Searched refs:tile_size_y (Results 1 – 3 of 3) sorted by relevance
| /external/tensorflow/tensorflow/compiler/xla/service/llvm_ir/ |
| D | kernel_tiling.cc | 119 absl::Span<const int64> dims_in_elems, int64 tile_size_y, int64 tile_size_x, in KernelMappingScheme() argument
124 tile_sizes_{1, tile_size_y, tile_size_x}, in KernelMappingScheme()
131 DCHECK_EQ(tile_size_y % num_threads_y_, 0);
|
| D | kernel_tiling.h | 98 KernelMappingScheme(absl::Span<const int64> dims_in_elems, int64 tile_size_y,
|
| /external/tensorflow/tensorflow/compiler/xla/service/gpu/ |
| D | ir_emitter_unnested.cc | 2324 int64 tile_size_y = mapping_scheme->GetTileSizeForDimensionY(); in EmitFullElementalTile() local
2334 /*end=*/llvm::ConstantInt::get(index_ty, tile_size_y), in EmitFullElementalTile()
2424 int64 tile_size_y = mapping_scheme->GetTileSizeForDimensionY(); in EmitTiledElementalCodeWithBoundsCheck() local
2432 builder->CreateICmpEQ(llvm::ConstantInt::get(index_ty, tile_size_y), in EmitTiledElementalCodeWithBoundsCheck()
3656 int64 tile_size_y = 1; in ComputeMappingSchemeAndReductionKind() local
3698 tile_size_y = kNumElementsPerPartialSum; in ComputeMappingSchemeAndReductionKind()
3723 dims_in_elem, tile_size_y, tile_size_x, req_block_sizes, num_threads_y, in ComputeMappingSchemeAndReductionKind()
|