search

NVIDIA / Fuser

A Fusion Code Generator for NVIDIA GPUs (commonly known as "nvFuser")
Other
271 stars 53 forks source link

Add TT, TN, NT, NN tests for HopperMultipleMatmulScheduler #3310

Closed rdspring1 closed 2 weeks ago

rdspring1 commented 3 weeks ago

This PR creates four tests for the HopperMultiMatmulScheduler. Each tests covers a different matmul layout - TT, TN, NT, and NN where the input arguments are already broadcasted.

rdspring1 commented 3 weeks ago

I created TN test with MNK ordering, added custom MatmulParams, and kept the original NT tests because two tests are better than one.

rdspring1 commented 2 weeks ago

!test