| /external/tensorflow/tensorflow/stream_executor/rocm/ |
| D | rocm_gpu_executor.cc | 90 return static_cast<GpuExecutor*>(stream->parent()->implementation()) in GetGpuContext()
94 GpuContext* ExtractGpuContext(GpuExecutor* rocm_exec) { in ExtractGpuContext()
99 GpuExecutor* ExtractGpuExecutor(StreamExecutor* stream_exec) { in ExtractGpuExecutor()
100 return static_cast<GpuExecutor*>(stream_exec->implementation()); in ExtractGpuExecutor()
103 GpuExecutor::~GpuExecutor() { in ~GpuExecutor()
116 bool GpuExecutor::UnloadModule(ModuleHandle module_handle) { in UnloadModule()
122 bool GpuExecutor::UnloadGpuBinary(const void* gpu_binary) { in UnloadGpuBinary()
144 void GpuExecutor::UnloadKernel(const KernelBase* kernel) { in UnloadKernel()
160 port::Status GpuExecutor::Init(int device_ordinal, in Init()
183 bool GpuExecutor::FindOnDiskForComputeCapability( in FindOnDiskForComputeCapability()
[all …]
|
| D | rocm_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 ROCMFft(GpuExecutor *parent) : parent_(parent) {} in ROCMFft()
113 GpuExecutor *parent_;
|
| D | rocm_rng.cc | 69 hiprandStatus_t operator()(GpuExecutor* parent, Args... args) { \
98 hiprandStatus operator()(GpuExecutor* parent, Args... args) { \
119 GpuRng::GpuRng(GpuExecutor* parent) : parent_(parent), rng_(nullptr) {} in GpuRng()
293 gpu::GpuExecutor* rocm_executor = in initialize_rocrand()
294 dynamic_cast<gpu::GpuExecutor*>(parent); in initialize_rocrand()
|
| D | rocm_blas.h | 61 class GpuExecutor;
75 explicit ROCMBlas(GpuExecutor *parent);
199 GpuExecutor *parent_;
|
| D | rocm_fft.cc | 51 hipfftResult operator()(GpuExecutor *parent, Args... args) { \
80 hipfftResult operator()(GpuExecutor *parent, Args... args) { \
146 bool SetStream(GpuExecutor *parent, hipfftHandle plan, Stream *stream) { in SetStream()
158 GpuExecutor *parent, Stream *stream, int rank, uint64 *elem_count, in Initialize()
318 port::Status ROCMFftPlan::Initialize(GpuExecutor *parent, Stream *stream, in Initialize()
599 gpu::GpuExecutor *rocm_executor = in initialize_rocfft()
600 dynamic_cast<gpu::GpuExecutor *>(parent); in initialize_rocfft()
|
| D | rocm_platform.cc | 109 return GpuExecutor::CreateDeviceDescription(ordinal); in DescriptionForDevice()
138 this, absl::make_unique<GpuExecutor>(config.plugin_config), in GetUncachedExecutor()
|
| D | rocm_dnn.h | 33 class GpuExecutor; variable
79 explicit MIOpenSupport(GpuExecutor* parent);
672 GpuExecutor* GetParentExecutor() { return parent_; } in GetParentExecutor()
687 GpuExecutor* parent_; // Parent executor object. Not owned.
|
| /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 port::Status GpuExecutor::LoadModuleFromCuBin(const char* cubin, in LoadModuleFromCuBin()
239 port::Status GpuExecutor::LoadModuleFromPtx(const char* ptx, CUmodule* module) { in LoadModuleFromPtx()
257 port::Status GpuExecutor::LoadModuleFromHsaco(const char* hsaco, in LoadModuleFromHsaco()
[all …]
|
| D | cuda_fft.h | 35 class GpuExecutor; variable
68 port::Status Initialize(GpuExecutor* parent, Stream* stream, int rank,
76 port::Status Initialize(GpuExecutor* parent, Stream* stream, int rank,
89 GpuExecutor* parent_;
111 explicit CUDAFft(GpuExecutor* parent) : parent_(parent) {} in CUDAFft()
117 GpuExecutor* parent_;
|
| D | cuda_blas.h | 44 class GpuExecutor; variable
58 explicit CUDABlas(GpuExecutor *parent);
165 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_gpu_executor.h | 30 using CUDAExecutor = gpu::GpuExecutor;
|
| D | cuda_platform.cc | 146 return GpuExecutor::CreateDeviceDescription(ordinal); in DescriptionForDevice()
175 this, absl::make_unique<GpuExecutor>(config.plugin_config), in GetUncachedExecutor()
|
| D | cuda_fft.cc | 65 bool SetStream(GpuExecutor *parent, cufftHandle plan, Stream *stream) { in SetStream()
78 GpuExecutor *parent, Stream *stream, int rank, uint64 *elem_count, in Initialize()
235 port::Status CUDAFftPlan::Initialize(GpuExecutor *parent, Stream *stream, in Initialize()
570 gpu::GpuExecutor *cuda_executor = in initialize_cufft()
571 dynamic_cast<gpu::GpuExecutor *>(parent); in initialize_cufft()
|
| D | cuda_dnn.h | 32 class GpuExecutor; variable
45 explicit CudnnSupport(GpuExecutor* parent);
582 GpuExecutor* parent_; // Parent executor object. Not owned.
|
| D | cuda_dnn.cc | 183 CudnnHandle GetHandle(GpuExecutor* executor, Stream* stream) { in GetHandle()
307 CudnnSupport::CudnnSupport(GpuExecutor* parent) : parent_(parent) {} in CudnnSupport()
1456 CudnnRnnSequenceTensorDescriptor(GpuExecutor* parent, int max_seq_length, in CudnnRnnSequenceTensorDescriptor()
1475 GpuExecutor* parent, int max_seq_length, int batch_size, int data_size, in Create()
1494 GpuExecutor* parent, int max_seq_length, int batch_size, int data_size, in Create()
1553 CudnnRnnStateTensorDescriptor(GpuExecutor* parent, int num_layers, in CudnnRnnStateTensorDescriptor()
4656 gpu::GpuExecutor* cuda_executor = in initialize_cudnn()
4657 dynamic_cast<gpu::GpuExecutor*>(parent); in initialize_cudnn()
|
| /external/tensorflow/tensorflow/core/util/ |
| D | rocm_solvers.cc | 43 using stream_executor::gpu::GpuExecutor;
53 rocblas_status operator()(GpuExecutor* parent, Args... args) { \
83 rocblas_status operator()(GpuExecutor* parent, Args... args) { \
101 explicit ROCmSolverHandles(GpuExecutor* parent, hipStream_t stream) { in ROCmSolverHandles()
116 GpuExecutor* parent_;
136 GpuExecutor* gpu_executor = static_cast<GpuExecutor*>( in ROCmSolver()
208 static inline Status TrsmImpl(GpuExecutor* gpu_executor, SolverFnT solver, in TrsmImpl()
234 GpuExecutor* gpu_executor = static_cast<GpuExecutor*>( \
|
| /external/tensorflow/tensorflow/stream_executor/gpu/ |
| D | gpu_stream.h | 29 class GpuExecutor; variable
37 explicit GpuStream(GpuExecutor* parent) in GpuStream()
78 GpuExecutor* parent() const { return parent_; } in parent()
81 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 | 37 class GpuExecutor; variable
45 explicit GpuRng(GpuExecutor* parent);
90 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;
340 SE_DISALLOW_COPY_AND_ASSIGN(GpuExecutor);
|
| D | gpu_event.cc | 25 GpuEvent::GpuEvent(GpuExecutor* parent) in GpuEvent()
|