VMD-L Mailing List

From: Axel Kohlmeyer (akohlmey_at_cmm.chem.upenn.edu)
Date: Mon Jul 21 2008 - 14:13:59 CDT

On Mon, 21 Jul 2008, Samir Unni wrote:

SU> Hi,
SU>
SU> That could prove problematic. Using the 'snapshot' renderer requires only
SU> about 10 seconds on this machine (octo-core 2.8 GHz Intel Xeon Mac Pro) with
SU> the command I mentioned earlier (using the modified vmdmovie.tcl). But the
SU> the 'tachyon' renderer requires upwards of 7 minutes. Is there no other
SU> renderer option that could be used without X/OpenGL?

there are several options and there are also ways to speed
up rendering at the expense of some detail or features.

in case you have a cluster of machines, you could consider
just rendering parts of the movie (while keeping the frames)
on different nodes and/or run the renderer in parallel (several,
including tachyon support this with good parallel efficiency)
and then build the movie afterwards.

i'm pretty certain that john has some good suggestions as well
(this is his special area of expertise. ;-) ).

cheers,
   axel.

SU>
SU> Thanks,
SU>
SU> Samir
SU>
SU>
SU> On Mon, Jul 21, 2008 at 11:44 AM, John Stone <johns_at_ks.uiuc.edu> wrote:
SU>
SU> >
SU> > Hi,
SU> > Ok, that's good to know. Yes, it should be possible to use the makemovie
SU> > method in pure text mode, but some state variables will have to be changed
SU> > and you won't be able to use the "snapshot" renderer since there's no
SU> > OpenGL window to take snapshots from. You'll probably want to try using
SU> > tachyon, I'll send you the necessary script tweaks to get it going for you.
SU> > I'll also be improving the text mode movie generator routine
SU> > for the next release. :)
SU> >
SU> > Cheers,
SU> > John
SU> >
SU> > On Mon, Jul 21, 2008 at 11:35:51AM -0500, Samir Unni wrote:
SU> > > Hi,
SU> > >
SU> > > Yes, that fixed the problem! Is the state variable that's missing one of
SU> > the
SU> > > two (or both) that I initialized in my modified version of vmdmovie.tcl?
SU> > >
SU> > > Also, even if this problem is fixed, it seems that the 'makemovie' method
SU> > > cannot be used unless a GUI is running (i.e. starting vmd with '-dispdev
SU> > > none' or '-dispdev text' won't work). That isn't an option in my case,
SU> > since
SU> > > I will be using the state files I'm creating with instances of VMD that
SU> > are
SU> > > on servers where X is not present. Is there a scriptable workaround for
SU> > this
SU> > > issue as well?
SU> > >
SU> > > Thanks,
SU> > >
SU> > > Samir
SU> > >
SU> > >
SU> > > On Mon, Jul 21, 2008 at 10:31 AM, John Stone <johns_at_ks.uiuc.edu> wrote:
SU> > >
SU> > > > I think I've reproduced the problem you encountered, try doing this
SU> > > > and see if you get different behavior:
SU> > > >
SU> > > > Run the command line script as before, and verify that you get
SU> > > > the 1 frame movie, as before.
SU> > > >
SU> > > > Then, after having done this, try opening the GUI for the movie
SU> > > > maker, through the Extensions->Visualization->Movie Maker menu.
SU> > > > Once the window is opened, try running the script at the command
SU> > > > line again. Let me know if it works correctly for you the second
SU> > > > time. I believe that the issue here is an internal state variable
SU> > > > that's only being properly initialized when the GUI has been opened.
SU> > > > It will be easy to workaround/fix with one or two extra lines of
SU> > > > scripting if that's all that's causing your problem.
