On Tue, Apr 29, 2014, Yin, Guowei wrote:
>
> During my MD runs, I got the situation that "vlimit exceeded for step
> ******; vmax = 20.4201". Once this appeared, the run was stopped. It
> happened normally after the simulation started about a few ns or a dozen
> ns later.
Yikes...look like we need to change to format of the warning message, now
that numbers of steps are larger than in the past. But a vlimit problem
itself does not stop a run...something else bad must have happened.
>
> I may use smaller step length to circumvent it. But to better assess the
> reason causing this "vlimit exceeding", I hope to extract the velocity
> from the restart file to see which atoms from which residues are with
> the higher velocity. Does anyone know how to do extract the velocity
> from the restart file and assign the velocity to each atom? Is there any
> commands regarding this?
Depends on the sort of restart file you have: if you used the default ntxo=1,
it's easy to find the velocities with a text editor; see the file formats
link on the Amber web site for details on what is in the file. But knowing
which atoms are involved may not be very helpful: vlimit problems are often
not very closely related to particular problems.
> In general, is there any suggestion to solve the "vlimit" problem?
No general rules, since it depends on the size of the system, whether or not
you are using a barostat or a thermostat (and with what parameters),
the step size, how well equilibrated your system is, whether you have
restraints or other artificial components, etc. Generally, longer values
of tautp and taup help, and (as you noted) shorter time steps help. But you
may have to do some experimentation.
...dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Apr 29 2014 - 18:30:03 PDT
