News for hackers. Stuff that matters.

A comment on MatX: Efficient C++17 GPU numerical computing library with Python-like syntax. Parent comment

Re:MatX: Efficient C++17 GPU numerical computing library wit... by cburdick13 on Tuesday October 03, 2023 @09:25PM (#37759799)
We typically support whatever the underlying library supports. For int8 the support would come from cuBLASLt currently. I don't believe that or Cutlass supports mixed precision inputs, but I can check.
[ Parent | on HN ]

There are no comments on this comment yet.