ci: add codes-ci-ubuntu24-mpich image (source-built MPICH)#260
Merged
caitlinross merged 1 commit intoJul 10, 2026
Conversation
Ubuntu 24.04's apt mpich (4.2.0-5build3) is built --with-device=ch4:ucx and is
broken on the GitHub runners: UCX aborts MPI_Init on the runner's verbs device
("ibv_create_srq() failed"), and even past that mpiexec launches size-1
singletons so optimistic (--sync=3) tests silently run sequential. There's no
fixed mpich in the noble archive, so build our own.
- ci/ubuntu24-mpich/Dockerfile: ubuntu:24.04 + MPICH from source with
--with-device=ch4:ofi (libfabric/TCP, no UCX/verbs), plus the toolchain,
cmake/ninja, flex/bison, clang, and lcov. Bumping MPICH_VERSION is the whole
update story.
- .github/workflows/ubuntu24-mpich-image.yml: publishes ghcr.io/codes-org/
codes-ci-ubuntu24-mpich (mirrors full-ci-image.yml). Make the GHCR package
public after the first publish so build.yml (incl. fork PRs) can pull it
anonymously.
build.yml's ubuntu-24.04 mpich legs are pointed at this image on a separate
branch. Validated locally: mpiexec forms a real 3-rank world and CODES's
optimistic ping-pong runs genuinely parallel (Total PEs 3, ~13% remote events,
~88k rollbacks) instead of degrading to sequential.
Member
Author
|
Note that this PR doesn't fix CI; it just creates the image (once merged) needed for #258 which will fix it. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Ubuntu 24.04's apt mpich (4.2.0-5build3) seems to have broken today and causing CI issues. It is now built --with-device=ch4:ucx and is broken on the GitHub runners: UCX aborts MPI_Init on the runner's verbs device ("ibv_create_srq() failed"), and even past that mpiexec launches size-1 singletons so optimistic (--sync=3) tests silently run sequential. So we'll just build a docker container that builds MPI from source instead of from apt.