"I didn't read the whole article, but the chart shows a P4 2.66 on par with a G5 2 GHz.
IIRC the 2 GHz is Apple's fastest, while P4 is available up to 3.2 GHz."
Quite true, however that just brings up a couple of relevant factors mentioned in the study.
Since the G5 was so new, the code was not optimized, and he estimated that optimized code would bring up the G5 another 20% or so (bringing the single G5 2.0 very close in performance to the single P4 3.2).
IBM has also released a new compiler for the G5 which brings up the speed on an order of anywhere from 60 to 120 percent.....the G5 hasn't even come close to having it's potential tapped yet.
Now extrapolate the performance of the Dual G5, and you have a clear winner, and the P4 is not capable of multiple processsors (you have to go to the Xeon).
The point I'm trying to make here is that the G5 is a very strong chip, and the reason it was chosen for the VTU cluster was based on price/performance.
"It's just a computer platform, people!"