llama.cpp
vulkan: disable FA mask_opt on GCN to improve performance
#24362

Merged

vulkan: disable FA mask_opt on GCN to improve performance #24362

0cc4m merged 2 commits into master from 0cc4m/vulkan-fa-mask-opt-gcn

0cc4m requested a review 110 days ago

github-actions added Vulkan

github-actions added ggml

vulkan: disable FA mask_opt on GCN to improve performance

d7382603

reenable mask opt over attention head size 256

78946e32

0cc4m force pushed from 6c2cbc4e to 78946e32 89 days ago

jeffbolznv approved these changes on 2026-07-08

ruixiang63 approved these changes on 2026-07-08

0cc4m merged a646006f into master 81 days ago

0cc4m deleted the 0cc4m/vulkan-fa-mask-opt-gcn branch 81 days ago

Reviewers

ruixiang63

jeffbolznv

Assignees

No one assigned

Labels

Vulkan ggml

Milestone

No milestone