]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
metal : use FA-vec kernel up to batch size 20 (#13496)
authorGeorgi Gerganov <redacted>
Tue, 13 May 2025 15:04:39 +0000 (18:04 +0300)
committerGitHub <redacted>
Tue, 13 May 2025 15:04:39 +0000 (18:04 +0300)
commitf0995d28ce3d15095b6845d94ce4465e46575873
tree5f3c3cc58806abaea8aa464297e4d3efc1f8b731
parentc252e0c4097b34666e5a81db9d0450d71fa3098f
metal : use FA-vec kernel up to batch size 20 (#13496)

* batched-bench : fix pp batch contents

* metal : optimize multi-sequence FA vec kernel

ggml-ci

* metal : use FA-vec kernel up to batch size 20

ggml-ci
ggml/src/ggml-metal/ggml-metal.m