> > Does anyone actually understand how the crazy ene / ener mess in sander
> > actually works?
>
> I'd advise checking with Ross -- I think he has spent quite a bit of time
> working on this. ^^^^
|
ha ha <sob...> :'( <-------
> There is a set of comments starting at line 1099 of force.f which might be
> a start some real code comments.
Yeah I looked at them but they aren't complete. The main issue is being able
to extend the array (add some extra terms on the end) without completely
breaking everything at the moment. The problem is that the size of the array
is hand coded in so any places that it is a disaster to update it. I might
try creating some kind of module and placing it in there so at least the
array size definition and all the comments etc will be in the same place. I
guess this will need a day with door locked and the coffee setup to feed
intravenously :-).
> I believe the use of an offset of 23 goes back to Chandra Singh(!), but it
> might be even older (i.e. Paul Weiner).
Ughh...
> In the earlier attempt to implement the charmm ff (e.g. Amber9), I just
> added the Urey-Bradley energy into the bond term, and the impropers into
> the
> dihedral term, thus avoiding the problem you point out. But that was just
> a
> continuation of the "path of least resistance" approach.
This is what we have been doing which is okay for the time being but it
makes debugging things harder it also just puts of the inevitable since we
will have to add some kind of additional energy term for CMAP anyway.
Although this could probably be added to the dihedral terms as well but that
seems kind of unsatisfying.
I tried creating a structure inside the CHARMM module to hold the charmm
specfic energy terms and then print these from within dynlib and printe but
at this point I realize that all sorts of additional stuff is going on with
respect to the ene arrays, calculating averages, RMS energies etc which of
course don't work if I keep the CHARMM energies somewhere neat and tidy.
Oh well, I'll see how I get on. I might just have to take Bob's advice and
gut the thing. It is certainly in need of some "rm " optimization.
Have a good weekend.
Ross
/\
\/
|\oss Walker
| Assistant Research Professor |
| San Diego Supercomputer Center |
| Tel: +1 858 822 0854 | EMail:- ross.rosswalker.co.uk |
|
http://www.rosswalker.co.uk | PGP Key available on request |
Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be read every day, and should not be used for urgent or sensitive issues.
Received on Fri Dec 05 2008 - 16:23:16 PST
