diff options
author | Kawrakow <iwankawrakow@gmail.com> | 2025-02-15 08:45:45 +0200 |
---|---|---|
committer | GitHub <noreply@github.com> | 2025-02-15 08:45:45 +0200 |
commit | 0551e7630b2e74dfd140aadbf6a136c58199f474 (patch) | |
tree | a96f395a2daac48d31cb1eedd1621b4943f61cac /examples/passkey/passkey.cpp | |
parent | 8e94b29e35d69a6721c35767222d16dc65df2db6 (diff) |
Moving 4D gemm logic from ggml.c to iqk_mul_mat.cpp (#207)
This allows us to optimize TG performance for GQA models.
E.g., for IQ4_XS L3-8B with 8k TG-64 goes from 8.6 to 10.26 t/s.
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Diffstat (limited to 'examples/passkey/passkey.cpp')
0 files changed, 0 insertions, 0 deletions