PR #30517 CUDA: faster MMVQ for low-bit types

CUDA: faster MMVQ for low-bit types #30517

thelittlefireman merged 2 commits into master from aman/vecdot-opt

CUDA: faster MMVQ for low-bit types

c869d1e3

am17an marked this pull request as ready for review 14 days ago

am17an requested a review 14 days ago

github-actions added ggml

github-actions added CUDA

pwilkin approved these changes on 2026-10-14

ggml-cuda: hip: small nwarps GCN fix on Q3_K(ncols=2,3), Q6_K(ncols=4)

6e9b69aa

IMbackK approved these changes on 2026-10-17

thelittlefireman merged 2744e48d into master 11 days ago

JohannesGaessler commented on 2026-10-17

Reviewers

IMbackK

pwilkin

JohannesGaessler

Assignees

No one assigned

Labels

ggml CUDA

Milestone

No milestone