Yup. That's the standard that they accept.
On a hybrid, I think what I'd do is run the HPL on the Opterons, then run one on the Cells. Not sure how I'd combine the results, though. They don't really go together linearly.
Mathematically and CPU-intensive elements are directed to the Cell processors. Each tri-blade unit can run at 400 billion operations per second (400 Gigaflops).