-
Notifications
You must be signed in to change notification settings - Fork 43
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/2023a] OpenFOAM v2312 #563
base: 2023.06-software.eessi.io
Are you sure you want to change the base?
{2023.06}[foss/2023a] OpenFOAM v2312 #563
Conversation
boegel
commented
May 7, 2024
•
edited
Loading
edited
Instance
|
Instance
|
bot: build repo:eessi.io-2023.06-software arch:x86_64/amd/zen3 |
Updates by the bot instance
|
Updates by the bot instance
|
New job on instance
|
easystacks/software.eessi.io/2023.06/eessi-2023.06-eb-4.9.1-2023a.yml
Outdated
Show resolved
Hide resolved
Co-authored-by: ocaisa <alan.ocais@cecam.org>
…io_OpenFOAM-v2312
bot: build repo:eessi.io-2023.06-software arch:x86_64/amd/zen3 |
Updates by the bot instance
|
Updates by the bot instance
|
New job on instance
|
bot: build repo:eessi.io-2023.06-software arch:x86_64/amd/zen3 |
Updates by the bot instance
|
Updates by the bot instance
|
New job on instance
|
Sanity check is failing because of missing binaries (e.g.
But I can't find any explanation for this in the logs... |
Oh, actually, it looks like the build step did fail, even though EB says it completed. In the log file of the build step itself I see lots of errors near the bottom, most of them are like this:
|
Ah, GMP is not included in the list of dependencies but it clearly is one |
Yep, and MPFR. |
Doing an interactive build now with both of them added to the deps:
It's worrying though that the build actually failed, but apparently this |
If I remember correctly this was one of the motivations for implementing the sanity check concept in EasyBuild :) |
Adding those dependencies didn't help, still getting the same issue/errors. |
It's probably related to CGAL, which could previously be built in different ways. Since version 5 it's header-only, but when using this with OpenFOAM, it looks like MPFR and GMP are still required. There seems to be functionality in OpenFOAM that should detect which variant of CGAL you have, but I find lots of similar issues, e.g. this one with Spack: https://develop.openfoam.com/Development/openfoam/-/issues/3101. |
Confirmed, it's related to CGAL. In Spack they even stick to CGAL 4, probably because of similar issues: I suspect it pops up now because we're filtering $LD_LIBRARY_PATH. As a sort of a quick/dirty workaround I've added a command to |
…M-v2312 move OpenFOAM to EB 4.9.2 easystack
bot: build repo:eessi.io-2023.06-software arch:aarch64/generic |
Updates by the bot instance
|
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
|
Looks like the path for skylake is too long... |
HI, just to double check, once this is issue is fixed, will we be able to run OpenFoam on zen4? |
@marconetto Yes and no, the Zen4 builds are happening separately as we still have to bring that stack in line with the other architectures. Having said that, zen4 can't go through without getting this PR over the line. It's not an easy one though as this is a difficult issue and will require a little design |
thanks for the clarification @ocaisa! |
Based on my comments in #296 (comment), I was hoping to fix this by only unsetting some environment variables. That didn't work though, and I'm still not fully sure why. The failing command does work fine when running it interactively, also with the same environment variables set as during the build. My only guess is that there's a bunch of makefiles calling/including other ones, which are also somehow increasing the size of the environment and apparently exceeding some limit. While analyzing and debugging this, I came up with a workaround though. The failing command is compiling
|
@bedroge Awesome work! Instead of implementing this in a hook, why don't we bake it into the OpenFOAM easyblock instead, since this should be a general fix, it's not specific to EESSI at all? |
Yes, you're absolutely right. I can open a PR in the easyblocks repo. |
I've opened an easyblocks PR: easybuilders/easybuild-easyblocks#3388. |
bot: build repo:eessi.io-2023.06-software arch:x86_64/intel/skylake_avx512 |
Updates by the bot instance
|
Updates by the bot instance
|
New job on instance
|