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 function
calculate_tile_n_k
def calculate_tile_n_k[a_type: DType, b_type: DType, c_type: DType, kernel_cols: Int](n: Int, k: Int) -> IndexList[2]
Helper heuristic function to decide on tile size to partition the matmul given the cache size and desired data layout.
Parameters:
- a_type (
DType): The dtype of the A tensor. - b_type (
DType): The dtype of the B tensor. - c_type (
DType): The dtype of the C tensor. - kernel_cols (
Int): The umber of columns of the micro kernel.
Returns:
IndexList[2]: The calculated tile size to partition the matmul as (TileN, TileK).
def calculate_tile_n_k[a_type: DType, b_type: DType, c_type: DType, kernel_cols: Int](global_tile_shape: GemmShape) -> IndexList[2]
Returns:
Was this page helpful?
Thank you! We'll create more content like this.
Thank you for helping us improve!