KILLED BY SIGNAL: 9 (Killed) during NpT simulation

[kai_oshiro]

Dear developers,

I encountered the following error when I switched from NVT simulation to NpT simulation.

===================================================================================
= BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
= RANK 0 PID 3276411 RUNNING AT ccc126
= KILLED BY SIGNAL: 9 (Killed)
===================================================================================

I think this error is related to insufficient memory, but I don't know why the calculation crashes only for the NpT simulation.

Is there any way to avoid running out of memory?

Best regards,
Kai

[jonathan_lahnsteiner2]

Dear kai_oshiro,

Signal 9 is sent by the Linux operating system to kill your job. This is usually not caused by the vasp code itself.
You can find some information about the sigkill command on this page:
https://komodor.com/learn/what-is-sigki ... ontainers/

I also checked your input files and they seem to be correct for what you want to do.
When running your job I checked the memory consumption. I am running on 8 cores and the
total amount is 2.5GB.
I would try to rerun the job and check if the error occurs again. Maybe you reached some time limit in your job scheduler.

If the error is still occurring please contact us again.


All the best Jonathan

[kai_oshiro]

Dear jonathan_lahnsteiner,

Thank you for the detailed reply.

I tried on different settings and sorted it out.

I usually employ a combination of MPI and OpenMP to conserve memory usage because I use a computer whose memory is limited to 1.875GB/core, and the limitation sometimes causes problems in my calculations.

However, I changed those settings to the standard MPI with NCORE again, and my calculation ran correctly now.

Sorry for the trouble.

Best regards,
Kai