Home
last modified time | relevance | path

Searched refs:hlo_module_config (Results 1 – 25 of 34) sorted by relevance

12

/external/tensorflow/tensorflow/compiler/xla/service/gpu/llvm_gpu_backend/
Dgpu_backend_lib.cc145 const HloModuleConfig& hlo_module_config, absl::string_view feature_str) { in GetTargetMachine() argument
164 switch (hlo_module_config.debug_options().xla_backend_optimization_level()) { in GetTargetMachine()
339 const HloModuleConfig& hlo_module_config, in NVPTXTargetModuleLinker() argument
353 hlo_module_config.debug_options().xla_gpu_ftz()); in NVPTXTargetModuleLinker()
356 if (hlo_module_config.debug_options().xla_gpu_ftz()) { in NVPTXTargetModuleLinker()
367 const HloModuleConfig& hlo_module_config) { in NVPTXGetTargetMachine() argument
371 hlo_module_config, "+ptx60"); in NVPTXGetTargetMachine()
378 const HloModuleConfig& hlo_module_config, in LinkAndOptimizeModule() argument
384 TF_RETURN_IF_ERROR(module_linker(module, gpu_version, hlo_module_config, in LinkAndOptimizeModule()
416 hlo_module_config.debug_options().xla_backend_optimization_level(); in LinkAndOptimizeModule()
[all …]
Dgpu_backend_lib.h44 const HloModuleConfig& hlo_module_config, const string& libdevice_dir_path,
54 const HloModuleConfig& hlo_module_config, const string& rocdl_dir_path);
DBUILD55 "//tensorflow/compiler/xla/service:hlo_module_config",
/external/tensorflow/tensorflow/compiler/xla/service/gpu/
Dnvptx_compiler.cc70 const HloModuleConfig& hlo_module_config) { in PrintCantFindCudaMessage() argument
74 for (const auto& dir : CandidateCudaRoots(hlo_module_config)) { in PrintCantFindCudaMessage()
84 string GetLibdeviceDir(const HloModuleConfig& hlo_module_config) { in GetLibdeviceDir() argument
85 for (const string& cuda_root : CandidateCudaRoots(hlo_module_config)) { in GetLibdeviceDir()
98 hlo_module_config); in GetLibdeviceDir()
343 int cc_minor, const HloModuleConfig& hlo_module_config, bool relocatable) { in CompileGpuAsmOrGetCachedResult() argument
372 auto ptxas_config = PtxOptsFromConfig(hlo_module_config); in CompileGpuAsmOrGetCachedResult()
386 if (!hlo_module_config.debug_options() in CompileGpuAsmOrGetCachedResult()
391 hlo_module_config); in CompileGpuAsmOrGetCachedResult()
414 hlo_module_config); in CompileGpuAsmOrGetCachedResult()
Dir_emitter_nested.cc38 IrEmitterNested::IrEmitterNested(const HloModuleConfig& hlo_module_config, in IrEmitterNested() argument
41 : IrEmitter(hlo_module_config, ir_emitter_context, /*is_nested=*/true), in IrEmitterNested()
45 const HloModuleConfig& hlo_module_config, in Create() argument
49 hlo_module_config, nested_computation, ir_emitter_context)); in Create()
Dir_emitter_nested.h43 const HloModuleConfig& hlo_module_config,
67 IrEmitterNested(const HloModuleConfig& hlo_module_config,
Dgemm_algorithm_picker.cc67 const HloModuleConfig& hlo_module_config = gemm->GetModule()->config(); in DoUncachedGemmAutotune() local
69 hlo_module_config.debug_options().xla_gpu_autotune_level() > 1; in DoUncachedGemmAutotune()
71 stream, allocator, PtxOptsFromConfig(hlo_module_config), in DoUncachedGemmAutotune()
74 BufferComparator comparator(gemm->shape(), hlo_module_config); in DoUncachedGemmAutotune()
Dstream_executor_util.cc226 se::GpuAsmOpts PtxOptsFromConfig(const HloModuleConfig& hlo_module_config) { in PtxOptsFromConfig() argument
228 hlo_module_config.debug_options().xla_gpu_asm_extra_flags(); in PtxOptsFromConfig()
232 hlo_module_config.debug_options().xla_gpu_disable_gpuasm_optimizations(), in PtxOptsFromConfig()
233 hlo_module_config.debug_options().xla_gpu_cuda_data_dir(), extra_flags); in PtxOptsFromConfig()
Dgpu_conv_algorithm_picker.cc389 const HloModuleConfig& hlo_module_config = instr->GetModule()->config(); in PickBestAlgorithmNoCacheCuda() local
391 hlo_module_config.debug_options().xla_gpu_autotune_level(); in PickBestAlgorithmNoCacheCuda()
404 stream, allocator, PtxOptsFromConfig(hlo_module_config)); in PickBestAlgorithmNoCacheCuda()
