| /external/tensorflow/tensorflow/contrib/reduce_slice_ops/kernels/ |
| reduce_slice_ops_gpu.cu.cc | 37 CUDA_AXIS_KERNEL_LOOP(x, config.virtual_thread_count.x, X) { \
38 CUDA_AXIS_KERNEL_LOOP(y, config.virtual_thread_count.y, Y) { \
39 CUDA_AXIS_KERNEL_LOOP(z, config.virtual_thread_count.z, Z) { \
40 Index outidx = x * config.virtual_thread_count.y * \
41 config.virtual_thread_count.z + \
42 y * config.virtual_thread_count.z + z; \
47 Index inidx = x * bound * config.virtual_thread_count.z + \
48 yin * config.virtual_thread_count.z + z; \
|
| /external/tensorflow/tensorflow/core/util/ |
| cuda_launch_config.h | 51 CUDA_1D_KERNEL_LOOP(x, config.virtual_thread_count) {
57 CUDA_AXIS_KERNEL_LOOP(x, config.virtual_thread_count, x) {
58 CUDA_AXIS_KERNEL_LOOP(y, config.virtual_thread_count, y) {
65 CUDA_AXIS_KERNEL_LOOP(x, config.virtual_thread_count, x) {
66 CUDA_AXIS_KERNEL_LOOP(y, config.virtual_thread_count, y) {
67 CUDA_AXIS_KERNEL_LOOP(z, config.virtual_thread_count, z) {
114 int virtual_thread_count = -1; member in struct:tensorflow::CudaLaunchConfig
129 const int virtual_thread_count = work_element_count; local
132 virtual_thread_count);
138 config.virtual_thread_count = virtual_thread_count
197 dim3 virtual_thread_count = dim3(0, 0, 0); member in struct:tensorflow::Cuda2DLaunchConfig
[all...] |
| cuda_kernel_helper_test.cu.cc | 42 CUDA_1D_KERNEL_LOOP(x, config.virtual_thread_count) { outbuf[x] = 0; }
47 CUDA_1D_KERNEL_LOOP(x, config.virtual_thread_count) {
55 CUDA_AXIS_KERNEL_LOOP(x, config.virtual_thread_count.x, X) {
59 CUDA_AXIS_KERNEL_LOOP(y, config.virtual_thread_count.y, Y) {
63 int idx = x * config.virtual_thread_count.y + y;
69 CUDA_AXIS_KERNEL_LOOP(x, config.virtual_thread_count.x, X) {
73 CUDA_AXIS_KERNEL_LOOP(y, config.virtual_thread_count.y, Y) {
77 CUDA_AXIS_KERNEL_LOOP(z, config.virtual_thread_count.z, Z) {
82 x * config.virtual_thread_count.y * config.virtual_thread_count.z
[all...] |
| /external/tensorflow/tensorflow/core/kernels/ |
| diag_op_gpu.cu.cc | 52 // CudaLaunchConfig uses an int for virtual_thread_count,
58 int virtual_thread_count = int(size * size); local
63 GetCudaLaunchConfig(virtual_thread_count, device);
65 device.stream()>>>(diag_config.virtual_thread_count, size,
105 0, device.stream()>>>(diag_config.virtual_thread_count,
|
| fused_batch_norm_op.cu.cc | 42 d.stream()>>>(config.virtual_thread_count,
64 d.stream()>>>(config.virtual_thread_count,
|
| matrix_set_diag_op_gpu.cu.cc | 78 config.virtual_thread_count, m, n, minsize, diag.data(),
84 config.virtual_thread_count, m, n, minsize, input.data(),
|
| eye_functor_gpu.cu.cc | 57 device.stream()>>>(config.virtual_thread_count, batch_size, m,
|
| matrix_band_part_op_gpu.cu.cc | 63 config.virtual_thread_count, batch_size, m, n, num_lower_diags,
|
| inplace_ops_functor_gpu.cu.cc | 54 cfg.virtual_thread_count, nrows, ncols, loc, src, dst);
|
| resize_bilinear_op_gpu.cu.cc | 169 config.virtual_thread_count, images.data(), height_scale,
198 config.virtual_thread_count, output_grad.data());
205 config.virtual_thread_count, input_grad.data(), height_scale,
|
| tile_functor_gpu.cu.cc | 76 cfg.virtual_thread_count, p, reinterpret_cast<const int32*>(dev_buf),
|
| depthtospace_op_gpu.cu.cc | 163 config.virtual_thread_count, input.data(), block_size, batch_size,
218 config.virtual_thread_count, input.data(), block_size, input_width,
|
| determinant_op_gpu.cu.cc | 134 config.virtual_thread_count, n, lu_factor.data(), pivots, nullptr,
155 config.virtual_thread_count, n, lu_factor.data(), pivots,
|
| spacetodepth_op_gpu.cu.cc | 159 config.virtual_thread_count, input.data(), block_size, batch_size,
214 config.virtual_thread_count, input.data(), block_size, output_width,
|
| crop_and_resize_op_gpu.cu.cc | 345 config.virtual_thread_count, image.data(), boxes.data(),
377 config.virtual_thread_count, grads_image.data());
386 config.virtual_thread_count, grads.data(), boxes.data(),
419 config.virtual_thread_count, grads_boxes.data());
428 config.virtual_thread_count, grads.data(), image.data(), boxes.data(),
|
| dilation_ops_gpu.cu.cc | 201 config.virtual_thread_count, input.data(), filter.data(), batch,
241 config.virtual_thread_count, input.data(), filter.data(),
281 config.virtual_thread_count, input.data(), filter.data(),
|
| svd_op_gpu.cu.cc | 66 CUDA_AXIS_KERNEL_LOOP(batch, config.virtual_thread_count.x, X) {
67 CUDA_AXIS_KERNEL_LOOP(i, config.virtual_thread_count.y, Y) {
78 CUDA_1D_KERNEL_LOOP(i, config.virtual_thread_count) {
|
| avgpooling_op_gpu.cu.cc | 95 config.virtual_thread_count, top_diff, num, height, width, channels,
|
| multinomial_op_gpu.cu.cc | 109 d.stream()>>>(config.virtual_thread_count, num_classes,
|
| spacetobatch_functor_gpu.cu.cc | 146 config.virtual_thread_count, const_cast<T*>(space_tensor.data()),
|
| bias_op_gpu.cu.cc | 82 config.virtual_thread_count, input, bias, output, bias_size);
86 config.virtual_thread_count, input, bias, output, bias_size,
|
| conv_ops_gpu_3.cu.cc | 449 config.virtual_thread_count, in.data(), combined_dims, out.data());
469 config.virtual_thread_count, in.data(), combined_dims, out.data());
499 config.virtual_thread_count, in.data(), input_dims, out.data(),
504 config.virtual_thread_count, in.data(), input_dims, out.data(),
[all...] |
| resize_nearest_neighbor_op_gpu.cu.cc | 158 d.stream()>>>(input_config.virtual_thread_count, input.data(),
|
| transpose_functor_gpu.cu.cc | 85 cfg.virtual_thread_count, p, reinterpret_cast<const int32*>(dev_buf),
|
