Can an M2090 work as a GPU AND output OpenGL over X11? Want to do some light visualisation work too

[zebraC]

Hi all.

I’m about to build a small GPU cluster using some Dell C410x series technology and some PCI-E array interconnects.

It’s come time to work out what types of GPU’s to get.

Initially I thought about getting two M2070Q’s, on the premise that I wanted to do some light visualisation work with these GPU’s as well as pathological GPU floating point operation intensive applications.

My question is simple.

Do I need the M2070Q’s to get access to an OpenGL frame buffer and general openGL functionality? I.e - displaying images/video/rendering etc. All I want to do is run some OpenGL forwarded over X11. Problem is, I’d much rather use the M2090’s, with a far higher stream processor count! I’d hate to think I need the Quadro chip just for this!

Hope that is clear enough. It’s my hope that nVidia haven’t done anything untoward in limiting the ability of a pure Telsla GPU to display/render OpenGL to internal frame buffers, or output over X11 windows in Linux!

Thank you.

z

[laughingrice]

With linux you can get away with setting up X on a GPU that is not connected to a display, something that you can’t do with windows. So basically the answer is yes, you can do it, but it takes some work. Note that forwarding OpenGL over X11 has it’s issues as well. Also note that once you setup X on the GPU you are going to get a runtime limit on your kernels, don’t know if it’s an issue or not.