So were you a little suspicious with Apple's Mac-vs.-Wintel bake-off at last July's Macworld Expo keynote address? You know the routine; every time Apple introduces faster processors (or, in this case, two processors for the price of one) at one of these big shindigs, federal law mandates that Steve drag out one of Apple's top machines and run it side-by-side with whatever Wintel system happens to be the fastest at the moment. The idea, of course, is to prove that megahertz ain't everything-- and in Steve's orchestrated demos, the Mac always wins hands down, while Phil Schiller's PC is left choking on the dust. When he demonstrated the dual-500's Pentium-crushing power last July, Steve proclaimed that its performance in Photoshop (one of those exceedingly rare Mac applications that is written to take advantage of both the G4's Velocity Engine and the presence of multiple processors) was twice that of a 1 GHz Pentium III.

The folks at PC Magazine, unsurprisingly, were a tad skeptical of those claims. So, as faithful viewer Wayne Parkhurst told us, they decided to run some tests and see for themselves. The results of the benchmarks are either surprisingly good (if you've always taken Apple's performance demos with a couple hundred grains of salt) or surprisingly bad (if you're so thoroughly susceptible to Reality Distortion Field energy that you seriously believe that a 500 MHz G4 will always trounce a 1 GHz Pentium in any scenario because it's a "supercomputer"). Since Apple's claim was that two 500 MHz G4s were twice as fast as a single 1 GHz Pentium III, PC Magazine decided to pit a dual-G4/500 against a similarly-configured Wintel box with two 1 GHz Pentiums. The result? The Mac won the Photoshop tests, albeit barely-- winning three of the seven races, losing two, and finishing the remaining two in a tie. Yes, for Photoshop use, apparently Apple's top-of-the-line Power Mac really is faster than even a dual-processor 1 GHz Pentium setup. Color us impressed.

Of course, it was the single-processor tests in which the Mac got stomped like a grape at a wine-making festival. In a battery of eight tests comprising actions in applications such as Illustrator and Bryce, the Mac lost all but one. But wait a minute, here-- we're not entirely sure these tests are on the up-and-up. For one thing, PC Magazine doesn't say which operating system the dual-1 GHz Pentium box was running. Windows NT/2000 supports symmetric multiprocessing, so were those single-processor tests really only using one processor on the PC side? And even more suspicious is PC Magazine's description of how the Mac was configured: "We turned virtual memory off for Power Mac applications and set each application's memory to 128MB minimum and 180MB preferred... both systems had 128MB of RAM installed." What's wrong with this picture? Say, perhaps, that if PC Magazine is telling the truth, the Mac wouldn't have been able to launch a single one of those applications after booting? We smell a rat...