Update packages: mpi, petsc, libmesh, code-server; fix openblas; download umpire by lindsayad · Pull Request #32404 · idaholab/moose

lindsayad · 2026-02-27T21:24:58Z

libMesh update libMesh/libmesh@`f8a1758...c889858`

Added FEType::p_refinement - this will allow MOOSE to independently enable or disable p refinement of variables even if they (otherwise) share the same finite element family and base polynomial order.
Added checks of hull integrity to Netgen interface, used by XYZDelaunay mesh generation in MOOSE
Fixed System::get_info() on uninitialized System objects
Overrode NodeElem::volume(), to return 1 more quickly and to explain that counter-intuitive behavior
Fixed crash when cleaning up from error exceptions while using umpire via PETSc
Added more test coverage for various MeshFunction and Variable features

Conda packages

`moose-mpi 2026.03.10`

Update compilers
Update openmpi to 5.0.10

`moose-petsc 3.24.5_0`

Update PETSc v3.24.4...v3.24.5
Do not download fblaslapack
Download openblas
Download umpire at v2025.12.0

`moose-libmesh 2026.03.03_c889858_0`

Update libmesh libMesh/libmesh@f8a1758...c889858

`moose-wasp 2025.09.19_02960f1_2`

Update compilers to be consistent with moose-mpi

`moose-seacas 2025.10.14_2`

Update compilers to be consistent with moose-mpi

`moose-dev 2026.03.10`

Bump due to dependency updates (moose-mpi, moose-petsc, moose-libmesh, moose-wasp)

Apptainer containers

`moose-mpi:2026.03.10`

Update rocky base images
Update openmpi to 5.0.10
Remove extraneous gcc 13 install in clang 20 container

`moose-petsc:3.24.5_0`

Update PETSc v3.24.4...v3.24.5
Do not download fblaslapack
Download openblas
Download umpire at v2025.12.0

`moose-libmesh:2026.03.03_c889858_0`

Update libmesh libMesh/libmesh@f8a1758...c889858

`moose-dev:2026.03.10`

Bump due to dependency updates (moose-petsc, moose-libmesh)
Fix USE_CUDA flag when building pytorch
Do not install openblas-devel
Add PETSc to CMAKE_INCLUDE_PATH and CMAKE_LIBRARY_PATH to use openblas from PETSc
Update code-server to 4.109.5

moosebuild · 2026-02-28T03:08:38Z

Job Documentation, step Docs: sync website on 77e0690 wanted to post the following:

View the site here

This comment will be updated on new commits.

moosebuild · 2026-02-28T06:23:33Z

Job Coverage, step Generate coverage on b197d08 wanted to post the following:

Framework coverage

	c89dc5	#32404 b197d0
	Total	Total	+/-	New
Rate	85.84%	85.84%	-0.00%	-
Hits	128436	128435	-1	0
Misses	21191	21192	+1	0

Diff coverage report

Full coverage report

Modules coverage

Coverage did not change

Full coverage reports

Reports

This comment will be updated on new commits.

.gitignore

- Use openblas instead of fblaslapack. I believe it's more optimized based on discussions I've had with Pierre - Download umpire which hypre suggests we should use - Make sure to download umpire version which supports modern cmake - Tell libtorch to use our PETSc downloaded openblas refs idaholab#32404

refs idaholab#32404

Not ready

libmesh

roystgnr

I'm happy with the libMesh parts of this.

libmesh

loganharbour · 2026-03-10T19:19:53Z

@cticenhour this is ready for a look when you have a moment (newsletter, conda)

cticenhour

Some weird behavior with moose-wasp build numbers observed. One requested minor change in the newsletter related to tech edit.

conda/wasp/meta.yaml

scripts/versioner.yaml

cticenhour · 2026-03-10T22:40:25Z

modules/doc/content/newsletter/2026/2026_03.md

+
 ## PETSc-level Changes

+The PETSc submodule has been updated from version 3.24.4 to 3.24.5. Find the GitLab generated set of


Suggested change

The PETSc submodule has been updated from version 3.24.4 to 3.24.5. Find the GitLab generated set of

