[gmx-users] simulation crashed without error information

[Qing Zhu]

Hi everyone,

I was running parallel MD simulations on a big DNA/water system which 
had about 300000 atoms total. After running for some certain time ( 
several hundred picoseconds), the trajectory files and log files stopped 
being written, and I thought this meant that the simulation had already 
crashed. But "top" command showed that the mdruns were still running in 
the system. After I restarted the runs using tpbconv, the same ting 
happened again. What I don't understand here is that why the mdrun was 
still running but nothing new was written into the output files. There 
was even no message in the log file, and I could not find any bad 
contact or overlap which can cause the crash in the system.

Does anyone know what is going on here? And how can I solve this problem?

Best regards,

Qing