Re-read Chapters 8 and 9. vglconnect is used only for the VGL Transport, which compresses the 3D images and forwards them to a remote X server. That isn't what you want, and in fact, by also specifying '-c proxy' as arguments to vglrun, you are bypassing the VGL Transport on the server end and instead causing VirtualGL to send the 3D images uncompressed to the remote X server, which definitely isn't what you want.
You simply need to use the same vglrun command line you're using below, but type it *after* you do 'export DISPLAY=localhost:1.0'. When using the X11 Transport (-c proxy), VirtualGL basically causes the 3D app to act like a 2D X11 app, so it uses DISPLAY the same way any X11 app would. The message about being unable to load libdlfaker.so and librrfaker.so is a problem, though. Make sure those libraries are installed in a system library path (for instance, /usr/lib) or that the path in which they're installed is in LD_LIBRARY_PATH. That message generally means that VirtualGL has not been loaded successfully, so it probably won't work until you resolve that issue. That message can also mean that you're running a 3D app that has setuid root permissions, in which case you need to read Chapter 13, but it would be odd for glxgears to be setuid root. On 11/29/12 12:19 AM, charles quarra wrote: > 2012/11/27 DRC <dcomman...@users.sourceforge.net>: >> On 11/27/12 2:25 PM, charles quarra wrote: >>> Because i'm thinking that you need turboVNC in order to composite the 3d >>> renderings inside the application windows? Are you saying that i can use >>> virtualGL to render remote apps (i.e: some 3d game) and then redirect the >>> output >>> to Xvfb without using turboVNC? >> >> Yes. >> >> > > This is great news. So i've already set up to run in the server with > ATI HD6800 hardware, and installed it in a few testing machines on the > LAN to behave as clients > > > i followed the documentation, first did vlgconnect to the server, then i did > vlgrun -c proxy -fps 10 -- VirtualGL/bin/glxgears > > this starts the window while connected over ssh to the server. My > question is, how do i direct this app inside the xvfb running on the > client machine? > > usually the process to run something inside Xvfb is like: > > 1) start Xvfb: > Xvfb :1 -screen 0 1024x768x24 -fbdir vfb_test/fbdir & > > 2) set default display > export DISPLAY=localhost:1.0 > > 3) start app > xterm -e '/usr/X11/bin/xcalc' & > > What would i need to change in how i start the glxgears sample so that > it will run inside Xvfb? > > > P.S: not related to this question, but vlgrun glxgears runs a bit > choppy on the client, and it reports high FPS (apparently ignoring the > -fps option) > > this is the output > > ERROR: ld.so: object 'libdlfaker.so' from LD_PRELOAD cannot be > preloaded: ignored. > ERROR: ld.so: object 'librrfaker.so' from LD_PRELOAD cannot be > preloaded: ignored. > Visual ID: 0xd3 > 18477 frames in 5.2 seconds = 3533.601 FPS > 3470 frames in 5.3 seconds = 652.749 FPS > > ------------------------------------------------------------------------------ > Keep yourself connected to Go Parallel: > VERIFY Test and improve your parallel project with help from experts > and peers. http://goparallel.sourceforge.net > _______________________________________________ > VirtualGL-Users mailing list > VirtualGL-Users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/virtualgl-users > ------------------------------------------------------------------------------ Keep yourself connected to Go Parallel: VERIFY Test and improve your parallel project with help from experts and peers. http://goparallel.sourceforge.net _______________________________________________ VirtualGL-Users mailing list VirtualGL-Users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/virtualgl-users
