I'm trying to get some OpenGL program to render on an X server on the network (actually it's from a VM) that has Nvidia hardware. So I configure X to allow clients from some IP, and on the client set the address of the X server:
X programs work properly, however I have problems with OpenGL. If I use the Mesa library (7.1) on the client, only OpenGL version 1.4 is reported by glxinfo and all version 2 extensions are missing (but it works). If I use nvidia's libGL on the client, version 2.1 is reported and some more extensions are available, but most are still missing, and Quake 3 renders some things black (while it worked with Mesa).
As far as I know Nvidia drivers implement AIGLX which allows for hardware-accelerated rendering using the GLX protocol. But why is half of the interface missing? Is it something with my configuration? Or is it really unimplemented?