llama.cpp
ggml: optimize concat op by replacing per-element memcpy with row-level memcpy
#24575

Merged

ggml: optimize concat op by replacing per-element memcpy with row-level memcpy #24575

ggerganov merged 7 commits into ggml-org:master from sirohikartik:optim/concat-forward

ggml: optimize concat op by replacing per-element memcpy with row-lev…

610531d9

sirohikartik requested a review from

ggerganov 76 days ago

github-actions added ggml

sirohikartik force pushed 9 days ago

CISC commented on 2026-08-19

sirohikartik force pushed 8 days ago

ggml: fix concat offsets for row-level copies

948c136e

sirohikartik force pushed to 948c136e 8 days ago

CISC approved these changes on 2026-08-20

ggerganov commented on 2026-08-21

ggml: add concat row contiguity asserts

b65a114c

sirohikartik force pushed to b65a114c 7 days ago

Merge branch 'master' into optim/concat-forward

7dc39268

ggml: move concat block size asserts

d557c6c0

ggml: remove redundant concat asserts

24added9

CISC commented on 2026-08-21

Update ggml/src/ggml-cpu/ops.cpp

f6c2a860

CISC requested a review from

ggerganov 6 days ago

ggerganov approved these changes on 2026-08-22

ggerganov merged 369e1cd6 into master 6 days ago

Reviewers

ggerganov

CISC

Assignees

No one assigned

Labels

ggml

Milestone

No milestone