Then I vote yes to put the zeros back in. I prefer catastrophic failure in a simulation any day when compared to a simulation that *seems* to work, but gives the wrong answer. YMMV.
:-) Lachele
Dr. B. Lachele Foley
Complex Carbohydrate Research Center
The University of Georgia
Athens, GA USA
lfoley.uga.edu
http://glycam.ccrc.uga.edu
________________________________________
From: Jason Swails [jason.swails.gmail.com]
Sent: Friday, November 11, 2011 3:46 PM
To: AMBER Developers Mailing List
Subject: Re: [AMBER-Developers] significant recent tleap changes
On Fri, Nov 11, 2011 at 3:34 PM, B. Lachele Foley <lfoley.uga.edu> wrote:
> "I noticed that the *current* practice for adding 1-4 scaling factors into
> the topology file is actually broken by the recent changes."
>
> You're talking just about the 1-4 scaling factors assigned to impropers?
> Since, typically, none of those atoms is actually a "4" at all, does it
> make sense to assign numbers? Do they just get ignored anyway?
>
They are *supposed* to get ignored. That was the rationale behind putting
a 0 there. If they're really getting ignored, then those numbers are never
used. However, if they are being used (when they shouldn't), the fact that
you have 1/0 makes energies blow up. It's a fail-safe, as I understand it.
My main question was should we put those back in?
All the best,
Jason
> I just added "Make tleap tests" to my calendar. Will have some soon.
>
> :-) Lachele
>
> Dr. B. Lachele Foley
> Complex Carbohydrate Research Center
> The University of Georgia
> Athens, GA USA
> lfoley.uga.edu
> http://glycam.ccrc.uga.edu
>
> ________________________________________
> From: Jason Swails [jason.swails.gmail.com]
> Sent: Friday, November 11, 2011 2:54 PM
> To: AMBER Developers Mailing List
> Subject: [AMBER-Developers] significant recent tleap changes
>
> Hello everyone,
>
> I'm a little concerned about tleap and all the changes it's been seeing
> recently (not that it's creating bad topologies, yet). The recent changes
> have caused a small test failure in the NAB directory due to the new IPOL
> additions. However, this alerted me to what I considered a rather
> deficient tleap test suite. The current tests
> ($AMBERHOME/AmberTools/test/leap) only test the creation of OFF library
> files and old-style topology files(!) From the look of it, these are the
> only tests we have for tleap, is this correct? (Except the indirect ones
> like those in NAB).
>
> I created a test in the leap directory (added to the default test suite) to
> create an amber7 style topology (solvated and GB-ready) using an AmberTools
> 1.5 (fully patched) version of tleap. I noticed that the *current*
> practice for adding 1-4 scaling factors into the topology file is actually
> broken by the recent changes. The behavior we discussed in
> http://dev-archive.ambermd.org/201108/0000.html is no more (impropers no
> longer get 0's, they get 1.2/2.0).
>
> Do we want to go back to the original behavior (which would require adding
> some of Arunima's original code back in) or just live with the new version?
>
> Thanks!
> Jason
>
> P.S. -- Another note -- A topology file created prior to the 1-4 scaling
> addition in tleap actually differed quite a bit from the topology file
> created after the 1-4 scaling feature patch was added. It appeared under
> slightly-more-than-just-skimming scrutiny to be a slight rearrangement of
> some dihedral pointers, but I didn't do an exhaustive check. In any case
> it'll be nice to be able to trace tleap changes through a more extensive
> test suite (I don't think what I added was sufficient, but better than
> _nothing_).
>
> --
> Jason M. Swails
> Quantum Theory Project,
> University of Florida
> Ph.D. Candidate
> 352-392-4032
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers
>
>
>
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers
>
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Fri Nov 11 2011 - 13:00:04 PST
