Searched refs:num_threads_y (Results 1 – 2 of 2) sorted by relevance
| /external/tensorflow/tensorflow/compiler/xla/service/gpu/ |
| D | kernel_mapping_scheme.h | 91 absl::Span<const int64> tile_sizes, int64 num_threads_y,
97 num_threads_y_(num_threads_y),
|
| D | ir_emitter_unnested.cc | 4011 llvm::Value* num_threads_y = constant(mapping_scheme.GetNumThreadsY()); in EmitTile() local
4058 num_threads_y), in EmitTile()
4061 thread_id_info.thread_id_y, b_.CreateMul(y_indvar, num_threads_y)); in EmitTile()
5320 int64 num_threads_y = reduction_dimensions.is_row_reduction ? 1 : kWarpSize; in ComputeReductionCodegenInfo() local
5388 {reduction_tiling[0], reduction_tiling[1] * num_threads_y, in ComputeReductionCodegenInfo()
5390 num_threads_y, num_threads_x, indexing_order, vector_size); in ComputeReductionCodegenInfo()
|