Re: ABF simulation with extendedLagrangian on keeps crashing

From: Jérôme Hénin (jhenin_at_ifr88.cnrs-mrs.fr)
Date: Thu May 12 2011 - 04:36:20 CDT

Hi Ajasja,

I see that you had the idea to avoid the singularity at theta = 0 or
180. What I've done recently is to keep the full grid from 0 to 180,
but set lowerWall to 5 and upperWall to 175 with a mild wall constant
(on the order of 1 kcal/mol/degree^2). This way you still get the full
PMF (the wall potentials have no effect on the measured free energy
gradient). This might help in the 2D simulation.

You should also check whether the extended system coordinate is
getting unstable as well. If it is the case, you can turn it off, now
that ABF for angles is available. Actually I recommend turning it off
in any case.

Cheers,
Jerome

On 12 May 2011 10:55, Ajasja Ljubetič <ajasja.ljubetic_at_gmail.com> wrote:
> Hi Jerome and Giacomo,
> A status update on the crashing theta and phi simulations.
> I moved the dummy atoms (so that they are inside the periodic cell) and
> also increased the limits on thetha from 0 to 180 to 0 to 10 to 170 degrees.
> Now a 1D theta or 1D phi ABF simulations runs without problems for 100 ns. I
> still have to analyse the results in grater detail.
> A 2D thetha phi ABF simulations crashes after about 30 ns with the following
> error:
>
> WRITING EXTENDED SYSTEM TO RESTART FILE AT STEP 687000
>
> FATAL ERROR: Low global exclusion count!  System unstable or pairlistdist or
> cutoff too small.
>
> FATAL ERROR: See http://www.ks.uiuc.edu/Research/namd/bugreport.html
>
> If I restart the simulations it runs ok for another ~30 ns seconds and then
> crashes again. If anybody is curios I'm attaching the colvars config file.
>
> Best regards,
>
> Ajasja
>

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 05:23:55 CST