Mojo function

tt_load_b_tr

tt_load_b_tr[mma_shape: IndexList[3]](tile: TileTensor[tile.dtype, tile.LayoutType, tile.origin, address_space=AddressSpace.SHARED, linear_idx_type=tile.linear_idx_type, element_size=tile.element_size]) -> SIMD[tile.dtype, 8]

Transposed B operand load for double-rate MFMA shapes.

Splits the tile along the K dimension into two halves (using tile[] instead of LayoutTensor's split[2]) and concatenates the results.

Returns:

SIMD

View source

Was this page helpful?

Thank you! We'll create more content like this.

Thank you for helping us improve!