/external/tensorflow/tensorflow/core/common_runtime/gpu/ |
gpu_util_platform_specific.cc | 25 void GPUDeviceContext::CopyCPUTensorToDevice(const Tensor* cpu_tensor, 29 GPUUtil::CopyCPUTensorToGPU(cpu_tensor, this, device, device_tensor, done); 34 Device* device, Tensor* cpu_tensor, 36 GPUUtil::CopyGPUTensorToCPU(device, this, device_tensor, cpu_tensor, done);
|
gpu_util.h | 45 // Copies the data in 'gpu_tensor' into 'cpu_tensor'. 47 // 'cpu_tensor' must be allocated to be of the same size as 51 const Tensor* gpu_tensor, Tensor* cpu_tensor, 91 static void CopyCPUTensorToGPU(const Tensor* cpu_tensor,
|
gpu_util.cc | 261 const Tensor* gpu_tensor, Tensor* cpu_tensor, 266 Status s = PrepareCopy(gpu_device, device_context, *gpu_tensor, cpu_tensor, 287 void* dst_ptr = GetBase(cpu_tensor); 304 void GPUUtil::CopyCPUTensorToGPU(const Tensor* cpu_tensor, 311 Status s = PrepareCopy(gpu_device, device_context, *cpu_tensor, gpu_tensor, 328 const int64 total_bytes = cpu_tensor->TotalBytes(); 331 void* src_ptr = GetBase(cpu_tensor); 336 // Use of cpu_tensor may outlive stack scope, so keep a ref. 337 TensorReference input_ref(*cpu_tensor);
|
/external/tensorflow/tensorflow/compiler/jit/ |
xla_device_context.cc | 57 void XlaTransferManager::CopyCPUTensorToDevice(const Tensor* cpu_tensor, 61 if (cpu_tensor->NumElements() > 0) { 63 << reinterpret_cast<const void*>(cpu_tensor->tensor_data().data()) 67 << " " << cpu_tensor->NumElements(); 69 void* src_ptr = const_cast<void*>(DMAHelper::base(cpu_tensor)); 70 const int64 total_bytes = cpu_tensor->TotalBytes(); 95 Tensor* cpu_tensor, 102 << reinterpret_cast<const void*>(cpu_tensor->tensor_data().data()) 105 const int64 total_bytes = cpu_tensor->TotalBytes(); 108 void* dst_ptr = DMAHelper::base(cpu_tensor); [all...] |
xla_device_context.h | 54 void CopyCPUTensorToDevice(const Tensor* cpu_tensor, Device* device, 58 Tensor* cpu_tensor, StatusCallback done); 74 void CopyCPUTensorToDevice(const Tensor* cpu_tensor, Device* device, 79 Tensor* cpu_tensor, StatusCallback done) override;
|
/external/tensorflow/tensorflow/core/debug/ |
debug_gateway.cc | 84 Tensor cpu_tensor(cpu_allocator, src_tensor->dtype(), src_tensor->shape()); 98 src_tensor, "TensorCopy", device, &cpu_tensor, 99 [node_name, cpu_tensor, copy_done_cb](const Status& s) { 101 copy_done_cb(&cpu_tensor); 111 cpu_tensor.UnsafeCopyFromInternal(*src_tensor, src_tensor->dtype(), 114 copy_done_cb(&cpu_tensor);
|
/external/tensorflow/tensorflow/core/common_runtime/sycl/ |
sycl_device_context.h | 34 void CopyCPUTensorToDevice(const Tensor *cpu_tensor, Device *device, 39 Device *device, Tensor *cpu_tensor,
|
sycl_device_context.cc | 25 void SYCLDeviceContext::CopyCPUTensorToDevice(const Tensor *cpu_tensor, 29 const int64 total_bytes = cpu_tensor->TotalBytes(); 31 const void *src_ptr = DMAHelper::base(cpu_tensor); 33 switch (cpu_tensor->dtype()) { 105 Tensor *cpu_tensor, 110 void *dst_ptr = DMAHelper::base(cpu_tensor);
|
/external/tensorflow/tensorflow/core/common_runtime/ |
gpu_device_context.h | 55 void CopyCPUTensorToDevice(const Tensor* cpu_tensor, Device* device, 60 Device* device, Tensor* cpu_tensor,
|
copy_tensor.cc | 280 Tensor* cpu_tensor = local 283 [cpu_tensor](StatusCallback done_, 286 delete cpu_tensor; 291 [delete_and_done, recv_dev_context, cpu_tensor, cpu_allocator, 299 CopyHostToDevice(cpu_tensor, cpu_allocator, out_allocator, edge_name, 305 cpu_tensor, send_dev_context,
|
/external/tensorflow/tensorflow/core/framework/ |
device_base.h | 75 // "cpu_tensor" is a tensor on a CPU. Copies "cpu_tensor" into 77 // must be allocated to be of the same size as "cpu_tensor". 78 virtual void CopyCPUTensorToDevice(const Tensor* cpu_tensor, Device* device, 85 // device_tensor into "cpu_tensor". "cpu_tensor" must be allocated 89 Tensor* cpu_tensor, StatusCallback done) {
|
/external/tensorflow/tensorflow/core/kernels/ |
stack_ops.cc | 295 Tensor* cpu_tensor = variable 298 &tensor, "StackPush", device, cpu_tensor, 299 [cpu_tensor, stack, ctx, done](const Status& s) { 303 ctx->SetStatus(stack->Push({*cpu_tensor, alloc_attrs, true})); 306 ctx->set_output(0, *cpu_tensor); 309 delete cpu_tensor; 414 Tensor* cpu_tensor = &value.tensor; variable 417 new Tensor(gpu_allocator, cpu_tensor->dtype(), cpu_tensor->shape()); 419 cpu_tensor, device, device_tensor [all...] |
dynamic_partition_op_gpu.cu.cc | 279 Tensor cpu_tensor; local 286 &cpu_tensor, alloc_attr), 292 ->ThenMemcpy(cpu_tensor.flat<int32>().data(), wrapped, 304 partition_ref, cpu_tensor, done]() { 306 this->AllocateOutputs(c, &data, &partitions, &cpu_tensor, &outputs, done);
|
/external/tensorflow/tensorflow/python/eager/ |
ops_test.py | 272 cpu_tensor = constant_op.constant(1.0) 273 gpu_tensor = cpu_tensor.gpu() 274 self.assertAllEqual(cpu_tensor + gpu_tensor, 2.0)
|