mlir.dialects.transform.gpu

Classes

MapForallToBlocks	Specialization for MapForallToBlocks class.
MapNestedForallToThreads	Specialization for MapNestedForallToThreads class.

Module Contents

class mlir.dialects.transform.gpu.MapForallToBlocks(result_type: Type, target: Operation | OpView | Value, *, grid_dims: Sequence[int] | Attribute | None = None, generate_gpu_launch: bool | Attribute | None = None, loc=None, ip=None)

class mlir.dialects.transform.gpu.MapForallToBlocks(target: Operation | OpView | Value, *, grid_dims: Sequence[int] | Attribute | None = None, generate_gpu_launch: bool | Attribute | None = None, loc=None, ip=None)

Bases: MapForallToBlocks

Specialization for MapForallToBlocks class.

class mlir.dialects.transform.gpu.MapNestedForallToThreads(result_type: Type, target: Operation | OpView | Value, *, block_dims: Sequence[int] | None = None, warp_size: Sequence[int] | None = None, sync_after_distribute: bool | None = None, loc=None, ip=None)

class mlir.dialects.transform.gpu.MapNestedForallToThreads(target: Operation | OpView | Value, *, block_dims: Sequence[int] | None = None, warp_size: Sequence[int] | None = None, sync_after_distribute: bool | None = None, loc=None, ip=None)

Bases: MapNestedForallToThreads

Specialization for MapNestedForallToThreads class.