sorry for the double post, but this might be a more appropriate forum to ask the question:
hardware: Asus P8-Z77, I7, 3xGTXx870, monitor connected to intel igpu, nothing plugged into the nvidias.
While improving/automating my CUDA installation process and trying out various options (runtime, deb, stripped xubuntu, ubuntu+xorg+xfce, etc), every once in a while, I was able to run the cuda OpenGL examples, but I was never able to duplicate the process. Originally I thought the problem was due to misconfigured/overwritten opengl lib, and followed the recommendations found on the web.
In the end I was able to get nvidia and cuda installed, not screw up opengl on the igpu and able to run glxgears succesfully, but cuda OpenGL examples would not run. All other cuda examples worked as expected. And it bugged me, because every once in a while the opengl example would run but I could never duplicate the installation process.
Now, after 300+ installation attempts I’m finally able to duplicate the process of installing Xubuntu 14.04, nvidia drivers, cuda, (from the nvidia’s deb packages) and have the OpenGL cuda examples run while the monitor is plugged into the motherboard’s integrated intel card. No special sauce, xubuntu straight out of the box, and cuda from nvidia’s repositories! Didn’t even touch xorg.conf
So now the (revised) question: should the OpenGL cuda examples run on intel’s igpu? And if they are not supposed to run on nothing but nvidia’s card (as it seems to be the common agreement) why are they running on the igpu? Something in the OpenGL specs?
thanks in advance