llama.cpp
hexagon: improve ARGSORT performance for small tensors
#25512

Merged

hexagon: improve ARGSORT performance for small tensors #25512

max-krasnyansky merged 4 commits into ggml-org:master from qualcomm:hexagon-argsort

hex-sort: add efficient bitomic sort in hvx regs up to 1024 elements

ac014740

hex-sort: fix inverted vrors

a2cc0539

hex-sort: specialize sort functions for the common cases

3bd734a3

github-actions added ggml

github-actions added Hexagon

hex-sort: add tracing and local context

7a5de3e7

max-krasnyansky marked this pull request as ready for review 5 days ago

max-krasnyansky requested a review 5 days ago

lhez approved these changes on 2026-07-10

ruixiang63 approved these changes on 2026-07-10

max-krasnyansky merged 67776eae into master 4 days ago

Reviewers

ruixiang63

lhez

Assignees

No one assigned

Labels

ggml Hexagon

Milestone

No milestone