SU> > > >
SU> > > > Cheers,
SU> > > > John Stone
SU> > > > vmd_at_ks.uiuc.edu
SU> > > >
SU> > > >
SU> > > > On Mon, Jul 21, 2008 at 07:57:59AM -0500, Samir Unni wrote:
SU> > > > > Hi,
SU> > > > >
SU> > > > > Sorry I'm so late in getting back to you. Anyway, that code doesn't
SU> > work
SU> > > > on
SU> > > > > the Mac either. It just briefly opens the video (for one frame, I
SU> > think)
SU> > > > and
SU> > > > > then closes. I'm using the 1.8.6 version of VMD?is that the one
SU> > you're
SU> > > > > using?
SU> > > > >
SU> > > > > Thanks,
SU> > > > >
SU> > > > > Samir
SU> > > > >
SU> > > > >
SU> > > > > On Thu, Jul 17, 2008 at 10:19 PM, Samir Unni <srunni_at_gmail.com>
SU> > wrote:
SU> > > > >
SU> > > > > > Hi,
SU> > > > > >
SU> > > > > > I tried that, but it didn't work for me. When I attempted to play
SU> > the
SU> > > > video
SU> > > > > > file with MPlayer, the window appeared for a second and then
SU> > > > disappeared,
SU> > > > > > just like before. I think only one frame is being recorded.
SU> > > > > >
SU> > > > > > Right now I'm trying this on my Linux machine at home, but I'll try
SU> > it
SU> > > > > > again on the Mac at work tomorrow morning and see if it works
SU> > there.
SU> > > > > > However, this was the exact same problem I was previously having on
SU> > the
SU> > > > Mac,
SU> > > > > > so I don't think it will. If it's working for you, it might have
SU> > > > something
SU> > > > > > to do with the state file I'm using, which I created by importing
SU> > the
SU> > > > DX
SU> > > > > > output file from APBS into VMD.
SU> > > > > >
SU> > > > > > I had meant to attached the modified version of vmdmovie.tcl to the
SU> > > > last
SU> > > > > > email; it's attached now, if you want to take a look.
SU> > > > > >
SU> > > > > > Thanks,
SU> > > > > >
SU> > > > > > Samir
SU> > > > > >
SU> > > > > >
SU> > > > > >
SU> > > > > > On Thu, Jul 17, 2008 at 9:56 PM, John Stone <johns_at_ks.uiuc.edu>
SU> > wrote:
SU> > > > > >
SU> > > > > >>
SU> > > > > >> Hi,
SU> > > > > >> I'm not sure how you changed your copy of the vmdmovie.tcl file,
SU> > > > > >> but I'd ask that before we delve any further, that you first try
SU> > the
SU> > > > > >> original vmdmovie plugin and the command I emailed earlier. If
SU> > that
SU> > > > > >> works for you, then we can proceed with debugging whatever has
SU> > been
SU> > > > > >> changed in your modified version. Starting with your modified
SU> > version
SU> > > > > >> is likely to add more confusion so I'd rather leave that for a
SU> > second
SU> > > > > >> step. Please try the example sequence I sent earlier and let me
SU> > know
SU> > > > > >> if that works with the unmodified script on your system, or not.
SU> > > > > >>
SU> > > > > >> Cheers,
SU> > > > > >> John Stone
SU> > > > > >> vmd_at_ks.uiuc.edu
SU> > > > > >>
SU> > > > > >> On Thu, Jul 17, 2008 at 09:18:24PM -0500, Samir Unni wrote:
SU> > > > > >> > Hi,
SU> > > > > >> >
SU> > > > > >> > That character resulted from an issue with the email formatting;
SU> > it
SU> > > > was
SU> > > > > >> not
SU> > > > > >> > actually part of the command that I used. The command I used
SU> > was:
SU> > > > > >> >
SU> > > > > >> > ::MovieMaker::makemovie /Users/srunni/dev/vmd_tests/1fas 1fas
SU> > > > ppmtompeg
SU> > > > > >> > snapshot 10
SU> > > > > >> >
SU> > > > > >> > Also, I sent an update a few hours after my initial email, but
SU> > it
SU> > > > seems
SU> > > > > >> not
