I use three monitors for editing. With Photoshop, I will commonly have the edit screen filling my IPS calibrated monitor plugged into my fast, 2gb video card; the menus and tools on a secondary (cheaper) monitor connected to a 2gb card; and another monitor available for any other purpose connected to that same 2gb card. Both cards are CUDA capable.
When editing on Premiere Pro, CUDA is available for the Mercury playback engine for either card. But of course, I'd want to make sure it's my primary card that bears the weight of the gpu processing. How do I know that's what's happening?
Can I presume if I've got the timeline and preview windows on the screen connected to the primary card that all the processing is being directed to that card? Or did Premiere Pro pick up one card or the other (perhaps depending on the PCI-E card slot) as it started and does all its processing on that card regardless of the monitor location of application elements?

