llama.cpp
~7% faster Q5_1 AVX2 code
#1477
Merged

~7% faster Q5_1 AVX2 code #1477

sw merged 4 commits into ggml-org:master from patch-1
ilyakurdyukov
ilyakurdyukov ~7% faster Q5_1 AVX2 code
2d61ed7f
ilyakurdyukov fixed AVX code
c68faa99
sw
ilyakurdyukov also applied to ggml_vec_dot_q4_1_q8_1
c1b3a8e2
ilyakurdyukov
ilyakurdyukov make use of upcoming AVXVNNIINT8
ff263841
ilyakurdyukov
sw
sw approved these changes on 2023-05-16
sw sw merged 42627421 into master 2 years ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone