From: Norman Geist (norman.geist_at_uni-greifswald.de)
Date: Wed Feb 22 2012 - 00:24:02 CST
Also, if the initial density of your solvents is to less, the periodic box shrinks to fast, u can then set margin 2 f.i., which prevents this error, with little influence on speed.
> -----Ursprüngliche Nachricht-----
> Von: Norman Geist [mailto:norman.geist_at_uni-greifswald.de]
> Gesendet: Dienstag, 21. Februar 2012 07:28
> An: 'namdlist_at_gmail.com'
> Cc: Namd Mailing List (namd-l_at_ks.uiuc.edu)
> Betreff: AW: namd-l: Keep PBC box size in Z-direction constant?
> You can hold it only in XY with "useconstantarea", so you could simply
> rotate your box and let pressure happen in Z. The reason why your
> simulation crashes might be fixed atoms at the boundaries of the
> periodic box, which cannot be rescaled or something very similar. Often
> this error comes with fixed atoms. You could try to increase the
> relaxation time of the pressure control if you do not have fixed atoms
> at the boundaries, but inside the system, as sometimes it happens that
> the atom coordinate rescaling due to pressure, when happening in too
> big step, creates heavy forces when rescaling free atoms into fixed
> Feel free to ask further.
> Good luck
> Norman Geist.
> > -----Ursprüngliche Nachricht-----
> > Von: owner-namd-l_at_ks.uiuc.edu [mailto:owner-namd-l_at_ks.uiuc.edu] Im
> > Auftrag von Bjoern Olausson
> > Gesendet: Montag, 20. Februar 2012 13:37
> > An: namd-l_at_ks.uiuc.edu
> > Betreff: namd-l: Keep PBC box size in Z-direction constant?
> > Hi NAMD users,
> > is it possible to keep the PBC-Box size in the Z-direction fixed?
> > I am asking because I want to run a NPT simulation in which the box
> > size is not allowed to shrink/extend along the Z-axis.
> > I tried to disable useFlexibleCell but after a few nanoseconds the
> > simulation crashes with the message "FATAL ERROR: Periodic cell has
> > become too small for original patch grid! ..."
> > To work around this I decreased the number of steps after which I
> > restart the simulation from a checkpoint to 5000ps and increased the
> > margin. But nothing helps. After 1.5ns the simulation just keeps
> > crashing no matter how small I set the individual run or how large I
> > choose the margin.
> > So is there a way to allow fluctuations along X any Y and keep Z
> > or any other workaround to prevent the above error?
> > Cheers,
> > Bjoern
This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:21:14 CST