PR #16857 CUDA: add expert reduce kernel

CUDA: add expert reduce kernel #16857

am17an merged 3 commits into ggml-org:master from am17an:expert-reduce

am17an requested a review from

slaren 103 days ago

am17an commented on 2025-10-30

am17an requested a review from

JohannesGaessler 103 days ago

CUDA: add expert reduce kernel

4999b215

am17an force pushed to 4999b215 103 days ago

github-actions added testing

github-actions added Nvidia GPU

github-actions added ggml

JohannesGaessler commented on 2025-10-30

contigous checks, better formatting, use std::vector instead of array

e765d9ad

am17an requested a review from

JohannesGaessler 102 days ago

JohannesGaessler approved these changes on 2025-10-31

use vector empty instead of size

2c10f1c4

am17an merged 4146d6a1 into master 102 days ago

am17an deleted the expert-reduce branch 102 days ago

Reviewers

JohannesGaessler

slaren

Assignees

No one assigned

Labels

testing Nvidia GPU ggml

Milestone

No milestone