VirtualGL initial setup (Fedora 30 server, Mac OS 10.14.6 host)

[Eric Sokolowsky]

I'm working my way through installing and configuring VirtualGL, to see if it's going to work for my OpenGL applications. I develop 3D applications in C++ (using OpenSceneGraph, which uses OpenGL for rendering), and my primary workstation at work is a Linux machine running Fedora 30. I logged in to the x11 session at the console before I started working at home full-time and it's important to me that I don't log it out or disturb the active X11 session if at all possible, since it will be a while before I can get back in to log in again if it's interrupted. I will be the only user of the machine. The machine has an nvidia Quadro graphics card and I know that 3D acceleration is working with the nvidia proprietary driver. I've verified that I am not running Wayland on the X11 session.

Is it still necessary to run "vglserver_config" in my case? Is there anything that I can do manually to add permissions or configure something to work for my case? I'm afraid that running this script will not work because of the active X11 session, or disturb said session.

Thanks for any insight you can provide.

[DRC]

No, if the "3D X server" (the X server on the server machine that is
connected to the GPU) will remain running under your user account, then
you don't need to use vglserver_config at all. The purpose of
vglserver_config is to grant access to the 3D X server for multiple
users while the server machine is sitting at the login prompt.

[Eric Sokolowsky]

On Thursday, May 14, 2020 at 6:28:15 PM UTC-4, DRC wrote:

No, if the "3D X server" (the X server on the server machine that is
connected to the GPU) will remain running under your user account, then
you don't need to use vglserver_config at all.  The purpose of
vglserver_config is to grant access to the 3D X server for multiple
users while the server machine is sitting at the login prompt.

This is great. Thanks!

[Eric Sokolowsky]

I now have VirtualGL running inside a TurboVNC session, running Gnome desktop. I am a bit confused by something. I have my 3D (hardware-accelerated) X server running on :1, and my TurboVNC server on :2. I'm running the TurboVNC session through ssh using port forwarding. To get an idea of performance I ran glxspheres64, and I got the following results:

> vglrun -d :1 glxspheres64
Polygons in scene: 62464 (61 spheres * 1024 polys/spheres)
Visual ID of window: 0x198
Context is Direct
OpenGL Renderer: Quadro K5200/PCIe/SSE2
1.000212 frames/sec - 0.721049 Mpixels/sec
0.999800 frames/sec - 0.720752 Mpixels/sec

> vglrun -d :2 glxspheres64
Polygons in scene: 62464 (61 spheres * 1024 polys/spheres)
Visual ID of window: 0x198
Context is Direct
OpenGL Renderer: llvmpipe (LLVM 8.0, 256 bits)
30.895017 frames/sec - 22.272094 Mpixels/sec

So I would have expected the first instance to be much faster, since it's using the 3D X11 server for rendering. Unless I'm misunderstanding what's going on. Any enlightenment will be appreciated! I also welcome suggestions on improving performance if that's possible.

[DRC]

I suspect you're encountering a known issue that was introduced with nVidia 440.xx.  Please see
https://github.com/VirtualGL/virtualgl/issues/120 for both a temporary and permanent workaround.

--
You received this message because you are subscribed to the Google Groups "VirtualGL User Discussion/Support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to virtualgl-use...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/virtualgl-users/ddc08704-b4fc-47c0-93b5-6baec66c5217%40googlegroups.com.