I’m rendering 4 different windows to 4 different displays.
My setup is this:
MacPro 2 x 2.8Ghz quad core
2 x Nvidia 8800 512mb graphics cards
Running 4 LG monitors at 1920x 1080
The computer boots up like this:
Card 1 – Display 1 and Display 4
Card 2 – Display 2 and Display 3
If I render all 4 jit.windows on Display 1, my framerate is good. 1 jit.window on Display 1 and 1 jit.window on Display 4 is ok. BUT as soon as I render jit.windows on Display 2 and 3 (card in slot 2), then framerate goes down to 12fps. Even when I’m not sending any video or images to the windows (just rendering blue). So, I guess it’s something to do with how I’m using the second graphics card.
All windows are rendered by separate jit.gl.render objects. But I guess I’m just doing this wrong.
Hey – thanks for the response. I figured it out thanks to some help from Anton. Yes, you’re right, the initial position of the windows was the problem. They need to be positioned on the correct windows for each GPU BEFORE rendering is started.
There might have been something going on with sharing video files across 2 GPU’s as well, but I’m really not sure right now. Anyway, I’ve gone from 12fps to 30/40 fps depending on activity. A mixture of slide show and video at 1080p (prores codec).
There might have been something going on with sharing video files across 2 GPU’s as well
you mean using the "shared_context" on jit.gl.render?
it’s only possible to share textures between two jit.gl.renders which are on the same gpu
so in your case display 1 & 4 and display 2 & 3
not display 1 & 3 etc…