-
Notifications
You must be signed in to change notification settings - Fork 705
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
TensorFlow builds fail on JURECA and JUWELS #13967
Comments
@SebastianAchilles Anything useful in the logs for the failing |
This looks like there is a problem with the creation of the threads. I assume this is caused by our (too?) strict limitations on the login nodes. |
It seems like
caused the problem on the login nodes. A larger values would be needed for running the tests. @boegel you use 10k, right? With
c.f. #13960 (comment) @boegel Do you have
|
What was the error when that was not added?
Bazel sends its greetings... Will send an easyblock PR to fix this. |
Did it fail on Dill? We use Also, does SciPy-bundle fail in the same way? There numpy is a |
Yes, it failed on Dill. On the lightweight OS of the backend nodes we do not have
Yes, I got the same error
Thank you! |
@SebastianAchilles Can you test with easybuilders/easybuild-easyblocks#2575? |
I will start a test report after 6PM. Today we have maintenance day on all systems. |
Follow up on the failing test reports in #13877
Building TensorFlow 2.4.1, 2.5.0 and 2.6.0 fails on JURECA and JUWELS, but not on my minimalist rocky 8.4 container.
EB Config and system info
Test Reports:
I am trying to understand why the builds fail on JURECA and JUWELS. And whether this is caused by to the specific configuration of these systems. Maybe @Flamefire or @boegel can give me a hint to better understand the problem?
The text was updated successfully, but these errors were encountered: