| /external/tensorflow/tensorflow/stream_executor/rocm/ |
| D | rocm_gpu_executor.cc | 91 return static_cast<GpuExecutor*>(stream->parent()->implementation()) in GetGpuContext()
95 GpuContext* ExtractGpuContext(GpuExecutor* rocm_exec) { in ExtractGpuContext()
100 GpuExecutor* ExtractGpuExecutor(StreamExecutor* stream_exec) { in ExtractGpuExecutor()
101 return static_cast<GpuExecutor*>(stream_exec->implementation()); in ExtractGpuExecutor()
104 GpuExecutor::~GpuExecutor() { in ~GpuExecutor()
116 bool GpuExecutor::UnloadModule(ModuleHandle module_handle) { in UnloadModule()
122 bool GpuExecutor::UnloadGpuBinary(const void* gpu_binary) { in UnloadGpuBinary()
139 void GpuExecutor::UnloadKernel(const KernelBase* kernel) { in UnloadKernel()
143 port::Status GpuExecutor::Init(int device_ordinal, in Init()
166 bool GpuExecutor::FindOnDiskForComputeCapability( in FindOnDiskForComputeCapability()
[all …]
|
| D | rocm_fft.h | 35 class GpuExecutor; variable
66 port::Status Initialize(GpuExecutor *parent, Stream *stream, int rank,
74 port::Status Initialize(GpuExecutor *parent, Stream *stream, int rank,
82 GpuExecutor *parent_;
102 explicit ROCMFft(GpuExecutor *parent) : parent_(parent) {} in ROCMFft()
108 GpuExecutor *parent_;
|
| D | rocm_rng.cc | 70 hiprandStatus_t operator()(GpuExecutor* parent, Args... args) { \
99 hiprandStatus operator()(GpuExecutor* parent, Args... args) { \
120 GpuRng::GpuRng(GpuExecutor* parent) : parent_(parent), rng_(nullptr) {} in GpuRng()
294 gpu::GpuExecutor* rocm_executor = in initialize_rocrand()
295 dynamic_cast<gpu::GpuExecutor*>(parent); in initialize_rocrand()
|
| D | rocm_blas.h | 38 class GpuExecutor; variable
52 explicit ROCMBlas(GpuExecutor *parent);
148 GpuExecutor *parent_;
|
| D | rocm_fft.cc | 51 hipfftResult operator()(GpuExecutor *parent, Args... args) { \
80 hipfftResult operator()(GpuExecutor *parent, Args... args) { \
134 bool SetStream(GpuExecutor *parent, hipfftHandle plan, Stream *stream) { in SetStream()
146 GpuExecutor *parent, Stream *stream, int rank, uint64 *elem_count, in Initialize()
338 port::Status ROCMFftPlan::Initialize(GpuExecutor *parent, Stream *stream, in Initialize()
594 gpu::GpuExecutor *rocm_executor = in initialize_rocfft()
595 dynamic_cast<gpu::GpuExecutor *>(parent); in initialize_rocfft()
|
| D | rocm_platform.cc | 138 this, MakeUnique<GpuExecutor>(config.plugin_config)); in GetUncachedExecutor()
|
| D | rocm_dnn.h | 32 class GpuExecutor; variable
43 explicit MIOpenSupport(GpuExecutor* parent);
620 GpuExecutor* GetParentExecutor() { return parent_; } in GetParentExecutor()
623 GpuExecutor* parent_; // Parent executor object. Not owned.
|
| D | rocm_blas.cc | 58 rocblas_status operator()(GpuExecutor *parent, Args... args) { \
91 rocblas_status operator()(GpuExecutor *parent, Args... args) { \
306 ROCMBlas::ROCMBlas(gpu::GpuExecutor *parent) in ROCMBlas()
2342 gpu::GpuExecutor *rocm_executor = in initialize_rocblas()
2343 dynamic_cast<gpu::GpuExecutor *>(parent); in initialize_rocblas()
|
| /external/tensorflow/tensorflow/stream_executor/cuda/ |
| D | cuda_gpu_executor.cc | 113 GpuContext* ExtractGpuContext(GpuExecutor* cuda_exec) { in ExtractGpuContext()
118 GpuExecutor* ExtractGpuExecutor(StreamExecutor* stream_exec) { in ExtractGpuExecutor()
119 return static_cast<GpuExecutor*>(stream_exec->implementation()); in ExtractGpuExecutor()
122 GpuExecutor::~GpuExecutor() { in ~GpuExecutor()
130 port::Status GpuExecutor::Init(int device_ordinal, in Init()
153 bool GpuExecutor::FindOnDiskForComputeCapability( in FindOnDiskForComputeCapability()
179 bool GpuExecutor::FindOnDiskForISAVersion(absl::string_view filename, in FindOnDiskForISAVersion()
220 bool GpuExecutor::LoadModuleFromCuBin(const char* cubin, CUmodule* module) { in LoadModuleFromCuBin()
