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

Adds Optional AVX2 Support, Cache Alignment, and Enhances Model Export Speed #94

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Adds Optional AVX2 Support, Cache Alignment, and Enhances Model Export Speed #94

wants to merge 1 commit into from

Commits on Jul 26, 2023

  1. Adds optional AVX2 support, performance enhancements, and cache align… 

    …ment in Config

This commit includes:
1. Optional AVX2 support for matmul and rmsnorm functions.
2. Fused matrix multiplies with new matmul2 and matmul3 functions.
3. Cache aligned allocations for better performance and compatibility with SIMD/Vector intrinsics.
4. Updated Config struct to support cache alignment. NOTE: Previous models should be re-exported due to this change.
5. Enhanced performance of serialization in Llama export code.
6. Updated Makefile to support AVX2 and OMP AVX2 builds.
    @Foundation42
    Foundation42 committed Jul 26, 2023
    Configuration menu
    Copy the full SHA
    28ca877 View commit details
    Browse the repository at this point in the history