NVIDIA X Info

[lewiscowles]

Hi, I have been reading your forum in search of answers, I know how to install nvidia in every linux distro that supports it, debian testing even built the kernel module from scratch I have a few questions

I have read that the nvidia driver has a number of connections within linux / *nix environments for example the kernel driver, the x driver, the apparently platform independent nvidia kernel that caused problems with being tied into the linux kernel with BSD (I Love that part)

My question is why is this three tiered interface implemented and does it mean that if I took the X bindings for opengl accel and applied them to a terminal program or GUI not using X I could have nvidia accelleration through that app instead.

basically the only part of X I want is the nvidia support I can't stand X but I'd like to do some dev work using OpenGL with my hardware acceleration and I'd like to know if the kernel module can do this without X or if the kernel module is effectively useless without X

Thanx

[AaronP]

No, OpenGL communicates with and relies on the X driver to manage the screen. X is pretty lightweight, though. If you just want to run an OpenGL app with minimal fuss, starting an X server and running a fullscreen OpenGL app as the only client is very easy.

The kernel module is not useless without X -- you can run CUDA applications on the GPU without X running.