For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /max/get-started.md).
Mojo struct
OutputTileStorage
struct OutputTileStorage[c_type: DType, c_dim0: Int, c_dim1: Int, num_output_stages: Int]
Storage for output tiles (C matrix).
Single source of truth for output tile array and storage. Separate from input tiles since output has different stage count.
All tiles use TileTensor natively. LayoutTensor conversion is kept for SMemEpilogueWriter compatibility.
Parameters
- c_type (
DType): Data type for C matrix tiles. - c_dim0 (
Int): First dimension for C tiles (OutputM). - c_dim1 (
Int): Second dimension for C tiles (OutputN). - num_output_stages (
Int): Number of output pipeline stages.
Fields
- c_tiles_storage (
OutputTileStorage[c_type, c_dim0, c_dim1, num_output_stages].CTileArray.Storage):
Implemented traits
AnyType,
ImplicitlyDestructible
comptime members
c_tile_layout
comptime c_tile_layout = Layout.row_major(c_dim0, c_dim1)
CTileArray
comptime CTileArray = SMemTileArray2DRowMajor[c_type, c_dim0, c_dim1, num_output_stages]
Methods
c_tiles
c_tiles(ref[AddressSpace._value] self) -> OutputTileStorage[c_type, c_dim0, c_dim1, num_output_stages].CTileArray
Get C tile array accessor.
Returns:
OutputTileStorage[c_type, c_dim0, c_dim1, num_output_stages].CTileArray
Was this page helpful?
Thank you! We'll create more content like this.
Thank you for helping us improve!