| /external/tensorflow/tensorflow/stream_executor/host/ |
| D | host_gpu_executor.cc | 76 bool HostExecutor::Memcpy(Stream *stream, DeviceMemoryBase *gpu_dst, in Memcpy() argument
78 void *dst_mem = gpu_dst->opaque(); in Memcpy()
87 DeviceMemoryBase *gpu_dst, in MemcpyDeviceToDevice() argument
90 void *dst_mem = gpu_dst->opaque(); in MemcpyDeviceToDevice()
130 port::Status HostExecutor::SynchronousMemcpy(DeviceMemoryBase *gpu_dst, in SynchronousMemcpy() argument
133 memcpy(gpu_dst->opaque(), host_src, size); in SynchronousMemcpy()
145 DeviceMemoryBase *gpu_dst, const DeviceMemoryBase &gpu_src, uint64 size) { in SynchronousMemcpyDeviceToDevice() argument
146 memcpy(gpu_dst->opaque(), gpu_src.opaque(), size); in SynchronousMemcpyDeviceToDevice()
|
| D | host_gpu_executor.h | 77 bool Memcpy(Stream *stream, DeviceMemoryBase *gpu_dst, const void *host_src,
79 bool MemcpyDeviceToDevice(Stream *stream, DeviceMemoryBase *gpu_dst,
97 port::Status SynchronousMemcpy(DeviceMemoryBase *gpu_dst,
102 port::Status SynchronousMemcpyDeviceToDevice(DeviceMemoryBase *gpu_dst,
|
| /external/tensorflow/tensorflow/stream_executor/cuda/ |
| D | cuda_driver.cc | 1103 CUdeviceptr gpu_dst, in SynchronousMemcpyH2D() argument
1110 CHECK(PointerIsValid(gpu_dst)) in SynchronousMemcpyH2D()
1111 << "Destination pointer is not actually on GPU: " << gpu_dst; in SynchronousMemcpyH2D()
1113 CUresult res = cuMemcpyHtoD(gpu_dst, host_src, size); in SynchronousMemcpyH2D()
1118 ToString(res).c_str(), absl::bit_cast<void*>(gpu_dst), host_src, size, in SynchronousMemcpyH2D()
1126 CUdeviceptr gpu_dst, in SynchronousMemcpyD2D() argument
1133 CHECK(PointerIsValid(gpu_dst)) in SynchronousMemcpyD2D()
1134 << "Destination pointer is not actually on GPU: " << gpu_dst; in SynchronousMemcpyD2D()
1136 CUresult res = cuMemcpyDtoD(gpu_dst, gpu_src, size); in SynchronousMemcpyD2D()
1141 ToString(res).c_str(), absl::bit_cast<void*>(gpu_dst), in SynchronousMemcpyD2D()
[all …]
|
| D | cuda_gpu_executor.cc | 604 port::Status GpuExecutor::SynchronousMemcpy(DeviceMemoryBase* gpu_dst, in SynchronousMemcpy() argument
606 return GpuDriver::SynchronousMemcpyH2D(context_, AsCudaDevicePtr(gpu_dst), in SynchronousMemcpy()
618 DeviceMemoryBase* gpu_dst, const DeviceMemoryBase& gpu_src, uint64 size) { in SynchronousMemcpyDeviceToDevice() argument
619 return GpuDriver::SynchronousMemcpyD2D(context_, AsCudaDevicePtr(gpu_dst), in SynchronousMemcpyDeviceToDevice()
662 bool GpuExecutor::Memcpy(Stream* stream, DeviceMemoryBase* gpu_dst, in Memcpy() argument
664 return GpuDriver::AsynchronousMemcpyH2D(context_, AsCudaDevicePtr(gpu_dst), in Memcpy()
670 DeviceMemoryBase* gpu_dst, in MemcpyDeviceToDevice() argument
673 return GpuDriver::AsynchronousMemcpyD2D(context_, AsCudaDevicePtr(gpu_dst), in MemcpyDeviceToDevice()
|
| /external/tensorflow/tensorflow/stream_executor/rocm/ |
| D | rocm_driver.cc | 908 GpuContext* context, hipDeviceptr_t gpu_dst, const void* host_src, in SynchronousMemcpyH2D() argument
912 gpu_dst, const_cast<void*>(host_src), size); in SynchronousMemcpyH2D()
917 ToString(res).c_str(), absl::bit_cast<void*>(gpu_dst), host_src, size, in SynchronousMemcpyH2D()
925 GpuContext* context, hipDeviceptr_t gpu_dst, hipDeviceptr_t gpu_src, in SynchronousMemcpyD2D() argument
928 hipError_t res = tensorflow::wrap::hipMemcpyDtoD(gpu_dst, gpu_src, size); in SynchronousMemcpyD2D()
