llama.cpp
cuda: guard the iq4_nl dequantize row kernel against short rows
#29683

Merged

cuda: guard the iq4_nl dequantize row kernel against short rows #29683

JohannesGaessler merged 1 commit into ggml-org:master from yeahdongcn:xd/cuda-iq4-nl-short-rows

cuda: guard the iq4_nl dequantize row kernel against short rows

908d76a7

github-actions added ggml

github-actions added CUDA

yeahdongcn marked this pull request as ready for review 10 days ago

yeahdongcn requested a review 10 days ago

yeahdongcn requested a review from

CISC 10 days ago

yeahdongcn requested a review from

slaren 10 days ago

yeahdongcn requested a review from

JohannesGaessler 10 days ago

CISC removed review request from

slaren 10 days ago

am17an approved these changes on 2026-09-30

JohannesGaessler approved these changes on 2026-09-30

JohannesGaessler merged f872b591 into master 9 days ago

Reviewers

JohannesGaessler

am17an

CISC

Assignees

No one assigned

Labels

ggml CUDA

Milestone

No milestone