Searched refs:kThreadsPerBlock (Results 1 – 3 of 3) sorted by relevance
| /external/tensorflow/tensorflow/core/kernels/ |
| D | maxpooling_op_gpu.cu.cc | 391 const int kThreadsPerBlock = 1024; in operator ()() local
395 (output_size + kThreadsPerBlock - 1) / kThreadsPerBlock, kThreadsPerBlock, in operator ()()
410 const int kThreadsPerBlock = 1024; in operator ()() local
415 <<<(output_size + kThreadsPerBlock - 1) / kThreadsPerBlock, in operator ()()
416 kThreadsPerBlock, 0, d.stream()>>>( in operator ()()
422 <<<(output_size + kThreadsPerBlock - 1) / kThreadsPerBlock, in operator ()()
423 kThreadsPerBlock, 0, d.stream()>>>( in operator ()()
438 const int kThreadsPerBlock = 1024; in operator ()() local
442 SetZero<<<(bottom_size + kThreadsPerBlock - 1) / kThreadsPerBlock, in operator ()()
443 kThreadsPerBlock, 0, d.stream()>>>(bottom_size, bottom_diff); in operator ()()
[all …]
|
| /external/tensorflow/tensorflow/core/util/ |
| D | cuda_launch_config.h | 210 const int kThreadsPerBlock = 256; in GetCuda2DLaunchConfig() local
211 int block_cols = std::min(xdim, kThreadsPerBlock); in GetCuda2DLaunchConfig()
213 int block_rows = std::max(kThreadsPerBlock / block_cols, 1); in GetCuda2DLaunchConfig()
218 const int max_blocks = std::max(physical_thread_count / kThreadsPerBlock, 1); in GetCuda2DLaunchConfig()
|
| /external/tensorflow/tensorflow/compiler/xla/service/gpu/ |
| D | ir_emitter_unnested.cc | 1297 const uint64 kThreadsPerBlock = kTileSize / 2; in HandleSort() local
1310 kThreadsPerBlock > in HandleSort()
1315 uint64 num_blocks = CeilOfRatio(num_iterations, kThreadsPerBlock); in HandleSort()
1316 LaunchDimensions tiled_launch_dimensions(num_blocks, kThreadsPerBlock); in HandleSort()
|