Searched refs:hsaco (Results 1 – 9 of 9) sorted by relevance
| /external/tensorflow/tensorflow/compiler/xla/service/gpu/llvm_gpu_backend/ |
| D | gpu_backend_lib.cc | 605 std::vector<uint8> hsaco; member
617 const std::string& gfx, std::vector<uint8>& hsaco);
619 const std::vector<uint8>& hsaco);
625 const std::string& gfx, std::vector<uint8>& hsaco) { in Find() argument
633 hsaco = x.hsaco; in Find()
646 const std::string& gfx, const std::vector<uint8>& hsaco) { in Add() argument
652 g_hsacoCache.cache.back().hsaco = hsaco; in Add()
763 std::vector<uint8> hsaco(hsaco_file_size); in EmitModuleToHsaco() local
765 hsaco_file.read(reinterpret_cast<char*>(&hsaco[0]), hsaco_file_size); in EmitModuleToHsaco()
772 return hsaco; in EmitModuleToHsaco()
[all …]
|
| /external/tensorflow/tensorflow/compiler/xla/service/gpu/ |
| D | amdgpu_compiler.cc | 129 std::vector<uint8> hsaco; in CompileTargetBinary() local
134 hsaco, amdgpu::CompileToHsaco(llvm_module, gpu_version, module_config, in CompileTargetBinary()
138 return std::pair<std::string, std::vector<uint8>>("", std::move(hsaco)); in CompileTargetBinary()
|
| /external/llvm-project/mlir/test/Conversion/GPUToROCm/ |
| D | lower-rocdl-kernel-to-hsaco.mlir | 1 // RUN: mlir-opt %s --test-kernel-to-hsaco -split-input-file | FileCheck %s
3 // CHECK: attributes {rocdl.hsaco = "HSACO"}
|
| /external/tensorflow/tensorflow/stream_executor/rocm/ |
| D | rocm_gpu_executor.cc | 256 const char* hsaco = spec.cuda_cubin_in_memory().bytes(); in GetKernel() local
258 module = in_memory_modules_[hsaco]; in GetKernel()
261 TF_RETURN_IF_ERROR(GpuDriver::LoadHsaco(context_, hsaco, &module)); in GetKernel()
263 kernel_to_gpu_binary_[kernel] = hsaco; in GetKernel()
396 port::Status GpuExecutor::LoadModuleFromHsaco(const char* hsaco, in LoadModuleFromHsaco() argument
399 std::tie(*module, module_refcount) = gpu_binary_to_module_[hsaco]; in LoadModuleFromHsaco()
402 TF_RETURN_IF_ERROR(GpuDriver::LoadHsaco(context_, hsaco, module)); in LoadModuleFromHsaco()
404 in_memory_modules_[hsaco] = *module; in LoadModuleFromHsaco()
405 VLOG(3) << "Loaded HSACO " << static_cast<const void*>(hsaco) in LoadModuleFromHsaco()
409 VLOG(3) << "HSACO " << static_cast<const void*>(hsaco) in LoadModuleFromHsaco()
[all …]
|
| /external/tensorflow/tensorflow/compiler/mlir/tools/kernel_gen/transforms/ |
| D | gpu_kernel_to_blob_pass.cc | 130 auto hsaco = hsaco_or.ValueOrDie(); in GetGpuBinaryBlob() local
135 return hsaco; in GetGpuBinaryBlob()
138 images.push_back({arch_str, std::move(hsaco)}); in GetGpuBinaryBlob()
|
| /external/llvm-project/mlir/test/Conversion/GPUCommon/ |
| D | lower-launch-func-to-gpu-runtime-calls.mlir | 2 // RUN: mlir-opt -allow-unregistered-dialect %s --gpu-to-llvm="gpu-binary-annotation=rocdl.hsaco" |…
11 nvvm.cubin = "CUBIN", rocdl.hsaco = "HSACO"
|
| /external/tensorflow/tensorflow/stream_executor/gpu/ |
| D | gpu_executor.h | 280 port::Status LoadModuleFromHsaco(const char* hsaco, GpuModuleHandle* module)
|
| /external/tensorflow/tensorflow/core/kernels/mlir_generated/ |
| D | build_defs.bzl | 29 "GPU binaries in either cubin format or hsaco format",
|
| /external/tensorflow/tensorflow/stream_executor/cuda/ |
| D | cuda_gpu_executor.cc | 257 port::Status GpuExecutor::LoadModuleFromHsaco(const char* hsaco, in LoadModuleFromHsaco() argument
|