Skip to content

Commit 082aa9a

Browse files
jiawenliu64facebook-github-bot
authored andcommitted
Boost performance of MXFP4 quantization with inline PTX (#4694)
Summary: Pull Request resolved: #4694 X-link: facebookresearch/FBGEMM#1720 - Add inline PTX to boost MXFP4 quantization kernel performance - Fix MXFP4 scaling factor in grouped GEMM Differential Revision: D80182398
1 parent 08a4c45 commit 082aa9a

File tree

3 files changed

+161
-184
lines changed

3 files changed

+161
-184
lines changed

0 commit comments

Comments
 (0)