| /external/tensorflow/tensorflow/stream_executor/host/ |
| D | host_gpu_executor.cc | 67 const DeviceMemoryBase &gpu_src, uint64 size) { in Memcpy() argument
70 void *src_mem = const_cast<void *>(gpu_src.opaque()); in Memcpy()
88 const DeviceMemoryBase &gpu_src, in MemcpyDeviceToDevice() argument
91 void *src_mem = const_cast<void *>(gpu_src.opaque()); in MemcpyDeviceToDevice()
138 const DeviceMemoryBase &gpu_src, in SynchronousMemcpy() argument
140 memcpy(host_dst, gpu_src.opaque(), 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 | 75 bool Memcpy(Stream *stream, void *host_dst, const DeviceMemoryBase &gpu_src,
100 const DeviceMemoryBase &gpu_src,
103 const DeviceMemoryBase &gpu_src,
|
| /external/tensorflow/tensorflow/stream_executor/cuda/ |
| D | cuda_driver.cc | 1080 CUdeviceptr gpu_src, in SynchronousMemcpyD2H() argument
1084 CHECK(PointerIsValid(gpu_src)) in SynchronousMemcpyD2H()
1085 << "Source pointer is not actually on GPU: " << gpu_src; in SynchronousMemcpyD2H()
1089 CUresult res = cuMemcpyDtoH(host_dst, gpu_src, size); in SynchronousMemcpyD2H()
1095 absl::bit_cast<void*>(gpu_src), size, size)); in SynchronousMemcpyD2H()
1127 CUdeviceptr gpu_src, in SynchronousMemcpyD2D() argument
1131 CHECK(PointerIsValid(gpu_src)) in SynchronousMemcpyD2D()
1132 << "Source pointer is not actually on GPU: " << gpu_src; in SynchronousMemcpyD2D()
1136 CUresult res = cuMemcpyDtoD(gpu_dst, gpu_src, size); in SynchronousMemcpyD2D()
1142 absl::bit_cast<void*>(gpu_src), size, size)); in SynchronousMemcpyD2D()
[all …]
|
| D | cuda_gpu_executor.cc | 611 const DeviceMemoryBase& gpu_src, in SynchronousMemcpy() argument
614 AsCudaDevicePtr(gpu_src), size); in SynchronousMemcpy()
618 DeviceMemoryBase* gpu_dst, const DeviceMemoryBase& gpu_src, uint64 size) { in SynchronousMemcpyDeviceToDevice() argument
620 AsCudaDevicePtr(gpu_src), size); in SynchronousMemcpyDeviceToDevice()
656 const DeviceMemoryBase& gpu_src, uint64 size) { in Memcpy() argument
658 AsCudaDevicePtr(gpu_src), size, in Memcpy()
671 const DeviceMemoryBase& gpu_src, in MemcpyDeviceToDevice() argument
674 AsCudaDevicePtr(gpu_src), size, in MemcpyDeviceToDevice()
|
| /external/tensorflow/tensorflow/stream_executor/rocm/ |
| D | rocm_driver.cc | 892 GpuContext* context, void* host_dst, hipDeviceptr_t gpu_src, uint64 size) { in SynchronousMemcpyD2H() argument
894 hipError_t res = tensorflow::wrap::hipMemcpyDtoH(host_dst, gpu_src, size); in SynchronousMemcpyD2H()
900 absl::bit_cast<void*>(gpu_src), size, size)); in SynchronousMemcpyD2H()
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()
934 absl::bit_cast<void*>(gpu_src), size, size)); in SynchronousMemcpyD2D()
942 hipDeviceptr_t gpu_src, in AsynchronousMemcpyD2H() argument
947 tensorflow::wrap::hipMemcpyDtoHAsync(host_dst, gpu_src, size, stream); in AsynchronousMemcpyD2H()
952 ToString(res).c_str(), host_dst, absl::bit_cast<void*>(gpu_src), size, in AsynchronousMemcpyD2H()
