AMBER Archive (2003)

Subject: Re: Gibbs stops after a finite number of window runs

From: caldwell_at_heimdal.compchem.ucsf.edu
Date: Sat Feb 01 2003 - 16:42:54 CST

What version of Gibbs is this? The older versions have a CPU
timer built in which doesn NOT default to unlimited time.
That could explain the "magic" 35 windows stopping point.

jim

On Sat, 1 Feb 2003, Seongeun Yang wrote:

> Thanks for your answer first.
> Yeah, I'm trying to do dynamically modified windows calulations.
> (I already tried using ISLDYN=-3 for the fixed window calculations.)
> With those input parameters, gibbs stopped after calculating from
> lambda=1 to lambda=0.85
> and the number of windows was 35.
> I tried again with the restart file starting from lambda=0.85, but gibbs
> stopped again
> around when lambda=0.65, and the number of windows was 35 also in this case.
> What I'm wondering about is whether this number is already set in the
> program as the default
> or there is something missing or wrong in my input parameters.
>
> Thanks again,
>
> Seongeun
>
> David A. Case wrote:
>
> >On Sat, Feb 01, 2003, Seongeun Yang wrote:
> >
> >
> >>I'm doing perturbation calculations using gibbs with the following input.
> >>But after calculating only 35 windows, gibbs finishes running for both
> >>FEP and TI.
> >>I can't see any reason for this finite number of windows.
> >>
> >>
> >
> >..
> >
> >
> >
> >>almda = 1.0, almdel = 0.02,
> >>almdl0 = 0.01,
> >>isldyn = -2, idifrg = 0,
> >>
> >>
> >
> >By setting isldyn=-2, you are asking for "dynamically modified windows".
> >This means that the number of windows is not determined in advance, but
> >is adjusted by the program. The value of almdel is ignored.
> >
> >You should look at the values of the lambdas that were used. If nothing
> >went wrong, they should roughly (but not uniformly) span the range from
> >1 to 0. I'm not sure why you were worried that there were "only" 35
> >windows...this is actually quite a respectable number.
> >
> >If you want to specify in advance the number of windows, set isldyn=-3.
> >
> >..hope this helps....dac
> >
> >
> >
>
>
> 

-- 

----------------------------------------------------------------------------
James W. Caldwell                         (voice) 415-476-8603
Department of Pharmaceutical Chemistry    (fax)   415-502-1411  
Mail Stop 0446                            (email) caldwell_at_heimdal.ucsf.edu
513 Parnassus Avenue                      
University of California                  
San Francisco, CA 94143-0446               

Starting Feb 11, 2003:

Department of Pharmaceutical Chemistry
UCSF Box 2240
600 16th Street
University of California
San Francisco, CA 94143-2240

----------------------------------------------------------------------------