| /external/tensorflow/tensorflow/core/kernels/rnn/ |
| D | blas_gemm.cc | 29 se::DeviceMemory<T> AsDeviceMemory(const T* cuda_memory) { in AsDeviceMemory() function
48 auto a_ptr = AsDeviceMemory(a); in operator ()()
49 auto b_ptr = AsDeviceMemory(b); in operator ()()
50 auto c_ptr = AsDeviceMemory(c); in operator ()()
|
| /external/tensorflow/tensorflow/core/kernels/ |
| D | matmul_op_impl.h | 282 se::DeviceMemory<T> AsDeviceMemory(const T* gpu_memory) {
369 a_device_memory.push_back(AsDeviceMemory(a_base_ptr));
370 b_device_memory.push_back(AsDeviceMemory(b_base_ptr));
371 c_device_memory.push_back(AsDeviceMemory(c_base_ptr));
377 a_device_memory.push_back(AsDeviceMemory(a_base_ptr + i * m * k));
378 b_device_memory.push_back(AsDeviceMemory(b_base_ptr + i * k * n));
379 c_device_memory.push_back(AsDeviceMemory(c_base_ptr + i * m * n));
388 a_device_memory.push_back(AsDeviceMemory(a_base_ptr + i * m * k));
391 b_device_memory.push_back(AsDeviceMemory(b_base_ptr + i * k * n));
394 c_device_memory.push_back(AsDeviceMemory(c_base_ptr + i * m * n));
[all …]
|
| D | cudnn_pooling_gpu.cc | 92 auto input_data = AsDeviceMemory(transformed_input.template flat<T>().data(), in Compute()
95 AsDeviceMemory(transformed_output.template flat<T>().data(), in Compute()
228 AsDeviceMemory(transformed_output.template flat<T>().data(), in Compute()
231 AsDeviceMemory(transformed_input.template flat<T>().data(), in Compute()
234 AsDeviceMemory(transformed_output_backprop.template flat<T>().data(), in Compute()
237 AsDeviceMemory(transformed_input_backprop.template flat<T>().data(), in Compute()
|
| D | lrn_op.cc | 216 auto input_data = StreamExecutorUtil::AsDeviceMemory<T>(in); in launch()
217 auto output_data = StreamExecutorUtil::AsDeviceMemory<T>(*output); in launch()
270 AsDeviceMemory(transformed_input.template flat<T>().data(), in launch()
273 AsDeviceMemory(transformed_output.template flat<T>().data(), in launch()
508 auto input_grads_data = StreamExecutorUtil::AsDeviceMemory<T>(in_grads); in launch()
509 auto input_image_data = StreamExecutorUtil::AsDeviceMemory<T>(in_image); in launch()
510 auto output_image_data = StreamExecutorUtil::AsDeviceMemory<T>(out_image); in launch()
511 auto output_grads_data = StreamExecutorUtil::AsDeviceMemory<T>(*output); in launch()
584 AsDeviceMemory(transformed_in_grads.template flat<T>().data(), in launch()
587 AsDeviceMemory(transformed_in_image.template flat<T>().data(), in launch()
[all …]
|
| D | cudnn_rnn_ops.cc | 293 const DeviceMemory<T> AsDeviceMemory(const Tensor* tensor) { in AsDeviceMemory() function
300 DeviceMemory<T> AsDeviceMemory(Tensor* tensor) { in AsDeviceMemory() function
466 return AsDeviceMemory<uint8>(handle_.AccessTensor(context_)); in AllocateBytes()
813 auto input_data = AsDeviceMemory<T>(input); in DoForward()
814 auto input_h_data = AsDeviceMemory<T>(input_h); in DoForward()
817 input_c_data = AsDeviceMemory<T>(input_c); in DoForward()
820 auto params_data = AsDeviceMemory<T>(params); in DoForward()
821 auto output_data = AsDeviceMemory<T>(output); in DoForward()
822 auto output_h_data = AsDeviceMemory<T>(output_h); in DoForward()
825 output_c_data = AsDeviceMemory<T>(output_c); in DoForward()
[all …]
|
| D | fused_batch_norm_op.cc | 860 y_ptr = StreamExecutorUtil::AsDeviceMemory<T>(*y); in operator ()()
878 y_ptr = StreamExecutorUtil::AsDeviceMemory<T>(y_transformed); in operator ()()
904 auto x_ptr = StreamExecutorUtil::AsDeviceMemory<T>(x_maybe_transformed); in operator ()()
905 auto scale_ptr = StreamExecutorUtil::AsDeviceMemory<U>(scale); in operator ()()
906 auto offset_ptr = StreamExecutorUtil::AsDeviceMemory<U>(offset); in operator ()()
908 StreamExecutorUtil::AsDeviceMemory<U>(estimated_mean); in operator ()()
910 StreamExecutorUtil::AsDeviceMemory<U>(estimated_variance); in operator ()()