242 bool GpuExecutor::LoadModuleFromPtx(const char* ptx, CUmodule* module) { in LoadModuleFromPtx()
262 bool GpuExecutor::LoadModuleFromHsaco(const char* hsaco, CUmodule* module) { in LoadModuleFromHsaco()
[all …]
|
| D | cuda_fft.h | 35 class GpuExecutor; variable
67 port::Status Initialize(GpuExecutor* parent, Stream* stream, int rank,
75 port::Status Initialize(GpuExecutor* parent, Stream* stream, int rank,
86 GpuExecutor* parent_;
107 explicit CUDAFft(GpuExecutor* parent) : parent_(parent) {} in CUDAFft()
113 GpuExecutor* parent_;
|
| D | cuda_blas.h | 41 class GpuExecutor; variable
55 explicit CUDABlas(GpuExecutor *parent);
150 GpuExecutor *parent_;
|
| D | cuda_rng.cc | 64 GpuRng::GpuRng(GpuExecutor* parent) : parent_(parent), rng_(nullptr) {} in GpuRng()
236 gpu::GpuExecutor* cuda_executor = in initialize_curand()
237 dynamic_cast<gpu::GpuExecutor*>(parent); in initialize_curand()
|
| D | cuda_fft.cc | 64 bool SetStream(GpuExecutor *parent, cufftHandle plan, Stream *stream) { in SetStream()
77 GpuExecutor *parent, Stream *stream, int rank, uint64 *elem_count, in Initialize()
233 port::Status CUDAFftPlan::Initialize(GpuExecutor *parent, Stream *stream, in Initialize()
541 gpu::GpuExecutor *cuda_executor = in initialize_cufft()
542 dynamic_cast<gpu::GpuExecutor *>(parent); in initialize_cufft()
|
| D | cuda_gpu_executor.h | 30 using CUDAExecutor = gpu::GpuExecutor;
|
| D | cuda_platform.cc | 172 this, MakeUnique<GpuExecutor>(config.plugin_config)); in GetUncachedExecutor()
|
| D | cuda_dnn.h | 33 class GpuExecutor; variable
45 explicit CudnnSupport(GpuExecutor* parent);
551 GpuExecutor* parent_; // Parent executor object. Not owned.
|
| D | cuda_dnn.cc | 184 CudnnHandle GetHandle(GpuExecutor* executor, Stream* stream) { in GetHandle()
298 CudnnSupport::CudnnSupport(GpuExecutor* parent) : parent_(parent) {} in CudnnSupport()
1253 CudnnRnnSequenceTensorDescriptor(GpuExecutor* parent, int max_seq_length, in CudnnRnnSequenceTensorDescriptor()
1276 GpuExecutor* parent, int max_seq_length, int batch_size, int data_size, in Create()
1295 GpuExecutor* parent, int max_seq_length, int batch_size, int data_size, in Create()
1368 CudnnRnnStateTensorDescriptor(GpuExecutor* parent, int num_layers, in CudnnRnnStateTensorDescriptor()
4123 gpu::GpuExecutor* cuda_executor = in initialize_cudnn()
4124 dynamic_cast<gpu::GpuExecutor*>(parent); in initialize_cudnn()
|
| /external/tensorflow/tensorflow/stream_executor/gpu/ |
| D | gpu_stream.h | 29 class GpuExecutor; variable
37 explicit GpuStream(GpuExecutor* parent) in GpuStream()
76 GpuExecutor* parent() const { return parent_; } in parent()
79 GpuExecutor* parent_; // Executor that spawned this stream.
|
| D | gpu_activation.cc | 25 GpuContext* ExtractGpuContext(GpuExecutor* gpu_exec);
26 GpuExecutor* ExtractGpuExecutor(StreamExecutor* stream_exec);
29 GpuExecutor* gpu_exec) in ScopedActivateExecutorContext()
|
| D | gpu_timer.h | 29 class GpuExecutor; variable
37 explicit GpuTimer(GpuExecutor* parent) in GpuTimer()
73 GpuExecutor* parent_;
|
| D | gpu_rng.h | 38 class GpuExecutor; variable
46 explicit GpuRng(GpuExecutor* parent);
91 GpuExecutor* parent_;
|
| D | gpu_activation.h | 34 class GpuExecutor; variable
41 explicit ScopedActivateExecutorContext(GpuExecutor* gpu_exec);
|
| D | gpu_event.h | 31 explicit GpuEvent(GpuExecutor* parent);
53 GpuExecutor* parent_;
|
| D | gpu_executor.h | 44 class GpuExecutor : public internal::StreamExecutorInterface {
48 explicit GpuExecutor(const PluginConfig& plugin_config) in GpuExecutor() function
60 ~GpuExecutor() override;
341 SE_DISALLOW_COPY_AND_ASSIGN(GpuExecutor);
|
| D | gpu_event.cc | 25 GpuEvent::GpuEvent(GpuExecutor* parent) in GpuEvent()
|