Hi, I was wondering if its possible or if there already is an option, When you have multiple cuda cards, the general idea is the more cards you have the faster it runs, would it be possible to use the other cards for there memory and not there speed?
Yes - as you have to manually distribute the work for the different cards anyway, that is entirely up to you.
Ok great. So this would basically be up to the programmer of the software and not driver options correct?
I’m going to be using the cards for 3D rendering (not games), so if its up to the software i guess I’ll have ask them to add support.
Oh - it seems I incorrectly assumed the programmer of the software to be you. Yes, this has to be supported in the program, the driver cannot do anything about it.