Skip to content

Update packages#32757

Merged
loganharbour merged 15 commits into
idaholab:nextfrom
nmnobre:petsc
May 10, 2026
Merged

Update packages#32757
loganharbour merged 15 commits into
idaholab:nextfrom
nmnobre:petsc

Conversation

@nmnobre
Copy link
Copy Markdown
Member

@nmnobre nmnobre commented Apr 13, 2026 •
edited by loganharbour
Loading

libMesh update libMesh/libmesh@0185b8b...aeab0b9

  • Improved Kokkos configuration and compatibility in MetaPhysicL
  • Updated ThreadedBuildingBlocks compatibility for modern oneTBB
  • Enabled future fine-grained control of subvector type in create_subvector()
  • Added Netgen control for tetrahedralization multithreading; set Netgen thread count to match libMesh n_threads
  • More robust build system behavior when rebuilding after Netgen updates
  • More assertions in AnalyticFunction
  • Fix for compiler warning in libMesh from OSX 26.4
  • Fix for compiler warning in TIMPI
  • Clarification in TIMPI parallel sync documentation
  • IntRange (the output type of make_range) is now usable as a Threads:: range
  • NumericVector<T>::l2_norm_diff() is now multithreaded
  • When building with METHOD=oprof, app binaries with oprof settings are now also built
  • Added --enable-sigint command-line option, telling libMesh applications to handle SIGINT (the Ctrl+C signal). The usual termination handler can then print stack trace and any incomplete PerfLog output, which can make threading and profiling easier when investigating application cases which run very long or hang entirely.

Conda packages

moose-petsc 3.25.1 [mpich,openmpi]_0

moose-libmesh 2026.05.06_aeab0b9 [mpich,openmpi]_0

moose-dev 2026.05.08 [mpich,openmpi]

  • Use updated PETSc from moose-petsc 3.25.1 [mpich,openmpi]_0
  • Use updated libMesh from moose-libmesh 2026.05.06_aeab0b9 [mpich,openmpi]_0

Apptainer containers

moose-petsc:3.25.1_0

  • Update petsc v3.24.6..v3.25.1
  • Workaround missing libnvidia-ml.so.1 by passing -DCMAKE_EXE_LINKER_FLAGS=-Wl,--as-needed to SLATE's CMake command to drop that library dependency
  • Override OpenBLAS build settings if using the oneAPI compilers by passing DYNAMIC_ARCH=1 NOFORTRAN=0 F_COMPILER=INTEL FEXTRALIB= to its make command

moose-libmesh:2026.05.06_aeab0b9_0

  • Use updated moose-petsc:3.25.1_0 container as base
  • Update libmesh libMesh/libmesh@0185b8b...aeab0b9
  • Workaround missing libnvidia-ml.so.1 by passing -Wl,--as-needed to netCDF's configure command to drop that library dependency

moose-dev:2026.05.08

  • Use updated moose-libmesh:2026.05.06_aeab0b9_0 container as base
  • Update mfem mfem/mfem@64ef39b...e52948f
  • Update conduit llnl/conduit@8543892...4ae5a69
  • Workaround missing libnvidia-ml.so.1 by passing -DPETSC_EXECUTABLE_RUNS=YES to MFEM's CMake command to ignore failing PETSc tests
  • Remove extra bashrc source in code-server-start which is causing issues on INL HPC
  • Use more robust conda activation of moose environment that works in shell
  • Update code-server to 4.118.0
  • Update miniforge to 26.3.2-0

@nmnobre nmnobre added C: libMesh C: PETSc/SLEPc Tickets pertaining to PETSc/SLEPc PR: Updates packages Pull requests that update versioner packages C: MFEM Tickets pertaining to MFEM Backend integration labels Apr 13, 2026
moosebuild
moosebuild reviewed Apr 13, 2026
View reviewed changes
Comment thread libmesh Outdated
moosebuild
moosebuild reviewed Apr 13, 2026
View reviewed changes
Comment thread libmesh Outdated
@moosebuild
Copy link
Copy Markdown
Contributor

moosebuild commented Apr 13, 2026 •
edited
Loading

Job Documentation, step Docs: sync website on dab3bad wanted to post the following:

View the site here

This comment will be updated on new commits.

@NamjaeChoi
Copy link
Copy Markdown
Contributor

NamjaeChoi commented Apr 13, 2026

@lindsayad Can we make sure your MetaPhysicL fix will be in with this update?

moosebuild
moosebuild reviewed Apr 14, 2026
View reviewed changes
Comment thread libmesh Outdated
@nmnobre nmnobre force-pushed the petsc branch 2 times, most recently from 214c9c3 to d0b39a6 Compare April 14, 2026 18:01
moosebuild
moosebuild reviewed Apr 14, 2026
View reviewed changes
Comment thread libmesh Outdated
@loganharbour loganharbour removed C: libMesh C: PETSc/SLEPc Tickets pertaining to PETSc/SLEPc C: MFEM Tickets pertaining to MFEM Backend integration labels Apr 14, 2026
@lindsayad
Copy link
Copy Markdown
Member

lindsayad commented Apr 14, 2026

@lindsayad Can we make sure your MetaPhysicL fix will be in with this update?

