Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

patch avx512 dispatch for 8bit_direct for PR 3853 #3871

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

patch avx512 dispatch for 8bit_direct for PR 3853 #3871

wants to merge 2 commits into from

Commits on Sep 18, 2024

  1. Generated from a GitHub Pull Request. Run 'jf sync' on this diff to l… 

    …oad the correct commit data.

Differential Revision: D62989543
    @facebook-github-bot
    mengdilin authored and facebook-github-bot committed Sep 18, 2024
    Configuration menu
    Copy the full SHA
    d06b600 View commit details
    Browse the repository at this point in the history

Commits on Sep 19, 2024

  1. patch avx512 dispatch for 8bit_direct for PR 3853 

    Summary:
Context in https://www.internalfb.com/diff/D62989543?dst_version_fbid=830141322266715&transaction_fbid=927659569187374

8bit_direct no longer has a huge regression https://www.internalfb.com/phabricator/paste/view/P1596996696 because we will branch to avx2 if `d` (size of the input vectors) is not a multiple of 32 but a multiple of 16.

Differential Revision: D62990470
    @mengdilin @facebook-github-bot
    mengdilin authored and facebook-github-bot committed Sep 19, 2024
    Configuration menu
    Copy the full SHA
    a71b5d2 View commit details
    Browse the repository at this point in the history