464 stream, allocator, PtxOptsFromConfig(hlo_module_config)); in PickBestAlgorithmNoCacheCuda()
565 comparator.emplace(result_shape, hlo_module_config); in PickBestAlgorithmNoCacheCuda()
640 !hlo_module_config.debug_options().xla_gpu_deterministic_ops()) { in PickBestAlgorithmNoCacheCuda()
Dstream_executor_util.h79 se::GpuAsmOpts PtxOptsFromConfig(const HloModuleConfig& hlo_module_config);
Dnvptx_compiler.h79 int cc_minor, const HloModuleConfig& hlo_module_config, bool relocatable);
Delemental_ir_emitter.cc73 const HloModuleConfig& hlo_module_config, llvm::Module* module, in GpuElementalIrEmitter() argument
76 hlo_module_config_(hlo_module_config), in GpuElementalIrEmitter()
Delemental_ir_emitter.h46 GpuElementalIrEmitter(const HloModuleConfig& hlo_module_config,
Dir_emitter.h116 explicit IrEmitter(const HloModuleConfig& hlo_module_config,
Dir_emitter_unnested.h150 const HloModuleConfig& hlo_module_config,
240 IrEmitterUnnested(const HloModuleConfig& hlo_module_config,
Dir_emitter.cc79 IrEmitter::IrEmitter(const HloModuleConfig& hlo_module_config, in IrEmitter() argument
87 hlo_module_config_(hlo_module_config) { in IrEmitter()
/external/tensorflow/tensorflow/compiler/xla/service/gpu/tests/
DBUILD117 "//tensorflow/compiler/xla/service:hlo_module_config",
144 "//tensorflow/compiler/xla/service:hlo_module_config",
171 "//tensorflow/compiler/xla/service:hlo_module_config",
198 "//tensorflow/compiler/xla/service:hlo_module_config",
225 "//tensorflow/compiler/xla/service:hlo_module_config",
252 "//tensorflow/compiler/xla/service:hlo_module_config",
277 "//tensorflow/compiler/xla/service:hlo_module_config",
315 "//tensorflow/compiler/xla/service:hlo_module_config",
327 "//tensorflow/compiler/xla/service:hlo_module_config",
358 "//tensorflow/compiler/xla/service:hlo_module_config",
[all …]
/external/tensorflow/tensorflow/compiler/xla/service/cpu/
Ddot_op_emitter.cc131 const HloModuleConfig& hlo_module_config,
240 mlir::MLIRContext* mlir_context, const HloModuleConfig& hlo_module_config, in DotOpEmitter() argument
251 hlo_module_config_(hlo_module_config), in DotOpEmitter()
1044 mlir::MLIRContext* mlir_context, const HloModuleConfig& hlo_module_config, in EmitNonBatchDotOperation() argument
1054 hlo_module_config, target_machine_features); in EmitNonBatchDotOperation()
1128 mlir::MLIRContext* mlir_context, const HloModuleConfig& hlo_module_config, in EmitBatchDotOperation() argument
1186 executable_run_options_value, b, mlir_context, hlo_module_config, in EmitBatchDotOperation()
1237 const HloModuleConfig& hlo_module_config, in EmitDotOperation() argument
1247 hlo_module_config, target_machine_features); in EmitDotOperation()
1253 hlo_module_config, target_machine_features); in EmitDotOperation()
Ddot_op_emitter.h68 const HloModuleConfig& hlo_module_config,
DBUILD394 "//tensorflow/compiler/xla/service:hlo_module_config",
461 "//tensorflow/compiler/xla/service:hlo_module_config",
493 "//tensorflow/compiler/xla/service:hlo_module_config",
522 "//tensorflow/compiler/xla/service:hlo_module_config",
1086 "//tensorflow/compiler/xla/service:hlo_module_config",
/external/tensorflow/tensorflow/compiler/mlir/tools/kernel_gen/transforms/
Dgpu_kernel_to_blob_pass.cc235 const xla::HloModuleConfig& hlo_module_config) { in GetLibdeviceDir() argument
237 hlo_module_config.debug_options().xla_gpu_cuda_data_dir())) { in GetLibdeviceDir()
DBUILD141 "//tensorflow/compiler/xla/service:hlo_module_config",
/external/tensorflow/tensorflow/compiler/xla/service/interpreter/
DBUILD50 "//tensorflow/compiler/xla/service:hlo_module_config",
119 "//tensorflow/compiler/xla/service:hlo_module_config",
/external/tensorflow/tensorflow/compiler/xla/service/llvm_ir/
DBUILD77 "//tensorflow/compiler/xla/service:hlo_module_config",
/external/tensorflow/tensorflow/stream_executor/tpu/
DBUILD55 "//tensorflow/compiler/xla/service:hlo_module_config",

12