]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: bea0452)
CUDA: Remove unneded bias/gate dims in fused mmvq (#16858)
authorOliver Simons <redacted>
Sat, 1 Nov 2025 05:13:26 +0000 (06:13 +0100)
committerGitHub <redacted>
Sat, 1 Nov 2025 05:13:26 +0000 (13:13 +0800)
commitd3dc9dd898be805c23a408cc36daed5b3bf29221
treeb4fc37de48ac81d7c21b88bfee3afb776f951a69tree
parentbea04522ff1a0d8559ccfd353aa018dcfbb608cccommit | diff
CUDA: Remove unneded bias/gate dims in fused mmvq (#16858)

* CUDA: Remove unneded bias/gate dims in fused mmvq

Pointed out
[here](https://github.com/ggml-org/llama.cpp/pull/16847#discussion_r2476798989)
that only a single value is needed per target col per thread

* Apply suggestions from code review

Co-authored-by: Johannes Gäßler <redacted>
* Fix "Error 991-D: extra braces are nonstandard" during compilation

---------

Co-authored-by: Johannes Gäßler <redacted>
ggml/src/ggml-cuda/mmvq.cu diff | blob | history
Packaging of ggml-org/llama.cpp