Issue with PETSc Installation: C++ compiler (/opt/homebrew/bin/mpicxx) appears non-compliant with C++14
Rathore, Saubhagya
Dear ATS users and developers,
I am having some trouble installing master
branch. Current ATS master branch requires petsc and geochemistry to be enabled.
The main problem lies in my PETSc configuration step.
It seems that mpicxx is wrapping
clang++, which doesn't
support -std=c++14 as required.
Any advice on the correct compiler setup or flags would be greatly appreciated.
I receive the following error:
*********************************************************************************************
UNABLE to CONFIGURE with GIVEN OPTIONS (see configure.log for details):
---------------------------------------------------------------------------------------------
C++ compiler (/opt/homebrew/bin/mpicxx) appears non-compliant with C++14 or didn't
accept:
- -std=c++14
Note, you have explicitly requested --with-cxx-dialect=c++14. If you do not need C++14,
then remove this flag and let configure choose the most appropriate flag for you.
If you DO need it, then (assuming your compiler isn't just old) try consulting your
compilers user manual. There may be other flags (e.g. '--gcc-toolchain') you must pass to
enable C++14
*********************************************************************************************
Here are my system details:
Chip: Apple M1 Max
macOS: 15.4 (Sequoia)
gcc -–version
Apple clang version 17.0.0 (clang-1700.0.13.3)
Target: arm64-apple-darwin24.4.0
Thread model: posix
InstalledDir: /Library/Developer/CommandLineTools/usr/bin
gfortran -–version
GNU Fortran (Homebrew GCC 15.1.0) 15.1.0
which gfortran
/opt/homebrew/bin/gfortran
mpicc -–version
Apple clang version 17.0.0 (clang-1700.0.13.3)
Target: arm64-apple-darwin24.4.0
Thread model: posix
InstalledDir: /Library/Developer/CommandLineTools/usr/bin
which mpicc
/opt/homebrew/bin/mpicc
Thanks,
Saubhagya
Rathore, Saubhagya
Update: It appears that PETSc is currently incompatible with the compilers available on my machine. I have tried a number of remedies — including reinstalling Command Line Tools, doing a clean Homebrew setup, removing Anaconda, and downgrading Python — but the issue persists. Interestingly, today I was able to install another branch with PETSc disabled, and both the installation and subsequent simulations completed successfully without any issues.
As a short-term solution, it would be very helpful if the master branch could be made installable with PETSc disabled, at least for cases like mine where PETSc is problematic.
Thanks,
Saubhagya
--
You received this message because you are subscribed to the Google Groups "Amanzi-ATS Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
ats-users+...@googlegroups.com.
To view this discussion visit
https://groups.google.com/d/msgid/ats-users/MN2PR09MB5547D230AED0B4C5DB6CB2AAC266A%40MN2PR09MB5547.namprd09.prod.outlook.com.
Rathore, Saubhagya
Update: Many thanks to Ethan – the master branch has now been updated and no longer requires PETSc to be enabled for installation.
Best regards,
Saubhagya
To view this discussion visit https://groups.google.com/d/msgid/ats-users/MN2PR09MB554721C930DE9556BC0E2DDEC266A%40MN2PR09MB5547.namprd09.prod.outlook.com.
Rathore, Saubhagya
Dear developers,
Another update: it appears that the master branch, when installed on Mac, is unable to run ATS in parallel. The simulation fails with the following error:
MESH_PartitionWithZoltan: Using partitioning method RCB for ZOLTAN
ERROR!! in MESH_AddGhost(): This is not a valid mstk file
--------------------------------------------------------------------------
prterun detected that one or more processes exited with non-zero status,
thus causing the job to be terminated. The first process to do so was:
Process name: [prterun-MAC133412-53462@1,0]
Exit code: 255
Fengming and Phong have confirmed the same issue on their machines. I am not yet sure whether this is an Amanzi or ATS issue, so I have not created a GitHub issue.
Thanks,
Saubhagya
From:
'Rathore, Saubhagya' via Amanzi-ATS Users <ats-...@googlegroups.com>
Date: Friday, May 30, 2025 at 6:44 AM
To: Amanzi-ATS Users <ats-...@googlegroups.com>
Subject: [EXTERNAL] Re: Issue with PETSc Installation: C++ compiler (/opt/homebrew/bin/mpicxx) appears non-compliant with C++14
Update: Many thanks to Ethan – the master branch has now been updated and no longer requires PETSc to be enabled for installation. Best regards, Saubhagya From: 'Rathore, Saubhagya' via Amanzi-ATS Users <ats-users@ googlegroups. com> Date:
To view this discussion visit https://groups.google.com/d/msgid/ats-users/MN2PR09MB5547EE30CF1EE1AE7ED3F5F5C261A%40MN2PR09MB5547.namprd09.prod.outlook.com.
Coon, Ethan
To: Amanzi-ATS Users <ats-...@googlegroups.com>
Subject: [EXTERNAL] Re: Issue with PETSc Installation: C++ compiler (/opt/homebrew/bin/mpicxx) appears non-compliant with C++14
Rathore, Saubhagya
We ran column_infil_orig from integrated hydrology tests. I also tried another model with exodus mesh. Single core run was fine, parallel run threw that error. I will try hillslope tests later today and report back. My system details:
Chip: Apple M1 Max
macOS: 15.5 (Sequoia)
Thanks,
Saubhagya
Coon, Ethan
So column_infil_orig isn’t a regression test in 04_integrated_hydro. But in general, column tests cannot be run in parallel – they are one column and we only parallelize in the horizontal dimensions.
Maybe a better error message would be good, but that’s not unexpected.
Can you try, e.g.
04_integrated_hydro/hillslope_inf-np2?
This should be a parallel test, and it runs for me on current master on my mac.
Ethan
Rathore, Saubhagya
HI Ethan,
The hillslope_inf-np2.xml test is now running fine in parallel—I had to reinstall ATS to get it working. Interestingly, the first time I ran it with mpiexec, it took nearly an hour to start the simulation, but subsequent runs have started normally. Thanks again for help.
Best regards,