sounds good

@nmnobre nmnobre force-pushed the petsc branch 2 times, most recently from bd0620d to 31b2ecc Compare April 14, 2026 22:16
moosebuild
moosebuild reviewed Apr 14, 2026
View reviewed changes
Comment thread libmesh Outdated
@nmnobre nmnobre force-pushed the petsc branch 2 times, most recently from 590522f to be31d30 Compare April 15, 2026 11:25
moosebuild
moosebuild reviewed Apr 15, 2026
View reviewed changes
Comment thread libmesh Outdated
moosebuild
moosebuild reviewed Apr 15, 2026
View reviewed changes
Comment thread libmesh Outdated
@nmnobre nmnobre force-pushed the petsc branch 2 times, most recently from 4604f4f to 69c6a60 Compare April 15, 2026 21:26
moosebuild
moosebuild reviewed Apr 15, 2026
View reviewed changes
Comment thread libmesh Outdated
@nmnobre nmnobre force-pushed the petsc branch 2 times, most recently from ec0df19 to 5cf5eee Compare April 15, 2026 22:17
@NamjaeChoi
Copy link
Copy Markdown
Contributor

NamjaeChoi commented May 7, 2026

What's the ETA of this? If it will take a few more days, can we include a small libMesh bug fix that Roy will make today or tomorrow?

@loganharbour
Copy link
Copy Markdown
Member

loganharbour commented May 8, 2026

What's the ETA of this? If it will take a few more days, can we include a small libMesh bug fix that Roy will make today or tomorrow?

We'll see. Because @roystgnr went on a unit test fix rampage I'm not sure if it'll be ready in time

loganharbour and others added 11 commits May 8, 2026 19:19
@loganharbour
Update mfem and conduit
15b7b44 
refs idaholab#32757
@nmnobre @loganharbour
Revert "Patch mfem::HypreGMRES/PCG to reset preconditioning matrix"
1ffe041 
This reverts commit d908ec3.

refs idaholab#32757
@nmnobre @loganharbour
Fix petsc.def:SLATE, libmesh.def:netCDF, moose-dev.def:MFEM build che…
7987a14 
…cks missing libnvidia-ml.so.1

For SLATE and netCDF we leverage the fact that the simple programs that
CMake and autotools use to test the compilers do not actually need
libnvidia-ml.so, they are just linked against it because that is an
inherited dependency from PETSc (as of 5268dc8a, !8817). We use
-Wl,--as-needed to drop that dependency at compile time, so the dynamic
linker does not complain at runtime. For MFEM, the check is on the PETSc
library itself, which we know must be working correctly since, by the
time we build MFEM, we have already compiled libMesh against PETSc. As
PETSc does require libnvidia-ml.so.1, we bypass the test altogether with
-DPETSC_EXECUTABLE_RUNS=YES.

refs idaholab#32757
@nmnobre @loganharbour
Override OpenBLAS build settings if using oneAPI compilers
b028344 
OpenBLAS uses a script, f_check, to determine what the Fortran compiler is.
Problem is, with FC=mpif90, even though mpif90 wraps ifx, OpenBLAS makes a
mess and assigns NOFORTRAN=1 and other flags relative to gfortran instead.
The net effect is that PETSc's post-install checks hang.

refs idaholab#32757
@nmnobre @loganharbour
Guarantee we use a matching pip for python 3.14
205815c 
refs idaholab#32757
@lindsayad @claude @loganharbour
Strip stale GCC toolchain paths from scalapack-targets.cmake in conda…
7a5eb3c 
… petsc recipe

The conda scalapack package's scalapack-targets.cmake embeds absolute CI
build-machine paths (/opt/civet0/...) and GCC emulated-TLS libs (-lemutls_w,
-lheapt_w) in INTERFACE_LINK_LIBRARIES. These propagate through MUMPS into
MFEM's link libraries, causing fatal link errors on macOS Apple Silicon.

Extend the existing Darwin GCC-stripping block to also cover
scalapack-targets.cmake: add it to the existing sed for -lemutls_w/-lheapt_w/
-lgcc_s.1, and add a second sed to strip the -Wl,-rpath and -L flags pointing
at /opt/civet0 (which don't appear in the PETSc files and so weren't handled
before).

refs idaholab#32757

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
@loganharbour
Remove code-server-start bashrc load
5a0a269 
This is causing crashes on INL HPC and I will do it elsewhere

refs idaholab#32757
@loganharbour
Use more robust conda activation
aa6b9a9 
This fixes conda activation when using sh instead
of bash in Docker

refs idaholab#32803
@loganharbour
Update code-server
c2d2eb4 
refs idaholab#32757
@loganharbour
Update miniforge
01d1e9b 
refs idaholab#32757
@loganharbour
Update libmesh
502451a 
Summary of changes:

