On Tue, Nov 27, 2012 at 6:20 AM, Elise Duboué-Dijon <
elise.duboue-dijon.ens.fr> wrote:
> We believe we have finally found how to run sander with up to 100 atom
> types.
> The required changes are:
> - in rdparms.F90, line 108 : nttyp > 5050 (instead of 1830 previously)
> - in parms.F90, lines 30-35: the dimension of the arrays cn1, cn2 should
> be
> increased from 1830 to 5050
> With those changes, the system seems to behave correctly.
>
> Best Regards
>
> Elise Duboué-Dijon
>
This isn't quite enough for parallel simulations. You also need to change
BC_PARMR to be the sum of all array lengths in the 'rparms' common block.
If you don't do this, then you will not properly broadcast all of the data
in parallel (you may or may not notice the effects, but there's also no
guarantee that it's 'correct').
Good luck,
Jason
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Nov 27 2012 - 05:30:02 PST
