[xplor-NIH] Adjustable timesteps in internal dynamics

[Mary Kirsten Frank]

---------- Forwarded message ----------
Date: Fri, 11 Oct 2002 11:23:06 -0400
From: Mary Kirsten Frank <kirsten at bach.cit.nih.gov>
To: mcmahon <mcmahon at ccnmr.mit.edu>

The internal dynamics has a special feature, the ability to adjust
timesteps to match the "graininess" of the potential surface. This is
actually a feature not a bug. The timestep you specify, timestep=0.002, is
the first timestep and then the algorithm adjusts the next one. If you
don't want this behavior, add "AdjustTS=false" into your internal dynamics
call. But this feature is fairly useful, so you may want to keep it. If
you find the timesteps getting quite small, like <1% of your original
setting, there is a problem with the dynamics, and it should be looked at.

On Thu, 10 Oct 2002, mcmahon wrote:

> Hi Kirsten,
> 
> I have another question, this one about the endtime command. How does this
> work exactly when running dynamics, it seems to produce a different number
> of steps for each change in temperature, and I didn't understand why this
> was the case. If i set the timestep to 0.002 and set the endtime to 1.612,
> shouldn't this
> produce the same amount of steps each time? I am using the
> domain2_tor_fin.inp example script.
> I am probably missing something simple, but hadn't seen this endtime used in
> CNS previously.
> 
> mike
> 
> 
> 
> 
>