MLIR 22.0.0git
mlir::gpu::GPUToNVVMPipelineOptions Struct Reference

Options for the gpu to nvvm pipeline. More...

#include "mlir/Dialect/GPU/Pipelines/Passes.h"

Inheritance diagram for mlir::gpu::GPUToNVVMPipelineOptions:

Public Attributes

PassOptions::Option< int64_tindexBitWidth
PassOptions::Option< std::string > cubinTriple
PassOptions::Option< std::string > cubinChip
PassOptions::Option< std::string > cubinFeatures
PassOptions::Option< std::string > cubinFormat
PassOptions::Option< std::string > cmdOptions
PassOptions::Option< int > optLevel
PassOptions::Option< boolkernelUseBarePtrCallConv
PassOptions::Option< boolhostUseBarePtrCallConv
PassOptions::Option< boolallowPatternRollback

Detailed Description

Options for the gpu to nvvm pipeline.

Definition at line 18 of file Passes.h.

Member Data Documentation

◆ allowPatternRollback

PassOptions::Option<bool> mlir::gpu::GPUToNVVMPipelineOptions::allowPatternRollback
Initial value:
{
*this, "allow-pattern-rollback",
llvm::cl::desc("Allow pattern rollback during dialect conversion"),
llvm::cl::init(true)}

Definition at line 61 of file Passes.h.

◆ cmdOptions

PassOptions::Option<std::string> mlir::gpu::GPUToNVVMPipelineOptions::cmdOptions
Initial value:
{
*this, "ptxas-cmd-options",
llvm::cl::desc(
"Command line options to pass to the downstream compiler."),
llvm::cl::init("")}

Definition at line 40 of file Passes.h.

◆ cubinChip

PassOptions::Option<std::string> mlir::gpu::GPUToNVVMPipelineOptions::cubinChip
Initial value:
{
*this, "cubin-chip", llvm::cl::desc("Chip to use to serialize to cubin."),
llvm::cl::init("sm_50")}

Definition at line 29 of file Passes.h.

◆ cubinFeatures

PassOptions::Option<std::string> mlir::gpu::GPUToNVVMPipelineOptions::cubinFeatures
Initial value:
{
*this, "cubin-features",
llvm::cl::desc("Features to use to serialize to cubin."),
llvm::cl::init("+ptx60")}

Definition at line 32 of file Passes.h.

◆ cubinFormat

PassOptions::Option<std::string> mlir::gpu::GPUToNVVMPipelineOptions::cubinFormat
Initial value:
{
*this, "cubin-format",
llvm::cl::desc("Compilation format to use to serialize to cubin."),
llvm::cl::init("fatbin")}

Definition at line 36 of file Passes.h.

◆ cubinTriple

PassOptions::Option<std::string> mlir::gpu::GPUToNVVMPipelineOptions::cubinTriple
Initial value:
{
*this, "cubin-triple",
llvm::cl::desc("Triple to use to serialize to cubin."),
llvm::cl::init("nvptx64-nvidia-cuda")}

Definition at line 25 of file Passes.h.

◆ hostUseBarePtrCallConv

PassOptions::Option<bool> mlir::gpu::GPUToNVVMPipelineOptions::hostUseBarePtrCallConv
Initial value:
{
*this, "host-bare-ptr-calling-convention",
llvm::cl::desc(
"Whether to use the bareptr calling convention on the host (warning "
"this should be false until the GPU layering is fixed)"),
llvm::cl::init(false)}

Definition at line 55 of file Passes.h.

◆ indexBitWidth

PassOptions::Option<int64_t> mlir::gpu::GPUToNVVMPipelineOptions::indexBitWidth
Initial value:
{
*this, "index-bitwidth",
llvm::cl::desc("Bitwidth of the index type for the host (warning this "
"should be 64 until the GPU layering is fixed)"),
llvm::cl::init(64)}

Definition at line 20 of file Passes.h.

◆ kernelUseBarePtrCallConv

PassOptions::Option<bool> mlir::gpu::GPUToNVVMPipelineOptions::kernelUseBarePtrCallConv
Initial value:
{
*this, "kernel-bare-ptr-calling-convention",
llvm::cl::desc(
"Whether to use the bareptr calling convention on the kernel "
"(warning this should be false until the GPU layering is fixed)"),
llvm::cl::init(false)}

Definition at line 49 of file Passes.h.

◆ optLevel

PassOptions::Option<int> mlir::gpu::GPUToNVVMPipelineOptions::optLevel
Initial value:
{
*this, "opt-level",
llvm::cl::desc("Optimization level for NVVM compilation"),
llvm::cl::init(2)}

Definition at line 45 of file Passes.h.


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