933 ToString(res).c_str(), absl::bit_cast<void*>(gpu_dst), in SynchronousMemcpyD2D()
963 hipDeviceptr_t gpu_dst, in AsynchronousMemcpyH2D() argument
969 gpu_dst, const_cast<void*>(host_src), size, stream); in AsynchronousMemcpyH2D()
974 ToString(res).c_str(), absl::bit_cast<void*>(gpu_dst), host_src, size, in AsynchronousMemcpyH2D()
984 hipDeviceptr_t gpu_dst, in AsynchronousMemcpyD2D() argument
[all …]
|
| D | rocm_gpu_executor.cc | 485 port::Status GpuExecutor::SynchronousMemcpy(DeviceMemoryBase* gpu_dst, in SynchronousMemcpy() argument
487 return GpuDriver::SynchronousMemcpyH2D(context_, AsROCmDevicePtr(gpu_dst), in SynchronousMemcpy()
499 DeviceMemoryBase* gpu_dst, const DeviceMemoryBase& gpu_src, uint64 size) { in SynchronousMemcpyDeviceToDevice() argument
500 return GpuDriver::SynchronousMemcpyD2D(context_, AsROCmDevicePtr(gpu_dst), in SynchronousMemcpyDeviceToDevice()
543 bool GpuExecutor::Memcpy(Stream* stream, DeviceMemoryBase* gpu_dst, in Memcpy() argument
545 return GpuDriver::AsynchronousMemcpyH2D(context_, AsROCmDevicePtr(gpu_dst), in Memcpy()
551 DeviceMemoryBase* gpu_dst, in MemcpyDeviceToDevice() argument
554 return GpuDriver::AsynchronousMemcpyD2D(context_, AsROCmDevicePtr(gpu_dst), in MemcpyDeviceToDevice()
|
| /external/tensorflow/tensorflow/stream_executor/ |
| D | stream_executor_internal.h | 219 virtual port::Status SynchronousMemcpy(DeviceMemoryBase *gpu_dst,
225 DeviceMemoryBase *gpu_dst, const DeviceMemoryBase &gpu_src,
237 virtual bool Memcpy(Stream *stream, DeviceMemoryBase *gpu_dst,
239 virtual bool MemcpyDeviceToDevice(Stream *stream, DeviceMemoryBase *gpu_dst,
|
| D | trace_listener.h | 56 DeviceMemoryBase* gpu_dst) {} in SynchronousMemcpyH2DBegin() argument
|
| D | stream.h | 1687 Stream &ThenMemcpy(DeviceMemoryBase *gpu_dst, const void *host_src,
1706 DeviceMemory<T> *gpu_dst) { in ThenMemcpyH2D() argument
1708 CHECK(gpu_dst->size() == 0 || gpu_dst->size() >= host_size); in ThenMemcpyH2D()
1709 return ThenMemcpy(gpu_dst, host_src.begin(), host_size); in ThenMemcpyH2D()
1715 Stream &ThenMemcpy(DeviceMemoryBase *gpu_dst, const DeviceMemoryBase &gpu_src,
1721 Stream &ThenMemcpyD2D(DeviceMemoryBase *gpu_dst, in ThenMemcpyD2D() argument
1723 return ThenMemcpy(gpu_dst, gpu_src, size); in ThenMemcpyD2D()
|
| D | stream.cc | 4793 Stream &Stream::ThenMemcpy(DeviceMemoryBase *gpu_dst, const void *host_src, in ThenMemcpy() argument
4795 VLOG_CALL(PARAM(gpu_dst), PARAM(host_src), PARAM(size)); in ThenMemcpy()
4798 CheckError(parent_->Memcpy(this, gpu_dst, host_src, size)); in ThenMemcpy()
4806 Stream &Stream::ThenMemcpy(DeviceMemoryBase *gpu_dst, in ThenMemcpy() argument
4808 VLOG_CALL(PARAM(gpu_dst), PARAM(gpu_src), PARAM(size)); in ThenMemcpy()
4811 CheckError(parent_->MemcpyDeviceToDevice(this, gpu_dst, gpu_src, size)); in ThenMemcpy()
|
| /external/tensorflow/tensorflow/stream_executor/gpu/ |
| D | gpu_executor.h | 127 port::Status SynchronousMemcpy(DeviceMemoryBase* gpu_dst,
134 port::Status SynchronousMemcpyDeviceToDevice(DeviceMemoryBase* gpu_dst,
148 bool Memcpy(Stream* stream, DeviceMemoryBase* gpu_dst, const void* host_src,
151 bool MemcpyDeviceToDevice(Stream* stream, DeviceMemoryBase* gpu_dst,
|
| D | gpu_driver.h | 283 GpuDevicePtr gpu_dst,
286 GpuDevicePtr gpu_dst,
295 static bool AsynchronousMemcpyH2D(GpuContext* context, GpuDevicePtr gpu_dst,
298 static bool AsynchronousMemcpyD2D(GpuContext* context, GpuDevicePtr gpu_dst,
|