Specifying which GPU to process on at run time in multi-GPU systems
Hello, Is there a way, at compile time or run time, to pick a specific GPU / video card to run the processing on when multiple video cards are installed? We have a workstation with one low end Nvidia k620, which the primary monitor is plugged into, and one Nvidia K5200 with 4 projectors plugged into it. The game repositions itself to the coordinates of the second video card at run time. We want the GPU processing to occur on the K5200, but MS Windows always assigns it to the K620 where the primary monitor is plugged in, and when viewing the GPU performance monitor in the Nvidia control panel, we see the k620 maxed out and very little processing occurring on the K5200.
This is possible using other game engines such as Unity, by passing the adaptor number as a command line argument at runtime. We are trying to find a similar solution.
Trying to point a process / application at a specific GPU using the Nvidia control panel does not seem to work with any engine or application.
asked Jul 20 '15 at 08:35 PM in Rendering
Follow this question
Once you sign in you will be able to subscribe for any updates here