On Tue, May 14, 2013 at 8:04 AM, Ross Walker <ross.rosswalker.co.uk> wrote:
> But for pmemd.cuda it should be good I believe? no?
>
Actually, it will work for sander, pmemd, pmemd.cuda, sander.MPI*, and
_maybe_ pmemd.cuda.MPI (testing or Scott's comment needed to confirm). It
will not work for pmemd.MPI.
-It- in this case being reproducible initial velocities for irest=0 and
ig=-1 if the assigned seed is known. Once you use a stochastic thermostat,
you lose sander.MPI's reproducibility as well (hence the *).
In any case, it's safer to simply assume that ig=-1 implies no
reproducibility.
All the best,
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 May 14 2013 - 05:30:05 PDT
