AMBER Archive (2006)

Subject: Re: AMBER: another question: is rms 2E-2 good for nmode ntrun=1?

From: David A. Case (case_at_scripps.edu)
Date: Fri May 12 2006 - 10:39:37 CDT

On Fri, May 12, 2006, Pan, Yongmei wrote:
>
> another question for starting structure for nmode ntrun=1. After 10000
> cycles of GB(igb=2) follwed by 10000 cycles of epsilon=4r sander
> minimization, the rms wandered around 2E-2, and it seems that it would take
> forever to reach E-3 magnitude. Is this because my molecule with 6420 atoms
> is too big to reach a small rms? and if the rms cannot be smaller, is rms
> value 2E-2 small enough for the nmode ntrun=1 calculation?

The four main options:

1. Be patient; that is, run *lots* of cycles of minimization.
2. Write a better minimizer.
3. Convince someone else to write a better minimizer.
4. If you have lots of memory available, NAB can minimize large molecules
    like this down to gradients of 10**-12 or so.
    <http://www.scripps.edc/case>.

Myself, I mainly use option 1. Options 2-4 are probably better, but it's
usually easier just to set maxcyc=100000 or so, and go have some coffee while
the computer is running....

...dac
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu