PR #25952 cuda: fuse MoE weighted expert reduction

cuda: fuse MoE weighted expert reduction #25952

ORippler merged 3 commits into ggml-org:master from anujj:cuda-moe-weighted-reduction-upstream

anujj requested a review from

ggerganov 95 days ago

anujj requested a review 95 days ago

github-actions added testing

github-actions added ggml

github-actions added CUDA

ORippler requested changes on 2026-07-21

am17an commented on 2026-07-23

ggerganov assigned

ggerganov 79 days ago

am17an commented on 2026-08-12

anujj force pushed 55 days ago

cuda : fuse MoE weighted reduction (mul + view + add)

eb24eb8a

anujj force pushed to eb24eb8a 54 days ago

Pruned the test matrix from 15 to 6

72743215

ORippler commented on 2026-09-01

am17an requested changes on 2026-09-01

Addressed the aman and olivers review comments

7a622743

am17an approved these changes on 2026-09-01

ORippler approved these changes on 2026-09-01

ORippler merged 3466812d into master 53 days ago

Reviewers

ORippler

am17an

ggerganov

Assignees

ggerganov

Labels

testing ggml CUDA

Milestone

No milestone