Now don't go interpreting this as a huge victory for SMP. It remains a costly, uncertain endeavor for boosting performance, because there is no way to know how well past and future applications will take advantage of it. Only professional 3D rendering apps seem at all interesting, and even there, the performance delta is not nearly what it theoretically should be for a CPU limitted process.
This new Office Bench test smells a lot like Bapco Webmark 2001: a benchmark engineered to ensure that a certain technology appears to achieve the performance gains it is supposed to achieve. In the case of Webmark, the test (which consists of "Internet applications" or office applications by any other name where the P4 has been proved weak) is tailored to make the Pentium 4 look better, even clock-for-clock better, than the Athlon or P3. And in the case of Office Bench, it is deliberately written to illustrate the performance benefits of SMP. If that's not synthetic, what is?
No, doubling the clock speed is always more powerful than doubling the chips. And given the sad state of their expensive, proprietary, poorly performing technology, Motorola/Apple don't deserve to mentioned in the same breath as AMD or Intel.
Modus