Searched refs:tile_size_x (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()
132 DCHECK_EQ(tile_size_x % num_threads_x_, 0);
|
| D | kernel_tiling.h | 99 int64 tile_size_x,
|
| /external/tensorflow/tensorflow/compiler/xla/service/gpu/ |
| D | ir_emitter_unnested.cc | 2299 int64 tile_size_x, int64 num_threads_x, in GetStartOffsetAndStepForX() argument
2309 x, llvm::ConstantInt::get(index_ty, tile_size_x / num_threads_x)); in GetStartOffsetAndStepForX()
2323 int64 tile_size_x = mapping_scheme->GetTileSizeForDimensionX(); in EmitFullElementalTile() local
2329 tile_size_x, num_threads_x, mapping_scheme, builder, x, index_ty); in EmitFullElementalTile()
2341 for (int64 j = 0; j < tile_size_x / num_threads_x; j++) { in EmitFullElementalTile()
2363 int64 tile_size_x = mapping_scheme->GetTileSizeForDimensionX(); in EmitPartialElementalTile() local
2368 tile_size_x, num_threads_x, mapping_scheme, builder, x, index_ty); in EmitPartialElementalTile()
2372 for (int64 j = 0; j < tile_size_x / num_threads_x; j++) { in EmitPartialElementalTile()
2423 int64 tile_size_x = mapping_scheme->GetTileSizeForDimensionX(); in EmitTiledElementalCodeWithBoundsCheck() local
2430 builder->CreateICmpEQ(llvm::ConstantInt::get(index_ty, tile_size_x), in EmitTiledElementalCodeWithBoundsCheck()
[all …]
|