llama.cpp
hexagon: dma optimizations (mostly fixing regressions)
#21137

Merged

hexagon: dma optimizations (mostly fixing regressions) #21137

max-krasnyansky merged 3 commits into ggml-org:master from qualcomm:hexagon-dma-opts

hex-fa: add simple dma cache for Mask

341916c4

hex-dma: unset in-order desc bit which caused signficant perf regression

c9b13642

hex-rope: update comment to clarify that we don't need in-order DMA c…

17a19c4e

max-krasnyansky requested a review 5 days ago

github-actions added ggml

github-actions added Hexagon

lhez approved these changes on 2026-03-29

ggerganov approved these changes on 2026-03-29

max-krasnyansky merged f5d1c417 into master 5 days ago

Reviewers

ggerganov

lhez

Assignees

No one assigned

Labels

ggml Hexagon

Milestone

No milestone