llama.cpp
IQ3_S improvements
#5829
Merged

IQ3_S improvements #5829

ggerganov merged 7 commits into master from ik/iq3_s_faster
ikawrakow
iq3_s: somewhat faster AVX2 dot product
39e3a429
iq3_s: somewhat faster ARM_NEON dot product
1e949891
iq3_s: another small ARM_NEON improvement
9c5b594c
iq3_s: minor improvement on Metal
7b629c3b
iq3_s: PPL improvement
11d4e099
iq3_s: use new grid everywhere
93bce3c9
Fix ARM_NEON
d4dfc250
Nindaleth
sorasoras
Artefact2
ikawrakow
Nindaleth
ikawrakow
ggerganov
ggerganov approved these changes on 2024-03-02
ggerganov ggerganov merged bbde6eb2 into master 2 years ago
JianbangZ
ikawrakow
JianbangZ
JianbangZ
Nexesenex

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone