Mdrun without openMP

User discussions
1

GROMACS version:2019
GROMACS modification: Yes/No
Here post your question

I am trying to execute gmx_mpi mdrun -v -deffnm md >& log in a cluster. But the job is aborted with an error message: “OpenMP threads have been requested with cut-off scheme Group, but these are
only supported with cut-off scheme Verlet”. I am using cutoff-scheme=group. In my system, some non-bonding energy calculations between energy groups are excluded using energygrp_excl in MDP file. If I am trying to use verlet cutoff-scheme, gromacs reports an error that energygrp_excl can only be used with group cutoff-scheme, has not yet been implemented in verlet cutoff scheme.

Is there a way I can run with group cutoff-scheme and energygrp_excl together without error using mdrun?

2

Hi mjm91,

no, this is not possible.

But if you tell us what you’re trying to evaluate with energy exclusion groups, we might be able to find a way around here.

3

Hi cblau, thank you for your response. The simulation system consists of a cavity and a ligand molecule, where we want to exclude any kind of intramolecular interactions. So that neither the cavity nor the ligand will change their conformation during the course of simulation. Position restraints are applied on the cavity and cavity atoms has also been frozen (freezegrps), but ligand is free to move. So, in order to allow its free movement and not having its conformation changed, we added the energygrp_excl, so that any non-bonding intramolecular interaction energies are excluded.

4

Ah okay, if I understand correctly, the ligand should move as a rigid body. That’s indeed a bit tricky - I know that’s a bit hacky, but would it work to alter the ligand parameters, so that you have very stiff dihedrals to emulate the rigid body behaviour?

5

Actually the ligand is a fullerene molecule. So, switching off intramolecular interactions seemed reasonable.