NPT Equilibration Stopped Due to Fatal PME Rank error

User discussions
1

GROMACS version: 2021.4
GROMACS modification: No

I did the NPT equilibration with a pressure of 1 bar using a Berendsen barostat for 50 ns. However, the NPT process crashed at 29 ns with an error message like the following:

Fatal errors:
43344 particles communicated to PME rank 0 are more than 2/3 times the cut-off out of the domain decomposition cell of their charge group in dimension x. This usually means that your system is not well equilibrated.

I’ve tried changing the barostat to Parrinello-Rahman and other parameters in grompp.mdp. However, it still stalled at 29 ns.
Here is my grompp.mdp:
npt.mdp (2.1 KB)

What should I do?
Please help me.
Thank You,
Ayyasy.

2

I suspect your crash is caused by the extremely short period of the Nose-Hoover barostat of 0.1 ps. That needs to be at least 10 times longer.

Generally we suggest to use the v-rescale thermostat and the c-rescale barostat.

1 Like
3

Thank you for your answer Sir.
Do you mean that the coupling time of Nose-Hoover thermostat need to be 1.0 ps?