llama.cpp
10+% performance improvement of ggml_vec_dot_q4_0 on AVX2
#654
Merged
Overview
Commits
3
Changes
View On GitHub

10+% performance improvement of ggml_vec_dot_q4_0 on AVX2 #654

sw merged 3 commits into ggml-org:master from SebastianApel:performance_improvement_q4_0_AVX2
SebastianApel
jart
rabidcopy
x02Sylvie
SebastianApel
rabidcopy
sw
SebastianApel
SebastianApel
sw
rabidcopy
SebastianApel
rabidcopy
howard0su
SebastianApel
SebastianApel SebastianApel changed the title ~1.5x performance improvement of ggml_vec_dot_q4_0 on AVX2 10+% performance improvement of ggml_vec_dot_q4_0 on AVX2 2 years ago
Ameobea
Ameobea
ggerganov
ggerganov commented on 2023-04-02
sw
sw commented on 2023-04-02
SebastianApel
SebastianApel
SebastianApel
SebastianApel SebastianApel force pushed from cb5cc711 to d8acf294 2 years ago
SebastianApel SebastianApel requested a review from sw sw 2 years ago
SebastianApel SebastianApel force pushed from d8acf294 to e621f62a 2 years ago
SebastianApel Performance improvement of AVX2 code
69ef03d5
SebastianApel SebastianApel force pushed from 9e62f03e to 69ef03d5 2 years ago
sw
sw commented on 2023-04-02
rabidcopy
SebastianApel Fixed problem with MSVC compiler
b589e34f
SebastianApel Reviewer comments: removed double semicolon, deleted empty line 1962
1ed8878a
SebastianApel SebastianApel requested a review from sw sw 2 years ago
sw
sw approved these changes on 2023-04-03
sw sw merged 437e7785 into master 2 years ago
rabidcopy
sw

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
None yet
Milestone
No milestone
Terms Of Use Privacy Policy Refund Policy Impressum