SU> > > > > >> > to have made it onto the mailing list. I will repost it here:
SU> > > > > >> >
SU> > > > > >> > --------------------
SU> > > > > >> > I've got a small update. I slightly changed vmdmovie.tcl, so
SU> > that
SU> > > > the
SU> > > > > >> > framerate can be passed to the 'makemovie' method. When I did
SU> > this,
SU> > > > > >> defined
SU> > > > > >> > 'framerate' as the extra variable being passed now, and defined
SU> > > > > >> 'numframes'
SU> > > > > >> > as the product of the framerate and the duration, a working
SU> > movie
SU> > > > was
SU> > > > > >> > created. Though it would seem that only 2 of these 3 variables
SU> > need
SU> > > > to
SU> > > > > >> be
SU> > > > > >> > defined for the program to be able to determine the 3rd, when I
SU> > > > > >> eliminated
SU> > > > > >> > any one of them, the movie was not generated correctly. I have
SU> > > > attached
SU> > > > > >> the
SU> > > > > >> > working version of vmdmovie.tcl, if anyone wants to look at it.
SU> > > > > >> >
SU> > > > > >> > I would still like to know if there is a way to generate a movie
SU> > > > from
SU> > > > > >> the
SU> > > > > >> > command line without modifying any files that are a part of VMD,
SU> > if
SU> > > > > >> > possible.
SU> > > > > >> > --------------------
SU> > > > > >> >
SU> > > > > >> > I still have not found a workaround for this issue. The command
SU> > I am
SU> > > > > >> using
SU> > > > > >> > now is:
SU> > > > > >> >
SU> > > > > >> > ::MovieMaker::makemovie /Users/srunni/dev/vmd_tests/1fas 1fas
SU> > > > ppmtompeg
SU> > > > > >> > tachyon 5 40
SU> > > > > >> >
SU> > > > > >> > Thanks,
SU> > > > > >> >
SU> > > > > >> > Samir Unni
SU> > > > > >> >
SU> > > > > >> >
SU> > > > > >> > On Thu, Jul 17, 2008 at 5:02 PM, John Stone <johns_at_ks.uiuc.edu>
SU> > > > wrote:
SU> > > > > >> >
SU> > > > > >> > >
SU> > > > > >> > > Hi,
SU> > > > > >> > > The ">" character you have in your example below shouldn't be
SU> > > > there,
SU> > > > > >> > > that would cause some trouble for sure.
SU> > > > > >> > >
SU> > > > > >> > > Here's a proper example command sequence to make a movie from
SU> > text
SU> > > > > >> mode:
SU> > > > > >> > > package require vmdmovie
SU> > > > > >> > > mkdir /tmp/mytmpdir
SU> > > > > >> > > ::MovieMaker::makemovie /tmp/mytmpdir example ppmtompeg
SU> > snapshot
SU> > > > 10
SU> > > > > >> > >
SU> > > > > >> > > This worked fine for me.
SU> > > > > >> > >
SU> > > > > >> > > Try that and let me know how it goes.
SU> > > > > >> > >
SU> > > > > >> > > Cheers,
SU> > > > > >> > > John Stone
SU> > > > > >> > > vmd_at_ks.uiuc.edu
SU> > > > > >> > >
SU> > > > > >> > > On Wed, Jul 16, 2008 at 09:32:37AM -0500, Samir Unni wrote:
SU> > > > > >> > > > Hi,
SU> > > > > >> > > >
SU> > > > > >> > > > I'm trying to create a Tcl script that will use output from
SU> > APBS
SU> > > > to
SU> > > > > >> > > create a
SU> > > > > >> > > > video file, with no user interaction. So far, I have managed
SU> > to
SU> > > > > >> manually
SU> > > > > >> > > > create a state file, to which I can then add a few lines of
SU> > code
SU> > > > for
SU> > > > > >> the
SU> > > > > >> > > > Movie Plugin, to automatically generate the video file and
SU> > then
SU> > > > exit
SU> > > > > >> VMD.
SU> > > > > >> > > > However, I can't seem to get VMD to properly generate the
SU> > video
SU> > > > file
SU> > > > > >> when
SU> > > > > >> > > I
SU> > > > > >> > > > use the script. It creates a file, but it doesn't seem to
SU> > > > contain
