Executing opengl applications remotely using vglrun

hi all,
I am attempting to run opengl applications remotely using turbo-vnc and virtualgl.
basically from a turbo-vnc session I execute the OF application as :
vglrun ./bin/<OFapp_name>

This works wonderfully for applications in OF_0.8.1. It does not work though in any application that I compile in OF_0.8.4 . even the same application if I carry it from OF .1 to .4 it will not work.

What happens is the application stalls. no window appears, nor any messages are printed in the command line.

In now very little of OpenGL internals, so reading through the change log did not help. What is different as far as opengl is concerned between the 2 OF versions??

thank you for your help.

hi synthnassizer… have you had any luck with this?

i think the problem here might be that glfw tries to use xwindows in a way that is buggy in some implementations, changing the window to glut or using the last glfw where there’s a workaround for this should work

not sure if this will work for your use case, but you should be able to get OF to start from the command line by specifying the DISPLAY variable to use before you run it.

ssh user@server
export DISPLAY=:0

you could wrap this all up in a shell script on the remote machine if you want as well.

1 Like

@TimS this is not same as using virtualGL.
If I am not mistaken, what you describe runs the executable on the remote machine and sends the final frames to the local machine. But I wonder, don’t you fail to start appplications when you run them over ssh? (lack of GL extentions on the ssh X tunnel and such ?)

VirtualGL intercepts GL commands to my local machine and lets the GL rendering take place here. This allows for higher fps.

@jmarsico no lack. I have not spend time to re-try this. There seems to be a new turbo-vnc version and perhaps virtualGL version though.

If you do try please let us know how it goes . Come to think of it, i’ll try to give it a try myself soon (hopefully during his weekend).

1 Like

If you build GLFW yourself, are you able to run the GLFW sample apps run via TurboVNC / VirtualGL?