llama.cpp
ggml-cpu: FA add GEMM microkernel
#19422
Merged

ggml-cpu: FA add GEMM microkernel #19422

am17an merged 10 commits into ggml-org:master from am17an:opt-fa-micro-gemm
am17an
am17an am17an requested a review from ggerganov ggerganov 105 days ago
github-actions github-actions added ggml
am17an ggml-cpu: FA add GEMM microkernel
c8b9c839
am17an add guard for sizeless vector types
a1e1420b
am17an fix case where DV % GGML_F32_EPR !=0
734f76fb
am17an am17an force pushed from c4c451a7 to 734f76fb 100 days ago
ggerganov
ggerganov commented on 2026-02-12
am17an move memset out of the loop
8debab31
ggerganov
am17an
ggerganov
am17an
ggerganov
ggerganov commented on 2026-02-13
ggerganov
ggerganov approved these changes on 2026-02-13
am17an move another memset out of the loop
9c660dda
ggerganov
ggerganov commented on 2026-02-13
am17an use RM=4 for arm
8d1be6c4
am17an simd_gemm: convert everything to int
d473b671
am17an am17an force pushed from 1b44835c to d473b671 99 days ago
am17an convert everything to size_t to avoid warnings
c34b1a48
am17an am17an force pushed from 27aa9286 to c34b1a48 99 days ago
am17an fixup
6aababd0
am17an add pragma for ignoring aggressive loop optimizations
9de8ba9c
am17an am17an merged 684b3610 into master 97 days ago
am17an am17an deleted the opt-fa-micro-gemm branch 97 days ago
ggerganov
ggerganov commented on 2026-02-15
Djip007
Djip007 commented on 2026-02-15
Djip007
Djip007 commented on 2026-02-15
Djip007
Djip007 commented on 2026-02-15
Djip007
Djip007 commented on 2026-02-15
Djip007
ggerganov
Djip007

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone