Or you could try running the PhysX stuff on the CPU, because our stuff is cross-platform. We support CPUs as well as the Wii, Xbox 360, and PlayStation 3, which is something that people tend to forget about.

On top of that, on the PC, if you're running PhysX on the GPU, it's also scalable, which is something you can't do on the CPU with competing solutions. So that means the more GPU power you have in your box, if you add a second or third card, then the amount of PhysX effects on the screen can scale up. For example, if a building blows up into 1,000 pieces and you add a second card, maybe that building blows up into 5,000 pieces.

GO: Is there a way for people to tell what that increase would be when they're weighing the cost of adding secondary or even tertiary hardware?

BDR: Think of it like SLI [, NVIDIA technology that allows you to combine more than one GPU in a computer to increase visual performance]. You have two chips and you almost get 2X performance, or three chips and it's almost 3X performance. In the past what we've done with SLI is that you've had to have two identical GPUs inside your machine for the scaling to take effect. Mixing and matching different generations of cards doesn't work because you'd end up dumbing down to the lowest common denominator.

With PhysX in an SLI configuration, you can designate one of those cards for PhysX rendering, and one for GPU rendering if you want. Or you can actually use an older card. Say you've just upgraded to a and you've got a sitting on your shelf. You can add that 9800 GTX back into your machine and use that as your PhysX renderer.