Manish Gupta 757275f279 Adding more Threadblock Tiles for Mixed-input TensorOp (BF16 * S8) in cutlass_library (#1132) ...

* Adding more tiles in the cutlass_library for mixed-input support.

* fix rebase issue

* more tiles to upcast a

2023-10-13 11:33:15 -04:00

220 KiB

Raw Blame History

View Raw