Doug, why the personal attack?
Power does equate with performance, generally, but obviously not in all cases. The best route to performance is by designing new pieces of logic that require more transistors. More transistors dissipate more heat. Thus, the direct correlation.
Intel and AMD both implement fairly well designed power systems inside their CPUs. Both use clock gating to some degree, low power idle states, etc. Unless you are comparing the power dissipation per logic transistor of a 3rd rate CPU supplier like VIA, I'd say the ratios for AMD and Intel are fairly similar (not counting caches, since they dissipate far less power per transistor than logic).
My assumption is that adding more transistors is going to lead to a proportionally greater amount of performance, based on likely design decisions from the architects. You don't waste 50% of your transistor budget on some new-fangled technique that only adds a couple percent benefit to your total performance. Thus, AMD and Intel are probably going to use transistor budgets fairly similarly.
Prescott is a higher power CPU, but I'm guessing it's because a much greater percentage of the transistors are being used efficiently, which is going to lead to higher IPC, and thus higher performance. It's not a law, but rather a rule of thumb. When you see enough cases of something, you tend to live by these rules of thumb, and many times, you can make more accurate predictions.
If you consider this worthless FUD, then I guess that's your opinion, but it's more like a personal attack, as far as I'm concerned. Ignore it if you want, but we'll see who's right. I think we'll see higher power CPUs in the future, both from AMD and Intel. In a couple of years, 100W is going to look rather quaint, but fortunately, I also see thermal solutions changing to keep up. They always have, so I'm guessing they always will.
See if you can prove me wrong, if you're so confident about your perspective.
