Dear abhilash,
Accelerated MD is not magically giving you more simulation steps in the
same time. The point is that the potential energy surface (and thus the
free energy surface) is perturbed such that - with fewer simulation steps
- you get conformational sampling that should be representative of much
longer timescales.
Please see the manual and the references therein.
Your performance seems to be as expected.
HTH,
Marc
On 11 Sep 2016 08:36, "Abhilash J" <md.scfbio.gmail.com> wrote:
> Hi!
>
> I am trying to preform aMD on a system.
> I was getting 8.06 ns/day using classical MD. When i tried using aMD,
> the speed dropped to 6.38 ns/day.
> My system has 238234 atoms (i.e. ~.2 million).
> I am using TIP3P water in a box with periodic boundaries.
> Protein being simulated is a dimer.
> I am using a K20 GPU card for this simulation.
> I am unable to understand why this speed drop happened when it should
> accelerate and what can be done to fix this.
> Please advice.
>
>
> Regards
>
> Abhilash
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Sep 11 2016 - 02:00:02 PDT
