Two GLGraphics Windows + Two Video Cards? Hardware requirement/optimization question.
in
Integration and Hardware
•
1 year ago
I am working on a project that will be using GLGraphics as the renderer, and utilizes two windows for output. One window is a small control panel window, and the second uses a GLTextureWindow to display everything rendered to a
GLGraphicsOffScreen object. The GLTextureWindow will span 8 displays (anywhere between 1280x720 each to 1920x1080 each), giving us a total resolution of the
GLGraphicsOffScreen canvas from 10240x720 to 15360x1080.
We plan to get a Radeon Eyefinity 6 to give us 6 outputs, and then use a Matrox TripleHead2Go on the 6th output to give us a total of 8. We also want to have the control panel window on a 9th display that will be installed with the computer for the exhibition, and allows us some control over parameters.
My concern is regarding the sharing of an OpenGL application across GPUs. Our options are (1) buy a secondary cheap card that will serve as the 9th output for the control panel, or (2) use a DoubleHead2Go on another one of the outputs to keep the rendering all on a single video card.
When I have a single window with an OpenGL context in it, and I run it on my workstation with 2 video cards, and I make it span two displays (each on their own card) I end up with very choppy results (which is to be expected). However, I have never created two windows before, one being the primary GLGraphics rendering and the other being the GLTextureWindow. Ultimately, my question is this: will rendering the two different windows on two different video cards result in the choppiness due to GPU sharing, or will they behave in a manner where one window uses one video card and the other window uses the other video card?
1