From: Norman Geist (norman.geist_at_uni-greifswald.de)
Date: Wed Feb 27 2013 - 01:24:05 CST
A timestep of 1 is the default and should be stable, no need for 0.1.
Please attach the output of the namd plot plugin from vmd from your
minimization output file, or the output file itself. Additionally, check
your structure for superimposed atoms. Are there some high energy values
like 999999 in your output?
> -----Ursprüngliche Nachricht-----
> Von: Mihaela Drenscko [mailto:quo.physics_at_gmail.com]
> Gesendet: Mittwoch, 27. Februar 2013 05:03
> An: Norman Geist
> Betreff: Re: namd-l: RATTLE algorithm
> Hi Norman,
> Thank you for your response.
> I reduced the timestep to 0.1 fs and increased the numsteps to 10000.
> Yet, I get the same problem with the algorithm. What else can I do?
> On Tue, Feb 26, 2013 at 1:38 AM, Norman Geist
> <norman.geist_at_uni-greifswald.de> wrote:
> > Hi Mihaela,
> > the RATTLE algorithm belongs to the "rigidbonds", needed to remove
> > degrees of freedom from the hydrogens, to prevent them from
> oscillating and
> > so save computation time, as nobody cares this osccilation (btw would
> > save the frames so fine grained to still see it later). The error you
> > points out, that a hydrogen moved that far within one timestep, that
> > couldn't be restrained to its original distance with its mother atom
> in the
> > maximum iterations or tolerance. And this, finally, happens when
> > move to fast. This can be caused by either a bad initial structure (a
> > minimization should help than) or a too large timestep ( > 1), maybe
> > due some other things, but you should check your minimization output,
> > there were some messages about bad contacts. Also you can plot your
> > minimization output file with VMD (NAMDplot), to check if your TOTAL
> > converged (export it to xmgrace, and zoom in), if not, try more
> > steps.
> > 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 Mihaela Drenscko
> >> Gesendet: Dienstag, 26. Februar 2013 00:56
> >> An: namd-l_at_ks.uiuc.edu
> >> Betreff: namd-l: RATTLE algorithm
> >> Hi,
> >> I am running a simulation and I get the following error:
> >> ERROR: Constraint failure in RATTLE algorithm for atom 82!
> >> ERROR: Constraint failure; simulation has become unstable.
> >> ERROR: Exiting prematurely; see error messages above.
> >> How do I fix it?
> >> Thanks,
> >> Mihaela
This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:20:57 CST