llama.cpp
[Performance] [metal] : Up to 18x speedup on CONV_2D
#28069
Merged

[Performance] [metal] : Up to 18x speedup on CONV_2D #28069

Sou-ly
Sou-ly metal : add implicit GEMM kernel for CONV_2D
60c12ca5
Sou-ly Sou-ly requested a review 79 days ago
github-actions github-actions added ggml
github-actions github-actions added Apple Metal
ggml-gh-bot
ggml-gh-bot ggml-gh-bot added draft
github-actions github-actions marked this pull request as draft 79 days ago
github-actions github-actions removed draft
Sou-ly Sou-ly changed the title metal : add implicit GEMM kernel for CONV_2D [Performance] [metal] : Up to 18x speedup on CONV_2D 79 days ago
Sou-ly Sou-ly marked this pull request as ready for review 79 days ago
ServeurpersoCom ServeurpersoCom requested a review from ServeurpersoCom ServeurpersoCom 1 day ago
ServeurpersoCom
ServeurpersoCom approved these changes on 2026-11-17
ServeurpersoCom ServeurpersoCom requested a review from ggerganov ggerganov 1 day ago
ServeurpersoCom ServeurpersoCom requested a review from ngxson ngxson 1 day ago
ggerganov
ggerganov approved these changes on 2026-11-17
ServeurpersoCom ServeurpersoCom merged e0a9118f into master 1 day ago
ServeurpersoCom

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone