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

Faster AVX2 matrix multiplications for lgacy quants #405

Merged
merged 4 commits into from
May 10, 2024
Merged

Faster AVX2 matrix multiplications for lgacy quants #405

merged 4 commits into from
May 10, 2024

Commits on May 10, 2024

  1. Matrix multiplications for legacy qunats

    @Kawrakow
    Kawrakow committed May 10, 2024
    Configuration menu
    Copy the full SHA
    8f7394f View commit details
    Browse the repository at this point in the history

  2. Very slightly faster Q5 dequantization 

    Somehow memcpy is kind of slow, so for
getting 4 bytes from 2-byte-aligned data
it is faster to just do or on two
consecutive 16-bit entries.
    @Kawrakow
    Kawrakow committed May 10, 2024
    Configuration menu
    Copy the full SHA
    897be80 View commit details
    Browse the repository at this point in the history

  3. Make it work after rebase 

    However, the way it is currently, we have lost the
zen4-tuned version.
    @Kawrakow
    Kawrakow committed May 10, 2024
    Configuration menu
    Copy the full SHA
    a3bca82 View commit details
    Browse the repository at this point in the history

  4. Restore faster AVX512VNNI+AVX512VL performance

    @Kawrakow
    Kawrakow committed May 10, 2024
    Configuration menu
    Copy the full SHA
    610a3e9 View commit details
    Browse the repository at this point in the history