Repository navigation
{2023.06}[foss/2023b] Rebuild of PETSc to pick up MUMPS dependency - #1784
Conversation
|
There is no reason this doesn't build so using the shotgun bot: build repo:eessi.io-2023.06-software instance:eessi-bot-mc-aws for:arch=x86_64/amd/zen2 |
|
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
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
Some failures, let's hope they were flukes bot: build repo:eessi.io-2023.06-software instance:eessi-bot-mc-aws for:arch=x86_64/amd/zen2 |
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
From the logs in the haswell tarball it seems that MUMPS is now being correctly picked up, now tough we are getting failures... |
|
Also not 100% sure but we might need a rebuild of petsc4py as well. In my case i definetily needed it since the RPATH of it would still point to EESSI and not the EESSI-extend dir. Depends if it actually builds stuff differently based on the presence of mumps EDIT: In case it is needed i would do it in another PR as the bindings build should be much faster |
|
All the failures i see are related to permission to chmod/chown files after the installation. The one for |
|
I think you know my answer bot: build repo:eessi.io-2023.06-software instance:eessi-bot-mc-aws for:arch=x86_64/amd/zen2 |
|
New job on instance
|
|
I tried putting the haswell tarball on top of EESSI with overlayfs and it seems it might not be needed to rebuild also Also tested running the app that was crashing due to the missing MUMPS on top of it |
|
bot:status last_build |
|
This is the status of all the
|
2 similar comments
|
This is the status of all the
|
|
This is the status of all the
|
|
This is the status of all the
|
5 similar comments
|
This is the status of all the
|
|
This is the status of all the
|
|
This is the status of all the
|
|
This is the status of all the
|
|
This is the status of all the
|
|
Staging PR has been merged, needs some time for the ingestion but PR is good to merge |
|
Can confirm it works without rebuilding |
The current build of PETSc (at least for x86_64/intel/haswell) does not have MUMPS support even if it is added in the list of dependencies.
From the EESSI build logs
Triggering a rebuild with EESSI-extend seems to pick it up so lets try a rebuild and see if it fixes this problem