MCPcopy Create free account
hub / github.com/DeepRec-AI/DeepRec / GetGpuDeviceInfo

Function GetGpuDeviceInfo

tensorflow/compiler/xla/service/gpu/gpu_compiler.cc:496–514  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

494}
495
496GpuDeviceInfo GetGpuDeviceInfo(se::StreamExecutor* stream_exec) {
497 GpuDeviceInfo gpu_device_info;
498 gpu_device_info.threads_per_block_limit =
499 stream_exec->GetDeviceDescription().threads_per_block_limit();
500 gpu_device_info.threads_per_warp =
501 stream_exec->GetDeviceDescription().threads_per_warp();
502 gpu_device_info.shared_memory_per_block =
503 stream_exec->GetDeviceDescription().shared_memory_per_block();
504 gpu_device_info.threads_per_core_limit =
505 stream_exec->GetDeviceDescription().threads_per_core_limit();
506 gpu_device_info.core_count = stream_exec->GetDeviceDescription().core_count();
507 gpu_device_info.block_dim_limit_x =
508 stream_exec->GetDeviceDescription().block_dim_limit().x;
509 gpu_device_info.block_dim_limit_y =
510 stream_exec->GetDeviceDescription().block_dim_limit().y;
511 gpu_device_info.block_dim_limit_z =
512 stream_exec->GetDeviceDescription().block_dim_limit().z;
513 return gpu_device_info;
514}
515
516StatusOr<std::unique_ptr<Executable>> GpuCompiler::RunBackend(
517 std::unique_ptr<HloModule> module, se::StreamExecutor* stream_exec,

Callers 1

RunBackendMethod · 0.85

Calls 2

core_countMethod · 0.80

Tested by

no test coverage detected