For the complete Mojo documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /docs/manual/basics.md).
copy_sram_to_local
def copy_sram_to_local[thread_layout: Layout[thread_layout.shape_types, thread_layout.stride_types], *, thread_scope: ThreadScope = ThreadScope.BLOCK, element_size: Int](dst: TileTensor[Storage=dst.Storage, address_space=AddressSpace.LOCAL, linear_idx_type=dst.linear_idx_type, element_size=element_size], src: TileTensor[Storage=src.Storage, address_space=AddressSpace.SHARED, linear_idx_type=src.linear_idx_type, element_size=element_size])
Synchronously copies a tile from SRAM (shared memory) to registers.
Delegates to SharedToLocalTileCopier. The axis-based distribution of
the legacy free function is not supported here.
Parameters:
- thread_layout (
Layout[thread_layout.shape_types, thread_layout.stride_types]): Warp layout describing how threads are organized over the copy. - thread_scope (
ThreadScope): Scope at which thread operations are performed. - element_size (
Int): Number of scalar elements per logical element; inferred from the source and destination tiles.
Args:
- dst (
TileTensor[Storage=dst.Storage, address_space=AddressSpace.LOCAL, linear_idx_type=dst.linear_idx_type, element_size=element_size]): Destination tile in local memory. - src (
TileTensor[Storage=src.Storage, address_space=AddressSpace.SHARED, linear_idx_type=src.linear_idx_type, element_size=element_size]): Source tile in shared memory.