Will the same CUDA driver handle different cards? My problem is that I have code that needs to be able to work on any compute level card, but take advantage of 2.x features if the user’s card supports them. Unfortunately, my current machine only has one PCI 16 slot, so to test the different cards (I have an 8800-level one installed, and just got a GTX550), I’ll have to swap them in and out. So:
-
Will the currently installed driver (fairly recent) handle the 550 if I just stick it in the slot?
-
If not, is there an easy way to deal with multiple drivers?
-
Or should I just resign myself to buying a new machine & dealing with all the installation hassles?