On Fri, 2014-05-09 at 19:49 +0530, Sourav Purohit wrote:
> Even After 50 ns long NPT simulation I saw that the pressure values are all
> negative and fluctuating about -500. The system consists of a large number
> of DNA fragments initially arranged in a crystal with a water buffer of 30
> Ang.. I wanted to see the type of LC phase the system attains.
> The density also attained a large value.
> I had heard from my colleagues about similar problems with multinode
> multigpu NPT simulations. I just wanted to make sure that all the results
> that I get are right. Do you feel that the observations of a high density
> and negative pressure even after a long run are okay for the system I am
> simulating?? Kindly help.
I have no idea. I haven't run a multi-node GPU NPT simulation before.
Try doing a density equilibration on the CPU (or a single GPU) and see
if the results you get there match the results you get with the
multi-node GPU. (It should not take 50 ns to equilibrate your density)
BTW, you also have to tell us what version of Amber you are using for
the best help, as well as which updates have been applied. This is
always a good idea. (And make sure that all updates have been applied;
although currently this only applies to Amber 12).
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri May 09 2014 - 09:00:04 PDT
