llama.cpp
CUDA: use registers instead of smem in topk-moe
#16647

Merged

CUDA: use registers instead of smem in topk-moe #16647

JohannesGaessler merged 1 commit into ggml-org:master from am17an:cuda_topk_optimize

CUDA: use registers instead of smem in topk-moe

06cd6bd4

github-actions added Nvidia GPU

github-actions added ggml

JohannesGaessler approved these changes on 2025-10-18

JohannesGaessler merged 38355c6c into master 276 days ago

am17an deleted the cuda_topk_optimize branch 276 days ago

Reviewers

JohannesGaessler

Assignees

No one assigned

Labels

Nvidia GPU ggml

Milestone

No milestone