On Sat, Apr 25, 2015, Soumendranath Bhakat wrote:
>
> Just a small query. I am planning to run two independent production run
> from a equlibrated system (equi.rst file).
>
> I am setting up md.in with keywords irest=0 and ig=-1. Am I doing right? I
> believe setting up irest=0 with ig==1 will start an independent simulation
> from the last restart file (equil.rst).
Probably the best is to set irest=0, tempi=300. (or whatever temperature you
want), and ig=-1. This will randomize the velocities at the beginning of the
run, and using ig=-1 will ensure that you have different random number
schemes.
>
> Or if I only add irest=1 and ig=-1 then the ig=-1 will add random seed to
> each simulation which will make them independent.
This will also work: the random nature of the thermostat will quickly make the
two trajectories independent. But it's easier to explain if you also
completely randomize the velocities, as suggested in my first paragraph.
...dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sat Apr 25 2015 - 07:30:02 PDT
