MLIR 22.0.0git
mlir::GPUPrintfOpToLLVMCallLowering Struct Reference

The lowering of gpu.printf to a call to an external printf() function. More...

#include "Conversion/GPUCommon/GPUOpsLowering.h"

Inheritance diagram for mlir::GPUPrintfOpToLLVMCallLowering:

Public Member Functions

 GPUPrintfOpToLLVMCallLowering (const LLVMTypeConverter &converter, int addressSpace=0, LLVM::cconv::CConv callingConvention=LLVM::cconv::CConv::C, StringRef funcName="printf")
LogicalResult matchAndRewrite (gpu::PrintfOp gpuPrintfOp, gpu::PrintfOpAdaptor adaptor, ConversionPatternRewriter &rewriter) const override
Public Member Functions inherited from mlir::ConvertOpToLLVMPattern< gpu::PrintfOp >
 ConvertOpToLLVMPattern (const LLVMTypeConverter &typeConverter, PatternBenefit benefit=1)
LogicalResult matchAndRewrite (Operation *op, ArrayRef< Value > operands, ConversionPatternRewriter &rewriter) const final
 Wrappers around the RewritePattern methods that pass the derived op type.
Public Member Functions inherited from mlir::ConvertToLLVMPattern
 ConvertToLLVMPattern (StringRef rootOpName, MLIRContext *context, const LLVMTypeConverter &typeConverter, PatternBenefit benefit=1)

Additional Inherited Members

Public Types inherited from mlir::ConvertOpToLLVMPattern< gpu::PrintfOp >
using OpAdaptor
using OneToNOpAdaptor
Protected Member Functions inherited from mlir::ConvertToLLVMPattern
LLVM::LLVMDialect & getDialect () const
 Returns the LLVM dialect.
const LLVMTypeConvertergetTypeConverter () const
Type getIndexType () const
 Gets the MLIR type wrapping the LLVM integer type whose bit width is defined by the used type converter.
Type getIntPtrType (unsigned addressSpace=0) const
 Gets the MLIR type wrapping the LLVM integer type whose bit width corresponds to that of a LLVM pointer type.
Type getVoidType () const
 Gets the MLIR type wrapping the LLVM void type.
Type getVoidPtrType () const
 Get the MLIR type wrapping the LLVM i8* type.
Type getPtrType (unsigned addressSpace=0) const
 Get the MLIR type wrapping the LLVM ptr type.
Value getStridedElementPtr (ConversionPatternRewriter &rewriter, Location loc, MemRefType type, Value memRefDesc, ValueRange indices, LLVM::GEPNoWrapFlags noWrapFlags=LLVM::GEPNoWrapFlags::none) const
 Convenience wrapper for the corresponding helper utility.
bool isConvertibleAndHasIdentityMaps (MemRefType type) const
 Returns if the given memref type is convertible to LLVM and has an identity layout map.
Type getElementPtrType (MemRefType type) const
 Returns the type of a pointer to an element of the memref.
void getMemRefDescriptorSizes (Location loc, MemRefType memRefType, ValueRange dynamicSizes, ConversionPatternRewriter &rewriter, SmallVectorImpl< Value > &sizes, SmallVectorImpl< Value > &strides, Value &size, bool sizeInBytes=true) const
 Computes sizes, strides and buffer size of memRefType with identity layout.
Value getSizeInBytes (Location loc, Type type, ConversionPatternRewriter &rewriter) const
 Computes the size of type in bytes.
Value getNumElements (Location loc, MemRefType memRefType, ValueRange dynamicSizes, ConversionPatternRewriter &rewriter) const
 Computes total number of elements for the given MemRef and dynamicSizes.
MemRefDescriptor createMemRefDescriptor (Location loc, MemRefType memRefType, Value allocatedPtr, Value alignedPtr, ArrayRef< Value > sizes, ArrayRef< Value > strides, ConversionPatternRewriter &rewriter) const
 Creates and populates a canonical memref descriptor struct.
Value copyUnrankedDescriptor (OpBuilder &builder, Location loc, UnrankedMemRefType memRefType, Value operand, bool toDynamic) const
 Copies the given unranked memory descriptor to heap-allocated memory (if toDynamic is true) or to stack-allocated memory (otherwise) and returns the new descriptor.
LogicalResult copyUnrankedDescriptors (OpBuilder &builder, Location loc, TypeRange origTypes, SmallVectorImpl< Value > &operands, bool toDynamic) const
 Copies the memory descriptor for any operands that were unranked descriptors originally to heap-allocated memory (if toDynamic is true) or to stack-allocated memory (otherwise).
Static Protected Member Functions inherited from mlir::ConvertToLLVMPattern
static Value createIndexAttrConstant (OpBuilder &builder, Location loc, Type resultType, int64_t value)
 Create a constant Op producing a value of resultType from an index-typed integer attribute.

Detailed Description

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. However not all backends use the same calling convention and function naming. For example, the LLVM SPIRV backend requires calling convention LLVM::cconv::CConv::SPIR_FUNC and function name needs to be mangled as "_Z6printfPU3AS2Kcz". Default callingConvention is LLVM::cconv::CConv::C and funcName is "printf" but they can be customized as needed.

Definition at line 154 of file GPUOpsLowering.h.

Constructor & Destructor Documentation

◆ GPUPrintfOpToLLVMCallLowering()

mlir::GPUPrintfOpToLLVMCallLowering::GPUPrintfOpToLLVMCallLowering ( const LLVMTypeConverter & converter,
int addressSpace = 0,
LLVM::cconv::CConv callingConvention = LLVM::cconv::CConv::C,
StringRef funcName = "printf" )
inline

Member Function Documentation

◆ matchAndRewrite()

LogicalResult GPUPrintfOpToLLVMCallLowering::matchAndRewrite ( gpu::PrintfOp gpuPrintfOp,
gpu::PrintfOpAdaptor adaptor,
ConversionPatternRewriter & rewriter ) const
override

The documentation for this struct was generated from the following files: