]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
server : print graphs reused in slot timings (#23279)
authorGeorgi Gerganov <redacted>
Tue, 19 May 2026 06:46:58 +0000 (09:46 +0300)
committerGitHub <redacted>
Tue, 19 May 2026 06:46:58 +0000 (09:46 +0300)
commit3c81c8deeabba01fa40869325ea80d07eef75fc6
tree8095f9ccc299678daf2298eef2b031dd1a5bbdee
parentcd963fee6a86387d598ebe3888017376d6e9e8f6
server : print graphs reused in slot timings (#23279)

Add graphs reused counter to the per-slot timing output, printed via
llama_perf_context().

Assisted-by: llama.cpp:local pi
Co-authored-by: ggerganov <redacted>
tools/server/server-context.cpp