Thanks to all for the help!
Actually, changing the cutoff value
to 9999.0 did prevent minimization from stopping before reaching either maxcyc
or drms.
I was thinking about upgrading to a
more recent version of Amber but for these xmin tests I need to use Amber10 to
make some comparisons with previous results. I will also try the NAB
implementation of lmod.
Kind regards,
Fabiana
> From: jason.swails.gmail.com
> To: amber.ambermd.org
> Date: Wed, 23 Apr 2014 07:27:03 -0400
> Subject: Re: [AMBER] xmin
>
> On Wed, 2014-04-23 at 12:15 +0200, Fabiana Caporuscio wrote:
> >
> > Dear Amber
> > users,
> >
> > I am
> > experiencing a problem that has been already reported in the list, but,
> > unfortunately, I was not able to find out a solution among the replies.
> >
> > I run some
> > tests with the Amber10 version of xmin and I found that the xmin minimizer
> > stops before it reaches either maxcyc or drms. I tried different drms and
> > maxcyc values but minimizations always stop at the same step returning the same
> > results. At every step RMS=0.0 values are reported except in the 'FINAL
> > RESULTS' section. This is one of my input files:
>
> Amber 10 is ancient in computer software years and has not been
> realistically supported since probably ~2010 or 2011. Please upgrade to
> the latest version of AmberTools 14 (very few of us even have a copy of
> Amber 10 that we can run tests on).
>
> >
> >
> >
> >
> >
> >
> >
> > &cntrl
> >
> > ntb=0, ntmin=3, maxcyc=5000, drms=0.01,
> >
> > imin=1, ntx=1, ntpr=10,
> >
> > ntf=1, dielc=1, cut=12.0,
>
> This is an awfully short cutoff for a non-periodic system. Consider
> using one a bit bigger... like infinity (1000 is the same thing for most
> systems).
>
> >
> > /
> >
> > &lmod
> >
> > xmin_method = 'PRCG'
> >
> > /
>
> I think lmod is broken in sander and has been for some time (but don't
> quote me on it).
>
> If you want to use lmod, I would suggest the NAB implementation, which
> has been working fine as far as I know. (There is a test case in
> $AMBERHOME/AmberTools/test/nab that tests lmod and it should pass).
>
> You can use that test case ($AMBERHOME/AmberTools/test/nab/tlmod.nab) as
> a starting point for your calculation and just modify it as necessary.
>
> Good luck,
> Jason
>
> --
> Jason M. Swails
> BioMaPS,
> Rutgers University
> Postdoctoral Researcher
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Apr 23 2014 - 08:30:02 PDT
