For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /get-started.md).

Mojo function

num_matrix_reg

def num_matrix_reg[dim_1: Int, dim_2: Int]() -> Int

Calculates the number of matrix registers required per thread.

Determines how many registers each thread in a warp needs to store a matrix of the given dimensions. This is calculated by dividing the total number of elements (dim_1 * dim_2) by the warp size, as the matrix is distributed across all threads in the warp.

Parameters:

• ​dim_1 (Int): First dimension of the matrix.
• ​dim_2 (Int): Second dimension of the matrix.

Returns:

Int: The number of matrix registers needed per thread.