FBGEMM icon indicating copy to clipboard operation
FBGEMM copied to clipboard
verified publisher icon pytorch

Simplify grouped gemm output allocations

Open jwfromm opened this issue 6 months ago • 2 comments

Summary: This very minor diff changes how output tensors in cutlass grouped gemm are allocated. Rather than treat them as a flat array, they now are allocated as standard matrices. This may help avoid an integer overflow error in the torch caching allocator.

Reviewed By: jiawenliu64

Differential Revision: D74852208

jwfromm avatar May 16 '25 00:05 jwfromm

Deploy Preview for pytorch-fbgemm-docs ready!

Name Link
Latest commit aa04e38e1c68501b5dd156b0c898246cb929b1e1
Latest deploy log https://app.netlify.com/projects/pytorch-fbgemm-docs/deploys/68268853fa1bf5000876cb98
Deploy Preview https://deploy-preview-4134--pytorch-fbgemm-docs.netlify.app
Preview on mobile
Toggle QR Code...

QR Code

Use your smartphone camera to open QR code link.

To edit notification comments on pull requests, go to your Netlify project configuration.

netlify[bot] avatar May 16 '25 00:05 netlify[bot]

This pull request was exported from Phabricator. Differential Revision: D74852208

facebook-github-bot avatar May 16 '25 00:05 facebook-github-bot