]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 0d135df)
server: refactor batch construction (#24843)
authorXuan-Son Nguyen <redacted>
Sun, 21 Jun 2026 12:16:11 +0000 (14:16 +0200)
committerGitHub <redacted>
Sun, 21 Jun 2026 12:16:11 +0000 (14:16 +0200)
commitbddfd2b1137cd6e51fbb939081caf50e9f496a66
tree0c48f57519c98b38f48cabb73ba29fd02ba23429tree
parent0d135df48ccee9a799fa9d9ea0ed494bd4fdd74fcommit | diff
server: refactor batch construction (#24843)

* server: refactor batch construction

* wip

* wip 2

* wip 3

* wip 4

* add abort_all_slots

* handle batch full more carefully

* fix assert

* rm debug log

* small nits

* (debug) add timings

* debug: force llama_synchronize for accurate timings

* address comments

* disable DEBUG_TIMINGS
tools/server/server-context.cpp diff | blob | history
Packaging of ggml-org/llama.cpp