AMBER Archive (2009)

Subject: Re: [AMBER] sqm x mopac in AmberTools 1.3

From: David A. Case (case_at_biomaps.rutgers.edu)
Date: Tue Dec 22 2009 - 21:50:35 CST

On Tue, Dec 22, 2009, Alan wrote:
>
> Then I tested antechamber on a ligand with sqm and mopac (which I just
> copied mopac and mopac.sh from amber10/bin into amber11/bin):
>
> First: with mopac, I got things done in 1min 20 s, with sqm it did in 15m 20
> s. Yes, 14 min more!
>
> The results for partial charges were almost identical.

The new sqm program has *much* more stringent convergence criteria than did
the old mopac program, in an attempt to make am1-bcc charges more
reproducible. This can increase computation times, as you have found.

See note 6 on p. 82 of the Users' Manual for a discussion. If computation
time is a problem, you can loosen the SCF and geometry convergence criteria,
as discussed there.

> Mopac saves a file mopac.pdb which is an optimised structure. Does sqm do
> the same? How?

The final coordinates are in the sqm.out file. You could also uncomment
line 555 in amber11/src/antechamber/charge.c, to have antechamber write
the optimized coordinates as it has done with mopac.

...dac

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