_kernel`/`spir_func` calling conventions used for kernels/functions. - `workgroup` attributions encoded as additional `llvm.ptr<3>` arguments. - No attribute used to annotate kernels - `reqd_work_group_size` attribute using to encode `gpu.known_block_size`. - `llvm.mlir.workgroup_attrib_size` used to encode workgroup attribution sizes. This will be attached to the pointer argument workgroup attributions lower to. **Note**: A notable missing feature that will be addressed in a follow-up PR is a `-use-bare-ptr-memref-call-conv` option to replace MemRef arguments with bare pointers to the MemRef element types instead of the current MemRef descriptor approach. --------- Signed-off-by: Victor Perez <victor.perez@codeplay.com>
176 lines
7.0 KiB
C++
176 lines
7.0 KiB
C++
//===- GPUOpsLowering.h - GPU FuncOp / ReturnOp lowering -------*- C++ -*--===//
|
|
//
|
|
// 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
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
#ifndef MLIR_CONVERSION_GPUCOMMON_GPUOPSLOWERING_H_
|
|
#define MLIR_CONVERSION_GPUCOMMON_GPUOPSLOWERING_H_
|
|
|
|
#include "mlir/Conversion/LLVMCommon/Pattern.h"
|
|
#include "mlir/Dialect/GPU/IR/GPUDialect.h"
|
|
#include "mlir/Dialect/LLVMIR/LLVMDialect.h"
|
|
|
|
namespace mlir {
|
|
|
|
/// Lowering for gpu.dynamic.shared.memory to LLVM dialect. The pattern first
|
|
/// create a 0-sized global array symbol similar as LLVM expects. It constructs
|
|
/// a memref descriptor with these values and return it.
|
|
struct GPUDynamicSharedMemoryOpLowering
|
|
: public ConvertOpToLLVMPattern<gpu::DynamicSharedMemoryOp> {
|
|
using ConvertOpToLLVMPattern<
|
|
gpu::DynamicSharedMemoryOp>::ConvertOpToLLVMPattern;
|
|
GPUDynamicSharedMemoryOpLowering(const LLVMTypeConverter &converter,
|
|
unsigned alignmentBit = 0)
|
|
: ConvertOpToLLVMPattern<gpu::DynamicSharedMemoryOp>(converter),
|
|
alignmentBit(alignmentBit) {}
|
|
|
|
LogicalResult
|
|
matchAndRewrite(gpu::DynamicSharedMemoryOp op, OpAdaptor adaptor,
|
|
ConversionPatternRewriter &rewriter) const override;
|
|
|
|
private:
|
|
// Alignment bit
|
|
unsigned alignmentBit;
|
|
};
|
|
|
|
struct GPUFuncOpLoweringOptions {
|
|
/// The address space to use for `alloca`s in private memory.
|
|
unsigned allocaAddrSpace;
|
|
/// The address space to use declaring workgroup memory.
|
|
unsigned workgroupAddrSpace;
|
|
|
|
/// The attribute name to use instead of `gpu.kernel`. Null if no attribute
|
|
/// should be used.
|
|
StringAttr kernelAttributeName;
|
|
/// The attribute name to to set block size. Null if no attribute should be
|
|
/// used.
|
|
StringAttr kernelBlockSizeAttributeName;
|
|
|
|
/// The calling convention to use for kernel functions.
|
|
LLVM::CConv kernelCallingConvention = LLVM::CConv::C;
|
|
/// The calling convention to use for non-kernel functions.
|
|
LLVM::CConv nonKernelCallingConvention = LLVM::CConv::C;
|
|
|
|
/// Whether to encode workgroup attributions as additional arguments instead
|
|
/// of a global variable.
|
|
bool encodeWorkgroupAttributionsAsArguments = false;
|
|
};
|
|
|
|
struct GPUFuncOpLowering : ConvertOpToLLVMPattern<gpu::GPUFuncOp> {
|
|
GPUFuncOpLowering(const LLVMTypeConverter &converter,
|
|
const GPUFuncOpLoweringOptions &options)
|
|
: ConvertOpToLLVMPattern<gpu::GPUFuncOp>(converter),
|
|
allocaAddrSpace(options.allocaAddrSpace),
|
|
workgroupAddrSpace(options.workgroupAddrSpace),
|
|
kernelAttributeName(options.kernelAttributeName),
|
|
kernelBlockSizeAttributeName(options.kernelBlockSizeAttributeName),
|
|
kernelCallingConvention(options.kernelCallingConvention),
|
|
nonKernelCallingConvention(options.nonKernelCallingConvention),
|
|
encodeWorkgroupAttributionsAsArguments(
|
|
options.encodeWorkgroupAttributionsAsArguments) {}
|
|
|
|
LogicalResult
|
|
matchAndRewrite(gpu::GPUFuncOp gpuFuncOp, OpAdaptor adaptor,
|
|
ConversionPatternRewriter &rewriter) const override;
|
|
|
|
private:
|
|
/// The address space to use for `alloca`s in private memory.
|
|
unsigned allocaAddrSpace;
|
|
/// The address space to use declaring workgroup memory.
