[gmx-users] Lincs Errors Crashing Simulation

[Eric Smoll]

Hello,

I am using gromacs 2018.1.

Although several steps of an all-atom simulation proceed without problem
(minimization, NVT and NPT runs with berendsen temperature and pressure
control), attempts to equilibrate the system with the md integrator under
NVT conditions with a nose-hoover thermostat and a 1 fs time-step generates
lincs errors because of extreme molecular distortion.

Viewing a molecule that distorts and crashes the simulation seems to
suggest that this molecule has normal dynamics for thousands of timesteps.
Then, for a few ps before the crash, the molecule starts to show increasing
amounts of distortion before snapping to an extremely distorted state.

Any suggestions on how I can troubleshoot this problem?

Best,
Eric