| /external/tensorflow/tensorflow/compiler/xla/service/gpu/ |
| D | nccl_all_to_all_thunk.cc | 87 se::gpu::GpuStreamHandle gpu_stream = se::gpu::AsGpuStreamValue(&stream); in RunAllToAll() local
120 gpu_stream)); in RunAllToAll()
123 gpu_stream)); in RunAllToAll()
144 /*rank=*/i, comm, gpu_stream)); in RunAllToAll()
146 /*rank=*/i, comm, gpu_stream)); in RunAllToAll()
|
| D | nccl_collective_permute_thunk.cc | 206 se::gpu::GpuStreamHandle gpu_stream = se::gpu::AsGpuStreamValue(&stream); in RunCollectivePermute() local
214 static_cast<const void*>(comm), gpu_stream); in RunCollectivePermute()
216 *target_id, comm, gpu_stream)); in RunCollectivePermute()
225 static_cast<const void*>(comm), gpu_stream); in RunCollectivePermute()
227 *source_id, comm, gpu_stream)); in RunCollectivePermute()
|
| D | nccl_all_gather_thunk.cc | 71 se::gpu::GpuStreamHandle gpu_stream = se::gpu::AsGpuStreamValue(&stream); in RunAllGather() local
89 gpu_stream); in RunAllGather()
92 send_buffer, recv_buffer, element_count, dtype, comm, gpu_stream)); in RunAllGather()
|
| D | custom_call_thunk.cc | 59 auto gpu_stream = se::gpu::AsGpuStreamValue(params.stream); in ExecuteOnStream() local
61 call_target_(gpu_stream, buffers.data(), opaque_.data(), opaque_.size(), in ExecuteOnStream()
|
| D | nccl_all_reduce_thunk.cc | 51 se::gpu::GpuStreamHandle gpu_stream = se::gpu::AsGpuStreamValue(&stream); in RunAllReduce() local
68 gpu_stream); in RunAllReduce()
72 comm, gpu_stream)); in RunAllReduce()
379 se::gpu::GpuStreamHandle gpu_stream = se::gpu::AsGpuStreamValue(&stream); in RunReduceScatter() local
407 gpu_stream); in RunReduceScatter()
410 comm, gpu_stream)); in RunReduceScatter()
|
| D | runtime_intrinsics.cc | 41 config.gpu_stream = stream_handle; in AssertOnGpu()
|
| D | cusolver_context.cc | 302 const gpuStream_t* gpu_stream = in Create() local
305 TF_RETURN_IF_ERROR(ConvertStatus(GpuSolverSetStream(handle, *gpu_stream))); in Create()
|
| D | bef_thunk.cc | 246 tfrt::gpu::wrapper::Stream(se_gpu_stream->gpu_stream())); in CreateGpuStream()
|
| /external/tensorflow/tensorflow/compiler/xla/stream_executor/ |
| D | executor_cache.cc | 75 if (config.gpu_stream) { in Get()
80 if (executor->FindAllocatedStream(config.gpu_stream)) { in Get()
87 absl::StrFormat("No executors own stream %p", config.gpu_stream)); in Get()
|
| D | platform.h | 83 void* gpu_stream = nullptr; member
|
| D | stream_executor_pimpl.h | 545 Stream* FindAllocatedStream(void* gpu_stream) { in FindAllocatedStream() argument
546 return implementation()->FindAllocatedStream(gpu_stream); in FindAllocatedStream()
|
| /external/tensorflow/tensorflow/compiler/xla/stream_executor/gpu/ |
| D | BUILD | 98 ":gpu_stream",
156 hdrs = if_gpu_is_configured(["gpu_stream.h"]),
165 name = "gpu_stream",
166 srcs = if_gpu_is_configured(["gpu_stream.cc"]),
167 hdrs = if_gpu_is_configured(["gpu_stream.h"]),
194 ":gpu_stream",
|
| D | gpu_stream.h | 70 GpuStreamHandle gpu_stream() const { in gpu_stream() function