|
|
unsigned workgroupAddrSpace;
|
|
|
|
/// The attribute name to use instead of `gpu.kernel`. Null if no attribute
|
|
/// should be used.
|
|
StringAttr kernelAttributeName;
|
|
/// The attribute name to to set block size. Null if no attribute should be
|
|
/// used.
|
|
StringAttr kernelBlockSizeAttributeName;
|
|
|
|
/// The calling convention to use for kernel functions
|
|
LLVM::CConv kernelCallingConvention;
|
|
/// The calling convention to use for non-kernel functions
|
|
LLVM::CConv nonKernelCallingConvention;
|
|
|
|
/// Whether to encode workgroup attributions as additional arguments instead
|
|
/// of a global variable.
|
|
bool encodeWorkgroupAttributionsAsArguments;
|
|
};
|
|
|
|
/// The lowering of gpu.printf to a call to HIP hostcalls
|
|
///
|
|
/// Simplifies llvm/lib/Transforms/Utils/AMDGPUEmitPrintf.cpp, as we don't have
|
|
/// to deal with %s (even if there were first-class strings in MLIR, they're not
|
|
/// legal input to gpu.printf) or non-constant format strings
|
|
struct GPUPrintfOpToHIPLowering : public ConvertOpToLLVMPattern<gpu::PrintfOp> {
|
|
using ConvertOpToLLVMPattern<gpu::PrintfOp>::ConvertOpToLLVMPattern;
|
|
|
|
LogicalResult
|
|
matchAndRewrite(gpu::PrintfOp gpuPrintfOp, gpu::PrintfOpAdaptor adaptor,
|
|
ConversionPatternRewriter &rewriter) const override;
|
|
};
|
|
|
|
/// The lowering of gpu.printf to a call to an external printf() function
|
|
///
|
|
/// This pass will add a declaration of printf() to the GPUModule if needed
|
|
/// and separate out the format strings into global constants. For some
|
|
/// runtimes, such as OpenCL on AMD, this is sufficient setup, as the compiler
|
|
/// will lower printf calls to appropriate device-side code
|
|
struct GPUPrintfOpToLLVMCallLowering
|
|
: public ConvertOpToLLVMPattern<gpu::PrintfOp> {
|
|
GPUPrintfOpToLLVMCallLowering(const LLVMTypeConverter &converter,
|
|
int addressSpace = 0)
|
|
: ConvertOpToLLVMPattern<gpu::PrintfOp>(converter),
|
|
addressSpace(addressSpace) {}
|
|
|
|
LogicalResult
|
|
matchAndRewrite(gpu::PrintfOp gpuPrintfOp, gpu::PrintfOpAdaptor adaptor,
|
|
ConversionPatternRewriter &rewriter) const override;
|
|
|
|
private:
|
|
int addressSpace;
|
|
};
|
|
|
|
/// Lowering of gpu.printf to a vprintf standard library.
|
|
struct GPUPrintfOpToVPrintfLowering
|
|
: public ConvertOpToLLVMPattern<gpu::PrintfOp> {
|
|
using ConvertOpToLLVMPattern<gpu::PrintfOp>::ConvertOpToLLVMPattern;
|
|
|
|
LogicalResult
|
|
matchAndRewrite(gpu::PrintfOp gpuPrintfOp, gpu::PrintfOpAdaptor adaptor,
|
|
ConversionPatternRewriter &rewriter) const override;
|
|
};
|
|
|
|
struct GPUReturnOpLowering : public ConvertOpToLLVMPattern<gpu::ReturnOp> {
|
|
using ConvertOpToLLVMPattern<gpu::ReturnOp>::ConvertOpToLLVMPattern;
|
|
|
|
LogicalResult
|
|
matchAndRewrite(gpu::ReturnOp op, OpAdaptor adaptor,
|
|
ConversionPatternRewriter &rewriter) const override;
|
|
};
|
|
|
|
namespace impl {
|
|
/// Unrolls op if it's operating on vectors.
|
|
LogicalResult scalarizeVectorOp(Operation *op, ValueRange operands,
|
|
ConversionPatternRewriter &rewriter,
|
|
const LLVMTypeConverter &converter);
|
|
} // namespace impl
|
|
|
|
/// Rewriting that unrolls SourceOp to scalars if it's operating on vectors.
|
|
template <typename SourceOp>
|
|
struct ScalarizeVectorOpLowering : public ConvertOpToLLVMPattern<SourceOp> {
|
|
public:
|
|
using ConvertOpToLLVMPattern<SourceOp>::ConvertOpToLLVMPattern;
|
|
|
|
LogicalResult
|
|
matchAndRewrite(SourceOp op, typename SourceOp::Adaptor adaptor,
|
|
ConversionPatternRewriter &rewriter) const override {
|
|
return impl::scalarizeVectorOp(op, adaptor.getOperands(), rewriter,
|
|
*this->getTypeConverter());
|
|
}
|
|
};
|
|
} // namespace mlir
|
|
|
|
#endif // MLIR_CONVERSION_GPUCOMMON_GPUOPSLOWERING_H_
|