On Tue, Feb 17, 2015, BERGY wrote:
> I used CPU pmemd and had no error msg.
Make sure that the cpu output looks like a stable trajectory, with reasonably
small energy changes.
> I also note that if ntc=2,ntf=2 I dont get any errors.
> But this was done without pmemd.cuda and not pmemd.cuda.MPI.
Not sure if you meant "with" (instead of "without") in the above line?
The business of ntf=1 vs. ntf=2 has come up before, but I can't remember the
details. Maybe someone else on the list can help.
You will probably have to make the actual files available so that someone
else can reproduce the problem. Or, run for a while with pmemd.cuda, and
see if changing later to pmemd.cuda.MPI will work.
....dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Feb 19 2015 - 05:30:04 PST
