llama.cpp
ggml-cuda: use universal launch bounds for MoE MMVQ kernel
#24547

Open

ggml-cuda: use universal launch bounds for MoE MMVQ kernel #24547

batot1 wants to merge 1 commit into ggml-org:master from batot1:issue24064-universal-moe-launchbounds

ggml-cuda: use universal launch bounds for MoE MMVQ kernel

8b153af4

batot1 requested a review 2 days ago

github-actions added Nvidia GPU

github-actions added ggml

Reviewers

No reviews

Assignees

No one assigned

Labels

Nvidia GPU ggml

Milestone

No milestone