llama.cpp
docker: parameterize CUDA version in build matrix, remove cuda-new.Dockerfile
#20920

Merged

docker: parameterize CUDA version in build matrix, remove cuda-new.Dockerfile #20920

CISC merged 1 commit into ggml-org:master from M1DNYT3:cuda-bump

M1DNYT3 requested a review from

ngxson 80 days ago

github-actions added devops

M1DNYT3 force pushed from 68fd6156 to 98c31743 71 days ago

docker: parameterize CUDA version in build matrix, remove cuda-new.Do…

5ef55acf

M1DNYT3 force pushed from 98c31743 to 5ef55acf 71 days ago

M1DNYT3 requested a review 71 days ago

M1DNYT3 changed the title ~~Bump CUDA base image 12.4.0 -> 12.9.1 (enables sm_120a-real for Blackwell)~~ docker: parameterize CUDA version in build matrix, remove cuda-new.Dockerfile 71 days ago

CISC approved these changes on 2026-04-02

CISC requested a review 71 days ago

taronaeo approved these changes on 2026-04-03

taronaeo added merge ready

CISC merged 277ff5ff into master 70 days ago

M1DNYT3 deleted the cuda-bump branch 66 days ago

Reviewers

taronaeo

CISC

ngxson

Assignees

No one assigned

Labels

devops merge ready

Milestone

No milestone

llama.cpp docker: parameterize CUDA version in build matrix, remove cuda-new.Dockerfile #20920 Merged

docker: parameterize CUDA version in build matrix, remove cuda-new.Dockerfile #20920

llama.cpp
docker: parameterize CUDA version in build matrix, remove cuda-new.Dockerfile
#20920

Merged