SU> > > > > >> any
SU> > > > > >> > > actual
SU> > > > > >> > > > data (it's only about 20 KB), and doesn't play correctly
SU> > when I
SU> > > > open
SU> > > > > >> it
SU> > > > > >> > > with
SU> > > > > >> > > > MPlayer. This is the line of code that seems to be the
SU> > problem:
SU> > > > > >> > > >
SU> > > > > >> > > > ::MovieMaker::makemovie /Users/srunni/dev/vmd_tests/1fas
SU> > 1fas
SU> > > > > >> ppmtompeg
SU> > > > > >> > > > > snapshot 10
SU> > > > > >> > > >
SU> > > > > >> > > >
SU> > > > > >> > > > When I load this exact same state file, but with that line
SU> > > > removed,
SU> > > > > >> and
SU> > > > > >> > > > manually use the Movie Plugin, there are no problems. I
SU> > figured
SU> > > > out
SU> > > > > >> the
SU> > > > > >> > > > syntax for the above code by taking a look at vmdmovie.tcl,
SU> > and
SU> > > > > >> imitating
SU> > > > > >> > > > the calls made when the Movie Plugin is launched from the
SU> > GUI.
SU> > > > As
SU> > > > > >> far as
SU> > > > > >> > > I
SU> > > > > >> > > > can tell, I have made no mistakes in the syntax. Does anyone
SU> > > > know
SU> > > > > >> why
SU> > > > > >> > > this
SU> > > > > >> > > > isn't working correctly?
SU> > > > > >> > > >
SU> > > > > >> > > > Thanks,
SU> > > > > >> > > >
SU> > > > > >> > > > Samir Unni
SU> > > > > >> > >
SU> > > > > >> > > --
SU> > > > > >> > > NIH Resource for Macromolecular Modeling and Bioinformatics
SU> > > > > >> > > Beckman Institute for Advanced Science and Technology
SU> > > > > >> > > University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
SU> > > > > >> > > Email: johns_at_ks.uiuc.edu Phone: 217-244-3349
SU> > > > > >> > > WWW: http://www.ks.uiuc.edu/~johns/>
SU> > <http://www.ks.uiuc.edu/%7Ejohns/>
SU> > > > <http://www.ks.uiuc.edu/%7Ejohns/><
SU> > > > > >> http://www.ks.uiuc.edu/%7Ejohns/>
SU> > > > > >> > > Fax: 217-244-6078
SU> > > > > >> > >
SU> > > > > >>
SU> > > > > >> --
SU> > > > > >> NIH Resource for Macromolecular Modeling and Bioinformatics
SU> > > > > >> Beckman Institute for Advanced Science and Technology
SU> > > > > >> University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
SU> > > > > >> Email: johns_at_ks.uiuc.edu Phone: 217-244-3349
SU> > > > > >> WWW: http://www.ks.uiuc.edu/~johns/>
SU> > <http://www.ks.uiuc.edu/%7Ejohns/><
SU> > > > http://www.ks.uiuc.edu/%7Ejohns/>
SU> > > > > >> Fax: 217-244-6078
SU> > > > > >>
SU> > > > > >
SU> > > > > >
SU> > > >
SU> > > > --
SU> > > > NIH Resource for Macromolecular Modeling and Bioinformatics
SU> > > > Beckman Institute for Advanced Science and Technology
SU> > > > University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
SU> > > > Email: johns_at_ks.uiuc.edu Phone: 217-244-3349
SU> > > > WWW: http://www.ks.uiuc.edu/~johns/><
SU> > http://www.ks.uiuc.edu/%7Ejohns/>
SU> > > > Fax: 217-244-6078
SU> > > >
SU> >
SU> > --
SU> > NIH Resource for Macromolecular Modeling and Bioinformatics
SU> > Beckman Institute for Advanced Science and Technology
SU> > University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
SU> > Email: johns_at_ks.uiuc.edu Phone: 217-244-3349
SU> > WWW: http://www.ks.uiuc.edu/~johns/ <http://www.ks.uiuc.edu/%7Ejohns/>
SU> > Fax: 217-244-6078
SU> >
SU> 

-- 
=======================================================================
Axel Kohlmeyer   akohlmey_at_cmm.chem.upenn.edu   http://www.cmm.upenn.edu
   Center for Molecular Modeling   --   University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.