The PETSc submodule has been updated from version 3.24.4 to 3.24.5. Find the GitLab-generated set of

This was copy pasta from your changes in the last newsletter ;)

I obviously changed my mind! :D

modules/doc/content/newsletter/2026/2026_03.md

refs idaholab#32404

libmesh

moosebuild · 2026-03-11T00:26:42Z

Job Precheck, step Versioner verify on 77e0690 wanted to post the following:

Versioner templates

Found 14 templates, 0 failed

Versioner influential files

Found 57 influential files, 20 changed, 0 added, 0 removed

package	status	file
mpi	CHANGE	apptainer/mpi.def
mpi	CHANGE	conda/mpi/conda_build_config.yaml
mpi	CHANGE	conda/mpi/meta.yaml
wasp	CHANGE	conda/wasp/conda_build_config.yaml
wasp	CHANGE	conda/wasp/meta.yaml
seacas	CHANGE	conda/seacas/conda_build_config.yaml
seacas	CHANGE	conda/seacas/meta.yaml
libmesh-vtk	CHANGE	conda/libmesh-vtk/conda_build_config.yaml
libmesh-vtk	CHANGE	conda/libmesh-vtk/meta.yaml
petsc	CHANGE	conda/petsc/conda_build_config.yaml
petsc	CHANGE	conda/petsc/meta.yaml
petsc	CHANGE	petsc
petsc	CHANGE	scripts/configure_petsc.sh
libmesh	CHANGE	conda/libmesh/conda_build_config.yaml
libmesh	CHANGE	conda/libmesh/meta.yaml
libmesh	CHANGE	libmesh
moose-dev	CHANGE	apptainer/moose-dev.def
moose-dev	CHANGE	conda/moose-dev/conda_build_config.yaml
moose-dev	CHANGE	conda/moose-dev/meta.yaml
moose-dev	CHANGE	scripts/update_and_rebuild_mfem.sh

Versioner versions

Found 9 packages, 7 changed, 0 failed

package	status	hash	to hash	version	to version
mpi	CHANGE	572cf81	7fc3ad9	2026.02.16	2026.03.10
wasp	CHANGE	ac33097	9ce59ea	2025.09.19_02960f1 build 4	2025.09.19_02960f1 build 5
seacas	CHANGE	9685477	2c80fef	2025.10.14 build 1	2025.10.14 build 2
libmesh-vtk	CHANGE	7e4f471	22871e0	9.6.0 build 0	9.6.0 build 1
petsc	CHANGE	d5b5601	0ea3b02	3.24.4 build 0	3.24.5 build 0
libmesh	CHANGE	47aa7d2	2e3352f	2026.02.18_f8a1758 build 0	2026.03.03_c889858 build 0
moose-dev	CHANGE	d2ceaf6	691ce2a	2026.02.20	2026.03.10

moosebuild · 2026-03-11T03:14:04Z

Job Test, step Results summary on 77e0690 wanted to post the following:

Framework test summary

Compared against 167d99f in job civet.inl.gov/job/3642141.

No change

Modules test summary

Compared against 167d99f in job civet.inl.gov/job/3642141.

No change

lindsayad · 2026-03-11T14:57:59Z

@joshuahansel could you think about whether the memory footprint of the two failing ~~sockeye~~ THM tutorial tests (https://civet.inl.gov/job/3645582/) could be reduced, or you could request more slots like @loganharbour suggested an an option on slack?

joshuahansel · 2026-03-11T15:30:59Z

I guess you mean the failing THM tests, not Sockeye tests 😄 Sockeye would never do such a thing.

Yes I'll address that now.

joshuahansel · 2026-03-11T16:10:44Z

I just ran those guys on my machine with current next:

tutorials/single_phase_flow.tutorial/05 ........................................................................................ OK [2.744s] [ 172MB]
tutorials/single_phase_flow.tutorial/06 ........................................................................................ OK [2.079s] [ 174MB]

Do we expect memory to be that different from the testing here?

