Gromacs: mpirun noticed that process rank 63 with PID 0 on node *** exited on signal 11

User discussions
1

GROMACS version: 2018.3
GROMACS modification: No
Here post your question
Hello!
I use a server with 72 logical cores (18 physical cores, RAM capacity approaching 256GB) to run Gromacs MD simulation and enter “mpirun -np 72 gmx_mpi mdrun -v -deffnm md” to run it. Before that, the system has been minimized and equilibrated. But it’s failture.
The error message is:
Signal: Segmentation fault(11)
Signal code: Invalid permissions(2)
Failing at address: 0x2b6cfefe0ab8
*
Primary job terminated normally, but 1 process returned a non-zero exit code. Per user-direction, the job has been aborted.
mpirun noticed that process rank 63 with PID 0 on node mn01 exit on signal 11 (Segmentation fault).

65fd72b2587928d52041db131346816
65fd72b2587928d52041db1313468161920×844 186 KB

I searched online and someone said it was an amount of RAM question. Should I expand the server’s memory？Is it helpful for me to reduce the -np value？

Thank you!

2

I don’t think you need more RAM, but it might of course depend on the size of your system. I don’t claim to have the solution to your problem, but to at least improve the performance I would reduce the number of MPI jobs. Are you sure that you have 72 logical cores on 18 hardware cores?

I would suggest:
mpirun -np 18 gmx_mpi mdrun -v -deffnm md
or even
mpirun -np 9 gmx_mpi mdrun -v -deffnm md

It might not help your problem, but at least I think you’ll get better performance when the problem is solved. Also, keep in mind that the GROMACS version you are using has not been supported for many years. If there is a bug, it has hopefully been fixed in recent releases, e.g. 2023.4.

3

Dear Mr. MagnusL
My true gratitude for your generous help.
I am sure I have 72 logical cores on 18 hardware cores.
1706257072832
I will accept your suggestions and update the version of gromacs.
Thank you.