957 << " bytes from " << absl::bit_cast<void*>(gpu_src) << " to " in AsynchronousMemcpyD2H()
[all …]
|
| D | rocm_gpu_executor.cc | 492 const DeviceMemoryBase& gpu_src, in SynchronousMemcpy() argument
495 AsROCmDevicePtr(gpu_src), size); in SynchronousMemcpy()
499 DeviceMemoryBase* gpu_dst, const DeviceMemoryBase& gpu_src, uint64 size) { in SynchronousMemcpyDeviceToDevice() argument
501 AsROCmDevicePtr(gpu_src), size); in SynchronousMemcpyDeviceToDevice()
537 const DeviceMemoryBase& gpu_src, uint64 size) { in Memcpy() argument
539 AsROCmDevicePtr(gpu_src), size, in Memcpy()
552 const DeviceMemoryBase& gpu_src, in MemcpyDeviceToDevice() argument
555 AsROCmDevicePtr(gpu_src), size, in MemcpyDeviceToDevice()
|
| /external/tensorflow/tensorflow/stream_executor/ |
| D | stream_executor_internal.h | 222 const DeviceMemoryBase &gpu_src,
225 DeviceMemoryBase *gpu_dst, const DeviceMemoryBase &gpu_src,
236 const DeviceMemoryBase &gpu_src, uint64 size) = 0;
240 const DeviceMemoryBase &gpu_src,
|
| D | trace_listener.h | 61 const DeviceMemoryBase& gpu_src, in SynchronousMemcpyD2HBegin() argument
|
| D | stream.h | 1680 Stream &ThenMemcpy(void *host_dst, const DeviceMemoryBase &gpu_src,
1694 Stream &ThenMemcpyD2H(const DeviceMemory<T> &gpu_src, in ThenMemcpyD2H() argument
1697 CHECK(gpu_src.size() == 0 || host_size >= gpu_src.size()); in ThenMemcpyD2H()
1698 return ThenMemcpy(host_dst.begin(), gpu_src, host_size); in ThenMemcpyD2H()
1715 Stream &ThenMemcpy(DeviceMemoryBase *gpu_dst, const DeviceMemoryBase &gpu_src,
1722 const DeviceMemoryBase &gpu_src, uint64 size) { in ThenMemcpyD2D() argument
1723 return ThenMemcpy(gpu_dst, gpu_src, size); in ThenMemcpyD2D()
|
| D | stream.cc | 4780 Stream &Stream::ThenMemcpy(void *host_dst, const DeviceMemoryBase &gpu_src, in ThenMemcpy() argument
4782 VLOG_CALL(PARAM(host_dst), PARAM(gpu_src), PARAM(size)); in ThenMemcpy()
4785 CheckError(parent_->Memcpy(this, host_dst, gpu_src, size)); in ThenMemcpy()
4788 << " did not memcpy device-to-host; source: " << gpu_src.opaque(); in ThenMemcpy()
4807 const DeviceMemoryBase &gpu_src, uint64 size) { 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()
4814 << " did not memcpy gpu-to-gpu; source: " << &gpu_src; in ThenMemcpy()
|
| /external/tensorflow/tensorflow/stream_executor/gpu/ |
| D | gpu_executor.h | 131 const DeviceMemoryBase& gpu_src,
135 const DeviceMemoryBase& gpu_src,
145 bool Memcpy(Stream* stream, void* host_dst, const DeviceMemoryBase& gpu_src,
152 const DeviceMemoryBase& gpu_src,
|
| D | gpu_driver.h | 281 GpuDevicePtr gpu_src, uint64 size);
287 GpuDevicePtr gpu_src, uint64 size);
293 GpuDevicePtr gpu_src, uint64 size,
299 GpuDevicePtr gpu_src, uint64 size,
|
| /external/tensorflow/tensorflow/ |
| D | tensorflow.bzl | 1332 for gpu_src in gpu_srcs:
1333 if gpu_src.endswith(".cc") and not gpu_src.endswith(".cu.cc"):
1335 .format(gpu_src))
|