Dear Professor David
Thanks for your help.
Your recommendations solved the problems.
How can I amend the number of LBFGS steps in "LBFGS TNGC method"?
Setting lbfgs_memory_depth value to 3 (default) would run LBFGS preconditioning in TNCG minimization?
Cheers
----- Original Message -----
From: David A Case <case.biomaps.rutgers.edu>
To: Anglea A. <a.anglea90.yahoo.com>; AMBER Mailing List <amber.ambermd.org>
Cc:
Sent: Wednesday, November 2, 2011 1:58 PM
Subject: Re: [AMBER] Minimization in sander
On Wed, Nov 02, 2011, Anglea A. wrote:
> The NSTEP is the final result.
> The input of my job is:
> &cntrl
>
> imin = 1,
> maxcyc = 100000,
> drms = 1.0E-5,
> ntb = 0,
> ntmin = 3,
> cut = 15.0,
> ntb=0,
> igb=0,
> /
> &lmod
> xmin_method = "LBFGS"
I'd recommend setting maxcyc to 500 or so, and ntpr to 10 (or even 1), and
following the course of the minimization. You have a vacuum system, so
using cut of 9999. (essentially infinite) might also be necessary.
The xmin defaults are almost always fine...I wouldn't worry about them.
The fact that your minimization ended after only 27 steps suggests some
problem (my guess is that it is the cutoff, but look in the output for clues.)
...dac
_______________________________________________
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 Thu Nov 03 2011 - 05:00:03 PDT
