On Wed, Jun 21, 2017, Alessandro Contini wrote:
>
> I'm using antechamber to compute bcc charges for several ligands. I've run
> the same simulation using ambertools15 and ambertools17, using either mopac
> (MOPAC2016.exe, version: 16.230L, under Ubuntu 16.04) or sqm. I noticed
> that, while sqm bcc charges were consistent among the two AmberTools
> versions, those computed by mopac were subtantially different.
We have not used or tested mopac in ages. Basically, we moved away from
mopac because we could not get consistent SCF convergence on different
machines. (But MOPAC2016 might be better than earlier versions; one
additional problem is there are literally dozens of version of mopac available
on the web, and it became impossible to check all of them.)
> I noticed that the mopac.in files for AT15/mopac and AT17/mopac differed for
> a blank line between title and coordinates. I've tried to insert the blank
> line back, by modifying the mopac.sh script, and the problem was fixed:
Please examine the mopcrt.c file in $AMBERHOME/AmberTools/src/antechamber.
In the wmopcrt() routine, there are comments in there about the problem with
an extra blank line. It looks like you can just uncomment the "original
correct statement:" line, and re-compile.
Scott: can you re-look at this? It looks like the patches made on Jan 3, 2017
may not be fully correct(?) According to this user, we should make sure
we retain the behavior of earlier versions of antechamber.
Alessandro: if modifying the mopcrt.c file fixes you problem, please let us
know, and we will correct our codes.
Thanks for the report....dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Jun 21 2017 - 06:00:15 PDT