913 ? StreamExecutorUtil::AsDeviceMemory<U>(*side_input) in operator ()()
915 auto batch_mean_ptr = StreamExecutorUtil::AsDeviceMemory<U>(*batch_mean); in operator ()()
917 auto batch_var_ptr = StreamExecutorUtil::AsDeviceMemory<U>(*batch_var); in operator ()()
[all …]
|
| D | conv_grad_input_ops.cc | 141 auto a_ptr = AsDeviceMemory(out_backprop.template flat<T>().data(), in operator ()()
143 auto b_ptr = AsDeviceMemory(filter.template flat<T>().data(), in operator ()()
145 auto c_ptr = AsDeviceMemory(in_backprop->template flat<T>().data(), in operator ()()
174 auto a_ptr = AsDeviceMemory(out_backprop.template flat<T>().data(), in operator ()()
176 auto b_ptr = AsDeviceMemory(filter.template flat<T>().data(), in operator ()()
178 auto c_ptr = AsDeviceMemory(in_backprop->template flat<T>().data(), in operator ()()
347 AsDeviceMemory(transformed_out_backprop.template flat<T>().data(), in operator ()()
350 AsDeviceMemory(transformed_filter.template flat<T>().data(), in operator ()()
353 AsDeviceMemory(pre_transformed_in_backprop.template flat<T>().data(), in operator ()()
|
| D | conv_ops_3d.cc | 260 auto a_ptr = AsDeviceMemory(input.template flat<T>().data(), in launch()
262 auto b_ptr = AsDeviceMemory(filter.template flat<T>().data(), in launch()
264 auto c_ptr = AsDeviceMemory(output->template flat<T>().data(), in launch()
287 auto a_ptr = AsDeviceMemory(input.template flat<T>().data(), in launch()
289 auto b_ptr = AsDeviceMemory(filter.template flat<T>().data(), in launch()
291 auto c_ptr = AsDeviceMemory(output->template flat<T>().data(), in launch()
467 auto input_ptr = AsDeviceMemory(input.template flat<T>().data(), in launch()
470 AsDeviceMemory(transformed_filter.template flat<T>().data(), in launch()
473 AsDeviceMemory(transformed_output.template flat<T>().data(), in launch()
|
| D | conv_grad_ops_3d.cc | 1191 auto a_ptr = AsDeviceMemory(out_backprop.template flat<T>().data(), in Compute()
1193 auto b_ptr = AsDeviceMemory(filter.template flat<T>().data(), in Compute()
1195 auto c_ptr = AsDeviceMemory(in_backprop->template flat<T>().data(), in Compute()
1221 auto a_ptr = AsDeviceMemory(out_backprop.template flat<T>().data(), in Compute()
1223 auto b_ptr = AsDeviceMemory(filter.template flat<T>().data(), in Compute()
1225 auto c_ptr = AsDeviceMemory(in_backprop->template flat<T>().data(), in Compute()
1385 AsDeviceMemory(transformed_out_backprop.template flat<T>().data(), in Compute()
1388 AsDeviceMemory(transformed_filter.template flat<T>().data(), in Compute()
1391 AsDeviceMemory(pre_transformed_in_backprop.template flat<T>().data(), in Compute()
1677 auto a_ptr = AsDeviceMemory(out_backprop.template flat<T>().data(), in Compute()
[all …]
|
| D | pooling_ops_common.cc | 377 AsDeviceMemory(reinterpret_cast<const typename RawType<T>::type*>( in Compute()
382 AsDeviceMemory(reinterpret_cast<const typename RawType<T>::type*>( in Compute()
692 AsDeviceMemory(transformed_output.template flat<T>().data(), in Compute()
695 AsDeviceMemory(transformed_input.template flat<T>().data(), in Compute()
698 AsDeviceMemory(transformed_output_backprop.template flat<T>().data(), in Compute()
700 auto input_backprop_data = AsDeviceMemory( in Compute()
|
| D | conv_grad_filter_ops.cc | 727 auto a_ptr = AsDeviceMemory(out_backprop.template flat<T>().data(), in operator ()()
733 auto b_ptr = AsDeviceMemory(input.template flat<T>().data(), in operator ()()
739 auto c_ptr = AsDeviceMemory(filter_backprop->template flat<T>().data(), in operator ()()
766 auto a_ptr = AsDeviceMemory(input.template flat<T>().data(), in operator ()()
768 auto b_ptr = AsDeviceMemory(out_backprop.template flat<T>().data(), in operator ()()
770 auto c_ptr = AsDeviceMemory(filter_backprop->template flat<T>().data(), in operator ()()
944 AsDeviceMemory(transformed_out_backprop.template flat<T>().data(), in operator ()()
