-
Notifications
You must be signed in to change notification settings - Fork 49
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
[2023.06]{foss/2021b} SciPy-bundle v2021.10 #306
Conversation
Instance
|
bot: build arch:aarch64/neoverse_v1 repo:eessi-2023.06-software |
Updates by the bot instance
|
New job on instance
|
@laraPPr do you have access to the AWS cluster? It would probably be faster to try and do this build yourself, rather than let the bot do it. It'll allow you to go through 'regular' EasyBuild debugging procedures (e.g. get the build environment with I have seen The real solution here is to address it with the OpenBLAS devs and get to the bottom of it. If it is expected / 'ok' that these tolerances are exceeded. If so, tests that are written on top will also have to be adjusted to account for these increased tolerances. In any case, you could address this in an issue to the |
I do not think I have enough access to the AWS cluster to build it myself. I've also never done 'regular' EasyBuild debugging procedures but I am willing to try. |
Yes there was a hook set for the installation of OpenBLAS on aarch64 that relaxed the allowed number of failing numerical LAPACK tests. |
In EESSI pilot 2021.12, we installed We have run into issues like this before in EasyBuild, see easybuilders/easybuild-easyconfigs#11959 I think we should try and asses how "serious" these test failures are, but most likely we'll need to move forward with ignoring the failing tests, at least for now, since fixing may require significant effort... |
bot: build repo:eessi-2023.06-software |
Updates by the bot instance
|
New job on instance
|
New job on instance
|
New job on instance
|
New job on instance
|
New job on instance
|
New job on instance
|
New job on instance
|
New job on instance
|
@laraPPr To ignore the failing |
bot: build arch:aarch64/neoverse_v1 repo:eessi-2023.06-software |
Updates by the bot instance
|
New job on instance
|
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.
@laraPPr You will also need to define pre_test_hook
that picks up on PRE_TEST_HOOKS
, similar to the pre_configure_hook
that's already there.
bot: build arch:aarch64/neoverse_v1 repo:eessi-2023.06-software |
Updates by the bot instance
|
New job on instance
|
Fix indentation line 267
bot: build arch:aarch64/neoverse_v1 repo:eessi-2023.06-software |
Updates by the bot instance
|
New job on instance
|
@laraPPr Hook looks good to me now, let's make sure builds are done for all CPU targets so we can deploy this? We should also open an internal issue with details on the failing tests, and start looking into reporting the problems upstream. Let's do at least the internal issue before merging this PR (to avoid forgetting about it). |
Should I than also look at other tests that might have failed and were skipped? |
The issue should cover all failing tests in this SciPy-bundle, yes, so perhaps also You should be able to spot if there's additional failing test suite easily by looking for " |
All staging PRs merged, so deployment is being done right now, should be fully in place in about an hour or so. |
@laraPPr Sorry, my mistake, the way to check for failing test steps is to check for lines that start with "``", since that's the message emitted by the
|
created issue for the numy test failures with SciPy-bundle: |
split Lmod hooks merge ok despite some stupidly failing CI workflow
No description provided.