Re: problem about pairlistdistance...

From: hl332_at_drexel.edu
Date: Tue Apr 10 2007 - 16:23:22 CDT

Hi Peter,
   Thanks a lot for such a detailed reply. This thing was being displayed in the output file from beginning (I just typed 100000 as an example). No, my simulation is in explicit solvent box and I am using 1fs time step. I did not find it in the output until i gave an explicit parameter in config file to output *info* about pairlist. As you said, it is not harmful and now I understand why it does that, it no more bothers me.
Thanks again
Harish

-------------------------------------------------
Harish Vashisth (Ph.D Candidate)
CAT-361,Chemical & Biological Engg.
Drexel University, Philadelphia, PA
office: 215-895-5823

----- Original Message -----
From: Peter Freddolino <petefred_at_ks.uiuc.edu>
Date: Tuesday, April 10, 2007 2:39 pm
Subject: Re: namd-l: problem about pairlistdistance...

> Hi Harish,
> anything in info tags, including this, is generally not considered
> harmful; indeed, I don't think this output is normally even given in
> NAMD 2.6. What this particular message means is that NAMD tried to
> dynamically shrink your pairlist distance to speed computation, found
> that such a change could cause problems, and backed off. It shouldn't
> have any effect on dynamics. However, it is odd that this doesn't show
> up until step 100000. Is there anything unusual about your simulation
> (e.g., is it in gas phase?) What time step are you starting it?
> You can enforce some extra padding on the pairlist distance by
> adding a
> nonzero margin if this bothers you.
> Peter
>
> hl332_at_drexel.edu wrote:
> > Hi All,
> > In my NVE simulation log file, i am getting some warnings
> about pairlistdistance parameter as:
> > info:Pairlistdist is too small for 20 patches during timestep 100000
> >
> > I looked through namd users guide and tutorial which tells it
> can be due to small pairlistdist or long cycle. My parameters are
> as follows:
> >
> > cutoff 12
> > switchdist 10
> > pairlistdist 16
> > steprspercycle 10
> >
> > As far as i know ~"4A" is enough cushion between cutoff and
> pairlistdist to account for all pairs. Before I had set it to 14
> but increased to 16. Still, these info:(warnings) haven't stopped.
> I tried changing cycle which didnt help either.
> >
> > If someone has more experience in playing around with them,
> Please advise.
> >
> > Thanks
> > Harish
> > -------------------------------------------------
> > Harish Vashisth (Ph.D Candidate)
> > CAT-361,Chemical & Biological Engg.
> > Drexel University, Philadelphia, PA
> > office: 215-895-5823
> >
>

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:44:33 CST