- Improved Kokkos configuration and compatibility in MetaPhysicL
- Updated ThreadedBuildingBlocks compatibility for modern oneTBB
- Enabled future fine-grained control of subvector type in create_subvector()
- Added Netgen control for tetrahedralization multithreading; set Netgen thread count to match libMesh n_threads
- More robust build system behavior when rebuilding after Netgen updates
- More assertions in AnalyticFunction
- Fix for compiler warning in libMesh from OSX 26.4
- Fix for compiler warning in TIMPI
- Clarification in TIMPI parallel sync documentation
- `restrict_int<T>` can now be used in place of `cast_int<T>`, to enable integer conversion range checking in optimized modes, not just in devel and debug modes.
- `restrict_int` is now used in many mesh input classes, to improve safety when libMesh builds with small `subdomain_id_type` configurations are asked to read mesh files with too-large subdomain ids.
- `IntRange` (the output type of `make_range`) is now usable as a `Threads::` range
- `NumericVector<T>::l2_norm_diff()` is now multithreaded
- When building with `METHOD=oprof`, app binaries with `oprof` settings are now also built
- Added `--enable-sigint` command-line option, telling libMesh applications to handle `SIGINT` (the Ctrl+C signal).  The usual termination handler can then print stack trace and any incomplete `PerfLog` output, which can make threading and profiling easier when investigating application cases which run very long or hang entirely.

refs idaholab#32757
loganharbour added a commit to nmnobre/moose that referenced this pull request May 9, 2026
@loganharbour
Update newsletter with package changes
52dd95d 
refs idaholab#32757
loganharbour pushed a commit to nmnobre/moose that referenced this pull request May 9, 2026
@nmnobre @loganharbour
Bump petsc, libmesh, and moose-dev
c9a14df 
refs idaholab#32757
loganharbour added a commit to nmnobre/moose that referenced this pull request May 9, 2026
@loganharbour
Bump versioner hashes
cbfca8d 
refs idaholab#32757
moosebuild
moosebuild reviewed May 9, 2026
View reviewed changes
Comment thread libmesh
@moosebuild
Copy link
Copy Markdown
Contributor

moosebuild commented May 9, 2026

Job Precheck, step Versioner verify on dab3bad wanted to post the following:

Versioner templates

Found 11 templates, 0 failed

Versioner influential files

Found 53 influential files, 14 changed, 0 added, 0 removed

package status file
petsc CHANGE apptainer/petsc.def
petsc CHANGE conda/petsc
petsc CHANGE conda/petsc/meta.yaml
petsc CHANGE petsc
petsc CHANGE scripts/configure_petsc.sh
libmesh CHANGE apptainer/libmesh.def
libmesh CHANGE conda/libmesh
libmesh CHANGE conda/libmesh/meta.yaml
libmesh CHANGE libmesh
moose-dev CHANGE apptainer/files/moose-dev
moose-dev CHANGE apptainer/moose-dev.def
moose-dev CHANGE conda/moose-dev/meta.yaml
moose-dev CHANGE framework/contrib/conduit
moose-dev CHANGE framework/contrib/mfem

Versioner versions

Found 10 packages, 3 changed, 0 failed

package status hash to hash version to version
petsc CHANGE 2fee174 b035b4c 3.24.6 build 0 3.25.1 build 0
libmesh CHANGE b73472f 5a8fab9 2026.04.13_0185b8b build 0 2026.05.06_aeab0b9 build 0
moose-dev CHANGE 0bf4d15 821fba8 2026.04.21 2026.05.08

@loganharbour loganharbour dismissed their stale review May 9, 2026 02:24

Changes addressed

@moosebuild
Copy link
Copy Markdown
Contributor

moosebuild commented May 9, 2026

Job Test, step Results summary on dab3bad wanted to post the following:

Framework test summary

Compared against 1ba05b6 in job civet.inl.gov/job/3806839.

No change

Modules test summary

Compared against 1ba05b6 in job civet.inl.gov/job/3806839.

No change

@NamjaeChoi
Copy link
Copy Markdown
Contributor

NamjaeChoi commented May 9, 2026

The libMesh bug fix was merged

@loganharbour
Copy link
Copy Markdown
Member

loganharbour commented May 10, 2026 •
edited
Loading

32bit dof ids failures will be skipped in #32897

@loganharbour loganharbour merged commit 25e8bc7 into idaholab:next May 10, 2026
113 of 116 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@moosebuild moosebuild moosebuild left review comments

@lindsayad lindsayad lindsayad left review comments

@loganharbour loganharbour loganharbour left review comments

@NamjaeChoi NamjaeChoi NamjaeChoi left review comments

@cticenhour cticenhour cticenhour approved these changes

@alexanderianblair alexanderianblair Awaiting requested review from alexanderianblair alexanderianblair is a code owner

@roystgnr roystgnr Awaiting requested review from roystgnr roystgnr is a code owner

Assignees

@nmnobre nmnobre

Labels

PR: AI assisted Pull requests that utilized AI PR: Updates packages Pull requests that update versioner packages

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

7 participants

@nmnobre @moosebuild @NamjaeChoi @lindsayad @loganharbour @roystgnr @cticenhour