[gmx-users] Fatal error: The X-size of the box times triclinic skew factor is smaller than the number of DD cells times the smallest allowed cell size
Justin Lemkul
jalemkul at vt.edu
Mon Jun 1 13:48:47 CEST 2015
On 6/1/15 6:39 AM, Kevin C Chan wrote:
> Dear Users,
>
> I have encountered "Fatal error: The X-size of the box times triclinic skew
> factor is smaller than the number of DD cells times the smallest allowed
> cell size" during a pre-equilibrium (NPT) run. It is NOT an annealing
> process and temperature was coupled to 300K, pressure to 1.0 bar. It was a
> 1000-water plus 100 amino acid system.
>
> I have gone through the mailing archive and it was reported that this might
> result from too-fast-shrinking systems. However I intended to shrink the
> box during the NPT run as initially it was generated using genbox tool and
> what I wanted to fix is only the number of solvent and additional
> particles. As I do not know the suitable box size for the system so I guess
> a little bigger box size (6 nm) and let it shrink during the NPT run. This
> was successful with a previous system (1000 water, 50 amino acids) and the
> box size shrank to 4.X nm.
>
> Would this be a wrong approach? And how could I deal with the reported
> error?
>
The error suggests indeed the box is collapsing, so your initial box is too
large. The box vectors are written to the .edr file so you can see this for
yourself if enough frames are written before the crash. Either (1) calculate a
proper size for the box, even if it is based on a rough estimate of the density
this is trivial to do and/or (2) relax the barostat coupling time to allow for a
more gentle equilibration.
-Justin
--
==================================================
Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow
Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201
jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul
==================================================
More information about the gromacs.org_gmx-users
mailing list