On Sat, Nov 8, 2014 at 11:45 AM, Ruchika Bhat <ruchikabhat31.gmail.com>
wrote:
> Hi everyone,
>
> For running MD simulations for 10 ns the changes in the parameters of
> heat.in and prod.in file are in the nstlim, dt and ntwx.
>
> If I change the parameters to
> nstlim=100000,
> dt=0.1,
>
This is a 100 fs time step. It is way too long. Use something like 0.001
or 0.002 assuming you SHAKE bonds including hydrogen.
ntpr=100,
> ntwx=100, in the prod.in file what change would be required in heat.in
> file's parameter as there needs to be consistency of running *prod* after
> the heat time is over.
>
I don't know what you are asking. Assuming the time step you choose is
not so long that it introduces integration errors (0.1 is certainly too
long), then how frequently you print out information has no bearing on the
actual simulation.
The only thing you need to make sure you keep "consistent" between the
heating and production phases is the potential energy function (you can
remove certain potential energy terms like positional restraints when you
go to do production simulations, but things like the cutoff, GB model,
nonbonded method, etc. need to be kept the same).
HTH,
Jason
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sat Nov 08 2014 - 09:30:02 PST
