Drude-compatible GROMACS code 2

Hello everyone,

during the NVT equilibration of my system I have a segmentation fault, even with a single molecule in a box. My system is a tetradecane molecule I have readapted as a residue from CHARMM FF. I guess the problem is the topology, but I don’t know what.

Thank you in advance

Actually this is the .log file:

There are: 7788 Atoms
There are: 2478 Shells

Will never predict shell positions
Atom distribution over 24 domains: av 427 stddev 47 min 342 max 485
Initial temperature: -nan K

Started mdrun on rank 0 Thu Nov 19 20:02:36 2020
Step Time
0 0.00000

And this is the fatal error:

Fatal error:
1979 particles communicated to PME rank 2 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.

If a single molecule fails then your topology is unsound. Parametrizing polarizable molecules requires great care and it is easy to generate a set of parameters that “works” (i.e. agrees with target data) but is ultimately unsuitable for a simulation due to overpolarization issues.

Thank you for the answer.
What about the charmm2gmx phyton script to convert? Is it available?

That won’t work for Drude, and we do not have generally available utilities for any such conversion.

I mean the header in aminoacids.rtp says:

; This force field generated by charmm2gmx_v5.2.multiple.drude.py from
; the following Drude force field files:
; 2013f protein and model compounds
; 2017c DNA/RNA
; 2017c lipid
; 2018a carbohydrate

I have scripts that convert the entire force field to GROMACS format. I do not have scripts that generate Drude topologies or that are suited to standalone conversions of CHARMM RTF entries to GROMACS residue definitions.