Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

CUDA CI Update, main branch (2024.04.17.) #544

Closed

Conversation

krasznaa
Copy link
Member

This is an attempt at bringing the CUDA runtime tests back into a working order. I don't expect the first try to succeed (CUDA 11.6 seems just as broken for these tests as 11.5), but I have some further ideas as well...

@krasznaa krasznaa added the cicd Changes related to the CI system label Apr 17, 2024
@krasznaa
Copy link
Member Author

@paulgessinger, are draft PRs not executed on the GitLab CI? 😕 I was hoping that I could keep this "cleanly" in a draft stage until I sorted out exactly how to fix the existing issue. 🤔

@paulgessinger
Copy link
Member

@krasznaa No. This is intentional to manage the load on our CI VMs. With the current resources, I don't see a great way around this.

@beomki-yeo
Copy link
Contributor

Can we fix the CI soon? it is not nice to leave this unfixed for long time.

@krasznaa krasznaa force-pushed the CUDACIUpdate-main-20240417 branch from 66acd10 to fe105e0 Compare May 18, 2024 11:24
@krasznaa krasznaa marked this pull request as ready for review May 18, 2024 11:24
@krasznaa
Copy link
Member Author

I'll try to find a working setup finally...

@krasznaa krasznaa force-pushed the CUDACIUpdate-main-20240417 branch 3 times, most recently from 4229511 to 88dc8bc Compare May 18, 2024 14:24
It no longer uses Acts Docker images, but rather uses minimal
images. With the build and test jobs picking up most of the
dependencies of the project from CVMFS.
@krasznaa krasznaa force-pushed the CUDACIUpdate-main-20240417 branch from 88dc8bc to da8a61e Compare May 18, 2024 14:25
@krasznaa
Copy link
Member Author

😦 Unfortunately the GPU CI machine is not set up to use CVMFS at the moment.

@paulgessinger, we'd need 2 things to make things "perfect":

  • The Acts GPU CI node would need to mount CVMFS into the containers that it starts;
  • We'd need to update the node to the latest NVIDIA GPU driver.

Let's maybe discuss offline what's the best way to make this happen. 🤔

@krasznaa
Copy link
Member Author

Replaced by #616.

@krasznaa krasznaa closed this Jun 18, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
cicd Changes related to the CI system
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants