From: Axel Kohlmeyer (akohlmey_at_cmm.chem.upenn.edu)
Date: Wed Nov 21 2007 - 16:35:50 CST

On Tue, 20 Nov 2007, Daniel Beaman wrote:

hello,

DB> $ xhost +computername.edu

please note that with this setting, _anybody_
logged into "computername.edu" can connect
to your laptop and monitor _everything_ that
you are doing, including logging every single
keystroke (including those when you type in
a password).

DB> $ ssh -X -l name etc...

this enables X11 tunneling which would _not_
be needed with the xhost setting. so you can
skip the xhost for reasons outlined above.
(... and i better not get started on "xhost +",
which is still use by far too many people).

DB> I think these two lines allow me to have graphical displays piped to my
DB> computer. What I'd like to have is the VMD GUI piped ( not sure if this is
DB> proper terminology? ) to my Mac from the Linux Machine. As VMD starts up on
DB> the Linux machine, it opens a window on my powerbook and shows me technical
DB> data about the cluster, ie. #CPUs, etc.. but then spits out the following
DB> errors:
DB>
DB> the x server does not support the opengl glx extension
DB> unable to create openGL window

this means, that your X server does not include
the "OpenGL over X" protocol which is required to
display VMD graphics over the net.

DB> From multiple hours of searching the internet I have learned a
DB> little about openGL (I'm a newbie to this type of stuff) but have
DB> absolutely no idea how I could enable this or get openGL loaded
DB> properly on my computer. The main

please open a terminal window and run "xdpyinfo" that
should give you a list of supported extensions to the X
protocol. if GLX is not amongst them, you'll have to
a find a way to "load" or configure it. that is the
problem of the MacOS.

DB> reason I would like to use these programs and computers in the way I'm
DB> describing is that the files are prohibitively large to be transfered to my
DB> machine (I can get VMD to run my powerbook from the Mac version of VMD that
DB> I installed, but this still isn't much of an option) and the files live on
DB> the Linux cluster where there are TBs worth of storage space. I also don't
DB> have direct access to the cluster so I can't sit in front of it and view the
DB> stuff that way. So, is there anyone out there who has done what I am trying
DB> to do (I would imagine this is somewhat common, the only exception being the
DB> Mac part of it) or anyone who has any ideas on how to get this to work? I
DB> can get this to work on a Lab mates computer who is running Linux and it
DB> seems to work fine. I've also looked into using Virtualbox but I am unable
DB> to because I have a powerpc rather than an intel chip. Any other ideas?

the alternative would be to have something like a webdav server
running on the cluster and then mount the file system to your
mac and thus keep the data remotely and process it locally.
the drawback is that in this case you have to pull the data
all the time across the net. it is usually more efficient for
small data sets and visualization and less efficient for analysis,
but in the latter case you can run those in text mode. i'm doing
this quite often these days: copy a small segment of the data
to my machine, write/debug my scripts and then run them in text
mode on the machine with the full trajectory and load and process
the data in batches.

cheers,
   axel.

DB>
DB> Thank you very much in advance.
DB> Dan Beaman
DB> University of Oregon
DB>
DB>

-- 
=======================================================================
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.