76 GpuStreamHandle cuda_stream() const { return gpu_stream(); } in cuda_stream()
|
| D | gpu_timer.cc | 76 parent_->gpu_context(), start_event_, stream->gpu_stream()); in Start()
85 parent_->gpu_context(), stop_event_, stream->gpu_stream()); in Stop()
|
| D | gpu_event.cc | 41 stream->gpu_stream()); in Record()
|
| D | gpu_stream.cc | 58 return AsGpuStream(stream)->gpu_stream(); in AsGpuStreamValue()
|
| D | gpu_executor.h | 302 Stream* FindAllocatedStream(void* gpu_stream) override { in FindAllocatedStream() argument
304 auto it = alive_gpu_streams_.find(gpu_stream); in FindAllocatedStream()
|
| /external/tensorflow/tensorflow/stream_executor/gpu/ |
| D | BUILD | 104 hdrs = if_gpu_is_configured(["gpu_stream.h"]),
109 name = "gpu_stream",
110 hdrs = if_gpu_is_configured(["gpu_stream.h"]),
111 deps = ["//tensorflow/compiler/xla/stream_executor/gpu:gpu_stream"],
|
| /external/tensorflow/tensorflow/compiler/xla/stream_executor/cuda/ |
| D | stream_search_test.cc | 35 config.gpu_stream = bad_ptr; in TEST_F()
59 c.gpu_stream = gpu_ptr; in TEST_F()
|
| D | cuda_platform.cc | 162 if (config.gpu_stream) { in GetExecutor()
|
| /external/tensorflow/tensorflow/core/profiler/utils/ |
| D | group_events_test.cc | 235 auto gpu_stream = device_plane_builder.GetOrCreateLine(0); in TEST() local
244 CreateXEvent(&device_plane_builder, &gpu_stream, kTF1GpuEvent, 200, 300, in TEST()
257 CreateXEvent(&device_plane_builder, &gpu_stream, kLegacyGpuEvent, 300, 400, in TEST()
269 CreateXEvent(&device_plane_builder, &gpu_stream, kEagerOpGpuEvent, 400, 500, in TEST()
281 CreateXEvent(&device_plane_builder, &gpu_stream, kEagerFuncGpuEvent, 500, 600, in TEST()
|
| /external/tensorflow/tensorflow/compiler/tf2xla/kernels/ |
| D | light_outside_compilation.cc | 325 const void* gpu_stream = reinterpret_cast<const void*>( in ReinitializeGpuDevice() local
328 context, gpu_stream, in ReinitializeGpuDevice()
413 config.gpu_stream = stream_handle; in CallTfKernel()
|
| /external/tensorflow/tensorflow/core/common_runtime/gpu/ |
| D | gpu_device.h | 61 void Reinitialize(OpKernelContext* context, const void* gpu_stream,
65 void Reinitialize(OpKernelContext* context, const void* gpu_stream,
|
| D | gpu_device.cc | 128 void Reinitialize(OpKernelContext* context, const gpuStream_t* gpu_stream, in Reinitialize() argument
139 stream_ = gpu_stream; in Reinitialize()
903 const void* gpu_stream, in Reinitialize() argument
911 ->Reinitialize(context, static_cast<const gpuStream_t*>(gpu_stream), in Reinitialize()
916 const void* gpu_stream, in Reinitialize() argument
921 ->Reinitialize(context, static_cast<const gpuStream_t*>(gpu_stream), in Reinitialize()
1165 const gpuStream_t* gpu_stream = reinterpret_cast<const gpuStream_t*>( in ReinitializeDevice() local
1167 concrete_device->Reinitialize(context, gpu_stream, tf_device_id_, allocator, in ReinitializeDevice()
|
| /external/tensorflow/tensorflow/stream_executor/rocm/ |
| D | rocm_platform.cc | 132 if (config.gpu_stream) { in GetExecutor()
|