[1.258s] [ 407MB]    ERROR tutorials/single_phase_flow.tutorial/06 FAILED (KILLED: OVER MEMORY) [max_cpus=1]
[20.10s] [ 382MB]    ERROR tutorials/single_phase_flow.tutorial/05 FAILED (KILLED: OVER MEMORY) [max_cpus=1]

loganharbour · 2026-03-11T16:26:34Z

I just ran those guys on my machine with current next:

tutorials/single_phase_flow.tutorial/05 ........................................................................................ OK [2.744s] [ 172MB]
tutorials/single_phase_flow.tutorial/06 ........................................................................................ OK [2.079s] [ 174MB]

Do we expect memory to be that different from the testing here?

[1.258s] [ 407MB]    ERROR tutorials/single_phase_flow.tutorial/06 FAILED (KILLED: OVER MEMORY) [max_cpus=1]
[20.10s] [ 382MB]    ERROR tutorials/single_phase_flow.tutorial/05 FAILED (KILLED: OVER MEMORY) [max_cpus=1]

I shared this with you on slack - but for public notice - how memory is tracked on linux vs on mac is very different. With that, it's not appropriate to compare those values across different systems. What we care about is setting a threshold for each recipe (which is system dependent), and that tests stay below those thresholds. Limits on macs and linux boxes will be different.

What you should try to do is keep a baseline on mac, make the test better, notice a decrease.

lindsayad marked this pull request as ready for review March 2, 2026 19:13

lindsayad requested review from cticenhour and loganharbour as code owners March 2, 2026 19:13

GiudGiud previously approved these changes Mar 3, 2026

View reviewed changes

loganharbour requested changes Mar 3, 2026

View reviewed changes

.gitignore Show resolved Hide resolved

lindsayad force-pushed the fix-cuda-typo branch from b197d08 to a183de9 Compare March 4, 2026 20:29

loganharbour changed the title ~~Fix cuda typo~~ Update packages: petsc, libmesh; fix openblas Mar 4, 2026

loganharbour changed the title ~~Update packages: petsc, libmesh; fix openblas~~ Update packages: petsc, libmesh, code-server; fix openblas Mar 4, 2026

loganharbour changed the title ~~Update packages: petsc, libmesh, code-server; fix openblas~~ Update packages: petsc, libmesh, code-server; fix openblas; download umpire Mar 4, 2026

loganharbour added the PR: Updates packages Pull requests that update versioner packages label Mar 4, 2026

loganharbour self-assigned this Mar 4, 2026

loganharbour added a commit to lindsayad/moose that referenced this pull request Mar 4, 2026

Update libmesh submodule

d766bdf

refs idaholab#32404

loganharbour added a commit to lindsayad/moose that referenced this pull request Mar 4, 2026

Update PETSc to v3.24.5

09b2f78

refs idaholab#32404

loganharbour added a commit to lindsayad/moose that referenced this pull request Mar 4, 2026

Update code-server

dd359af

refs idaholab#32404

loganharbour pushed a commit to lindsayad/moose that referenced this pull request Mar 4, 2026

Bump petsc, libmesh, moose-dev

d7b2d62

refs idaholab#32404

loganharbour added a commit to lindsayad/moose that referenced this pull request Mar 4, 2026

Update newsletter for package updates

7b1a78b

refs idaholab#32404

loganharbour added a commit to lindsayad/moose that referenced this pull request Mar 4, 2026

Update versioner hashes

1452632

refs idaholab#32404

loganharbour force-pushed the fix-cuda-typo branch from a183de9 to 1452632 Compare March 4, 2026 23:40

loganharbour requested review from grmnptr and roystgnr as code owners March 4, 2026 23:40

loganharbour pushed a commit to lindsayad/moose that referenced this pull request Mar 4, 2026

Bump petsc, libmesh, moose-dev

8b95b12

refs idaholab#32404

loganharbour added a commit to lindsayad/moose that referenced this pull request Mar 4, 2026

Update newsletter for package updates

9397e17

refs idaholab#32404

loganharbour added a commit to lindsayad/moose that referenced this pull request Mar 4, 2026

Update versioner hashes

20ce21b

refs idaholab#32404

loganharbour force-pushed the fix-cuda-typo branch from 1452632 to 20ce21b Compare March 4, 2026 23:43

moosebuild reviewed Mar 4, 2026

View reviewed changes