llama.cpp
Remove redundant CUDA copies after gated_delta_net.
#23940

Merged

Remove redundant CUDA copies after gated_delta_net. #23940

gaugarg-nv merged 2 commits into ggml-org:master from gaugarg-nv:fuse_gated_delta_net_with_copy

gaugarg-nv requested a review 135 days ago

github-actions added Nvidia GPU

github-actions added ggml

JohannesGaessler commented on 2026-06-01

ggerganov assigned

ggerganov 133 days ago

gaugarg-nv force pushed 124 days ago

ggerganov unassigned

ggerganov 112 days ago

gaugarg-nv force pushed 109 days ago

github-actions added CUDA

gaugarg-nv force pushed 109 days ago

ggerganov assigned

ggerganov 109 days ago

gaugarg-nv force pushed 108 days ago

am17an commented on 2026-07-02

Remove redundant CUDA copies after gated_delta_net.

5850b4ed

Address review comments

71daf7d4

gaugarg-nv force pushed to 71daf7d4 103 days ago

am17an approved these changes on 2026-07-03

pwilkin approved these changes on 2026-07-03

gaugarg-nv merged 5a460dea into master 102 days ago

gaugarg-nv deleted the fuse_gated_delta_net_with_copy branch 102 days ago

Reviewers

pwilkin

am17an

JohannesGaessler

Assignees

ggerganov

Labels

Nvidia GPU ggml CUDA

Milestone

No milestone