Quote Originally Posted by trinibwoy View Post
The Fermi pipeline is optimized to switch to below 25 microseconds, a significant improvement over last generation GPUs. Besides improved performance, this allows developers to create applications that take greater advantage of frequent kernel-to-kernel communication, such as fine-grained interoperation between graphics and PhysX applications.
I'm afraid you're the one who needs to learn how to read. The above quote is talking about running PhysX and graphics on a single chip, hence the need for of the architecture for fast switching between kernels. This is different then using a GPU as a standalone PhysX card.