AMBER Archive (2009)

Subject: Re: [AMBER] using sqm from AmberTools 1.3

From: David A. Case (case_at_biomaps.rutgers.edu)
Date: Wed Dec 23 2009 - 08:16:04 CST

On Wed, Dec 23, 2009, Alan wrote:

> QMMM: WARNING!
> QMMM: Unable to achieve self consistency to the tolerances specified
> QMMM: No convergence in SCF after 1000 steps.
> QMMM: Job will continue with unconverged SCF. Warning energies
> QMMM: and forces for this step will not be accurate.
> QMMM: E = -0.1216E+07 DeltaE = -0.8180E+02 DeltaP = 0.2721E+00
> QMMM: Smallest DeltaE = -0.2763E-02 DeltaP = 0.1572E-01 Step = 730
>
>
> QMMM: WARNING!
> QMMM: Unable to achieve self consistency to the tolerances specified
> [snip]
>
> First, I note that E = -0.1210E+07 doesn't change, but I am afraid that
> with standard criteria for convergence given by sqm "qm_theory=’AM1’,
> grms_tol=0.0002, tight_p_conv=1, scfconv=1" it will never converge anyway.

>
> So, what I am trying to figure out is if I see for others cases things
> similar to that above, is it safe to kill the process and try to restart
> with a looser criteria as scfconv to 1.d-8 and/or tight_p_conv to 0?

Yes, I'd certainly suggest killing the process and running with looser
criteria. If you are willing to post the problem molecule, that would help
us...we may want to change the default criteria for am1-bcc charges:
basically, problems are more likely to occur if you start with a bad geometry
than with a pretty good one; it may also depend on the electronic structure
of the molecule in question.

...dac

_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber