From: John Stone (
Date: Thu Sep 14 2017 - 21:35:59 CDT

The latest version of the VMD tutorials has elementary examples
of how to use the "parallel xxx" commands to schedule analysis
scripts across large numbers of compute nodes. These examples are
at the very end of the tutorial:

By far the biggest hurdle is usually the task of getting VMD compiled
with MPI support enabled on whatever cluster or supercomputer you
are planning to use. We make VMD binaries available for several
of the national supercomputers in the U.S., but if you're running
on a local cluster then at present you'll need to compile VMD yourself
for the particular version of MPI you're using.

Axel Kohlmeyer also posted just the other week about another MPI
interface that he has made that can be loaded as a plugin into any
Tcl interpreter (including VMD of course), so you may want to
have a look at that also.

Best regards,
  John Stone

On Fri, Sep 15, 2017 at 11:20:43AM +1000, Monika Madhavi wrote:
> Hi all,
> I used to write sequential tcl codes to analyze my MD trajectories with
> VMD. However, I would like to speed up my code and need to make it able to
> run in parallel. I am totally new to parallel computing and therefore I
> would like to have a very simple guide on how to do a small analysis in
> parallel.
> For example, if I want to calculate the distance between (say) two atoms
> in every frame and write it into an output file, is it possible for vmd to
> process few frames at one time and wait till all frames are finished
> before writing the output? What would be the format of a sample tcl code
> for such a task?
> Thank you.
> Best regards,
> Monika
> --
> W.A.Monika Madhavi
> Lecturer (Probation),
> Department of Physics,
> University of Colombo.

NIH Center for Macromolecular Modeling and Bioinformatics
Beckman Institute for Advanced Science and Technology
University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801           Phone: 217-244-3349