-
Notifications
You must be signed in to change notification settings - Fork 1.3k
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
[CI] Add a GPU runner. #16580
[CI] Add a GPU runner. #16580
Conversation
hageboeck
commented
Oct 2, 2024
•
edited
Loading
edited
- Add a new job to run with cuda
- Select GPU-enabled runners
- Pass host GPU into container
Test Results 17 files 17 suites 4d 1h 29m 6s ⏱️ For more details on these failures, see this check. Results for commit 2694d5d. ♻️ This comment has been updated with latest results. |
88919dc
to
fba5543
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It's green! 🎉
abf35da
to
1ae17f0
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for this improvement of the CI. For me the PR is ready to be merged, once we verify once again it works fine.
Maybe we want to consider restricting the testing only to the tests that use the GPU (a set we'll be expanding in the next few months, now that we have the build), in order to optimise the usage of the device w/o spending 30 minutes running the full test suite on CPU.
6db9141
to
88696d4
Compare
- Add a new run with cuda=On. - Create a build config file for ubuntu2404-cuda. - Map nvidia GPUs into the container when running with cuda.
RooBatchCompute is using host pointers in a kernel, which is hard to detect. This assertion helps to catch the error a lot earlier.
The choice should be made based on the input size, and not based on the output size.
When webgui=Off, a simple notebook test times out.