From: Axel Kohlmeyer (akohlmey_at_gmail.com)
Date: Wed Jan 12 2011 - 11:40:12 CST

dudo,

a few comments below.

On Wed, Jan 12, 2011 at 7:41 AM, Dudo <dudomail_at_gmail.com> wrote:
> Hi,
> I met the same problems with visualization of VTF trajectories as some folks
> some time ago (seek in the thread).
> I don't know how those guys solved the problem, or how they do visualize
> their trajectories.
> The thing is, that the VTF pluggin in VMD is most likely malfunctional. It
> doesn't work with the default inputs from
> their webpage.

rather than complaining in this very unspecific way,
can you provide examples or pointers to examples
that "don't work" and then explain what does not
work how, and what you would have expected instead.
this way, somebody else has a chance to reproduce
the situation and come up with suggestions.
when you refer to a thread on the mailing list, please
provide an URL.
also, it is of vital importance to know which version
of VMD you are using and on what platform.
in general, the easier it is for somebody else to
reproduce your problem, the easier you will be helped.

furthermore, have you considered contacting the author
of the VTF plugin, olaf lenz, about the problems. he would
certainly be _most_ interested to know of problems with
it and try to resolve them.

[...]

> at this point VMD looks to me however as more straightforward tool, where
> you can get your visualization in the fastest way.
> So, it appeared to me as a simplest possibility to convert from VTF to other
> format, particularly LAMMPSTRJ.
> For those, who find this a convenient option, I'm providing the source code
> in the attachment.
> If you compute trajectories on a cluster and then download the trajectory to
> visualize how the structure evolves,
> as I do, you can convert the trajectory on that machine.
> Hence, in Linux you compile the code:
> ifort vtf2lammps.f90
> next, rename compilation output file
> cp a.out vtf2lammps
> make it executable

how complicated!

ever head of using the -o flag of the compiler?
or using "mv" to rename files?

> chmod +x vt*
> You may run it by typing
> ./vtf2lammps
> then the program asks for an input file name. It changes its extension to
> .lammpstrj so you don't provide a name for an output file.
> you may use it also directly as a command with argument
> ./vtf2lammps [filename]

> in the VMD the grains will appear as tiny dots, I don't know why. It works

that usually happens, when VMD has insufficient information
about the system and thus cannot make any guesses to
assign elements and bonds. the LAMMPS trajectory file
by default contains limited information.

> like this also for the default inputs from the pluggin website.
> hence, you just click on the VMD main window, Graphics tab, from the top
> menu, Representations,
> choose CPK and increase the size of the spheres as you wish.

a better way would be to provide a "topology" file that assigns
elements and bonds, or running some tcl scripting that will assign
atom radii, names, masses and more.

> anyways, for a quick look you don't have to do that. it's sufficient to see,
> that polymer is a polymer and e.g. doesn't fall apart.

> one more thing, i was not sure about meaning of required inputs for PBC
> BOUNDS in the lammps format.

this is discussed in the lammps documentation of the dump command. RTFM!

axel.

> so if you have to correct it, it can be easily done in the source file, and
> then please let me know too :-))
> bye,
> Dudo Racko
> --
> ____________________
> Ing. Dusan Racko, PhD
> Polymer Institute of the Slovak Academy of Sciences
> Dubravska cesta 3
> 845 41 Bratislava, Slovak Republic
> tel: +421 2 5477 3448
>

-- 
Dr. Axel Kohlmeyer
akohlmey_at_gmail.com  http://goo.gl/1wk0
Institute for Computational Molecular Science
Temple University, Philadelphia PA, USA.