947 AsDeviceMemory(pre_transformed_filter_backprop.template flat<T>().data(), in operator ()()
949 auto input_ptr = AsDeviceMemory(transformed_input.template flat<T>().data(), in operator ()()
|
| D | fft_ops.cc | 344 se::DeviceMemory<T> AsDeviceMemory(const T* cuda_memory) { in AsDeviceMemory() function
351 se::DeviceMemory<T> AsDeviceMemory(const T* cuda_memory, uint64 size) { in AsDeviceMemory() function
387 AsDeviceMemory(temporary_memory.flat<uint8>().data(), in AllocateBytes()
532 auto src = AsDeviceMemory<InT>(in.flat<InT>().data()); in DoFFTInternal()
533 auto dst = AsDeviceMemory<OutT>(out->flat<OutT>().data()); in DoFFTInternal()
|
| D | conv_ops.cc | 687 auto a_ptr = AsDeviceMemory(input.template flat<T>().data(), in operator ()()
689 auto b_ptr = AsDeviceMemory(filter.template flat<T>().data(), in operator ()()
691 auto c_ptr = AsDeviceMemory(output->template flat<T>().data(), in operator ()()
715 auto a_ptr = AsDeviceMemory(input.template flat<T>().data(), in operator ()()
717 auto b_ptr = AsDeviceMemory(filter.template flat<T>().data(), in operator ()()
719 auto c_ptr = AsDeviceMemory(output->template flat<T>().data(), in operator ()()
951 auto input_ptr = AsDeviceMemory(input.template flat<T>().data(), in operator ()()
954 AsDeviceMemory(transformed_filter.template flat<T>().data(), in operator ()()
957 AsDeviceMemory(transformed_output.template flat<T>().data(), in operator ()()
|
| D | ctc_loss_op.cc | 342 auto probs_data = StreamExecutorUtil::AsDeviceMemory<float>(*inputs); in Compute()
343 auto costs_data = StreamExecutorUtil::AsDeviceMemory<float>(*loss); in Compute()
344 auto grads_data = StreamExecutorUtil::AsDeviceMemory<float>(*gradient); in Compute()
|
| D | conv_ops_gpu.h | 86 AsDeviceMemory(temporary_memory.flat<uint8>().data(), in AllocateBytes()
|
| D | gpu_utils.h | 63 inline se::DeviceMemory<T> AsDeviceMemory(const T* cuda_memory, uint64 size) { in AsDeviceMemory() function
|
| D | conv_ops_fused_impl.h | 630 return AsDeviceMemory(t.template flat<T>().data(),
641 AsDeviceMemory(static_cast<T*>(nullptr), 0);
|
| /external/tensorflow/tensorflow/core/kernels/linalg/ |
| D | matrix_triangular_solve_op_impl.h | 53 se::DeviceMemory<Scalar> AsDeviceMemory(const Scalar* gpu_memory) { in AsDeviceMemory() function
267 auto src_device_mem = AsDeviceMemory(in_y.template flat<Scalar>().data());
268 auto dst_device_mem = AsDeviceMemory(out->template flat<Scalar>().data());
286 auto src_device_mem = AsDeviceMemory(b_tmp_ptrs[b_batch_indices[i]]);
288 AsDeviceMemory(out->template flat<Scalar>().data() + i * m * n);
|
| D | tridiagonal_solve_op_gpu.cu.cc | 71 se::DeviceMemory<Scalar> AsDeviceMemory(const Scalar* cuda_memory) { in AsDeviceMemory() function
80 auto src_device_mem = AsDeviceMemory(src); in CopyDeviceToDevice()
81 auto dst_device_mem = AsDeviceMemory(dst); in CopyDeviceToDevice()
|
| /external/tensorflow/tensorflow/core/util/ |
| D | stream_executor_util.h | 33 static se::DeviceMemory<T> AsDeviceMemory(const Tensor& t) { in AsDeviceMemory() function
|
| /external/tensorflow/tensorflow/core/common_runtime/gpu/ |
| D | gpu_util.h | 73 static se::DeviceMemory<T> AsDeviceMemory(const Tensor& t) { in AsDeviceMemory() function
|
| /external/tensorflow/tensorflow/core/nccl/ |
| D | nccl_manager_test.cc | 156 auto in_gpu_mem = AsDeviceMemory(in_gpu.flat<Scalar>().data()); in MakeReductionTestCase()
197 auto in_gpu_mem = AsDeviceMemory(in_gpu.flat<Scalar>().data()); in MakeGatherTestCase()
233 auto in_gpu_mem = AsDeviceMemory(in_gpu.flat<Scalar>().data()); in MakeBroadcastTestCase()
269 auto out_gpu_mem = AsDeviceMemory(out_gpu.flat<Scalar>().data()); in VerifyResults()
439 static se::DeviceMemory<Scalar> AsDeviceMemory(const Scalar* cuda_memory) { in AsDeviceMemory() function in tensorflow::NcclManagerTest
|