Test simulation issue

[JihoonKimKorea]

Hi,

I am trying to run an example simulation after compilation.

It seems there is an issue with ParMETIS related to MPI segmentation.

I would greatly appreciate it if someone could guide me on how to resolve this issue.

I've attached the error message below.

ucns3d_p will run on 2 MPI tasks.
Authorization required, but no authorization protocol specified
Authorization required, but no authorization protocol specified
i read the bleed BC
RESTARTING 0 0.0000000000000000 0

                    ParMETIS Initiated                               

---

[kucfd-System:909616] *** An error occurred in MPI_Comm_rank
[kucfd-System:909616] *** reported by process [1937178625,0]
[kucfd-System:909616] *** on communicator MPI_COMM_WORLD
[kucfd-System:909616] *** MPI_ERR_COMM: invalid communicator
[kucfd-System:909616] *** MPI_ERRORS_ARE_FATAL (processes in this communicator will now abort,
[kucfd-System:909616] *** and potentially your MPI job)
[kucfd-System:909609] 1 more process has sent help message help-mpi-errors.txt / mpi_errors_are_fatal
[kucfd-System:909609] Set MCA parameter "orte_base_help_aggregate" to 0 to see all help / error messages

[TakisCFD]

can you share some more info:

1. platform used
2. compiled natively or through docker
3. UCNS3D.DAT file
4. OMP_NUM_THREADS= ?

[TakisCFD]

I assume that there might be an issue with the parmetis since this is when the code crashes.
Please check the following:

1. If you are running a 2D case do you have the 2D option in the UCNS3D.DAT file?
2. check the chmod of ucns3d_p since the authorisation required message might indicate something related to that?
3. Check what compilers combo you are using (Gnu with OpenMPI/MPICH should work, Intel fortran with Intel MPI should work, GNU with CrayMPI/HPE mpi should work, however when switching with intel fortran and openmpi or gnu fortran with intel mpi you might get these issue.

[dudugang]

I got the same error with Gnu compiler and OpenMPI