Gee, what a surprise; Apple publishes benchmark results which show that the Power Mac G5 outscores the fastest Intel-based systems, and, as faithful viewer moff moff points out, everybody and his Windows-using grandmother comes scurrying out of the woodwork to cry foul. And why not? After all, the only reason benchmarks exist in the first place is to give the geeks something to fight about, because they sure as heck aren't for proving which hardware is actually faster or anything like that. So let's watch the fur fly, shall we?

Lots of the hubbub from the Intel apologists seems to stem from spl's soapbox at Haxial, where the allegedly Mac-using author makes several seemingly valid points about ways in which Apple stacked the deck in its own favor. He took a look through the details of the tests performed by Veritest and found all sorts of fishy fun facts-- for example, the way that the tested G5 used a "high-performance, single-threaded malloc library... geared for speed rather than memory efficiency...[and] unsuitable for many uses." Well, that doesn't sound too kosher, does it?

Moreover, on certain tests, the Intel Xeon's Hyperthreading feature was disabled, and used a compiler option that seemingly disabled performance-boosting SSE2 instructions as well. Finally, the G5's registers were manually tweaked "to enable memory Read By-pass," which implies that this is a benchmark-friendly feature that won't be turned on by default in actual shipping Power Macs. Meanwhile, The Register pointed out that Apple's claimed SPEC scores for the dual Xeon Dell system were much, much lower than Dell's reported scores for the same box. And since it couldn't possibly be Dell, obviously Apple's the filthy pack of cheating bottomfeeders trying to put one over on us all.

Or not. Believe it or not, Apple cares enough about the hardcore geeks trash-talking its benchmark methodology that hardware product marketing czar Greg Joswiak actually contacted Slashdot to try to set the record straight. Apparently memory read by-pass may well be enabled in shipping G5s, which is why is was enabled for the benchmarks. Yes, Apple's reported scores for the Dell were lower-- because they used the gcc compiler on all systems in an attempt to level the playing field, while Dell used Intel's own compiler, which one would certainly expect to deliver higher scores; "the Apple numbers could be higher with a different compiler, too." (Codewarrior, anyone?) Yes, Apple did turn off Hyperthreading in the Xeon-- because the Xeon scored higher without it. And no, Apple didn't disable SSE2; people who thought so simply misinterpreted the gcc options. In short, Joz claims that Apple went out of its way to be fair and maximize Intel's scores within the context of using the cross-platform gcc compiler. (No word on the super-secret special malloc library, though.)

Of course, you're allowed to believe whomever you want. Personally, given all the variables present when calculating SPEC scores, we've never considered them much beyond their marketing value. What we tend to trust is real-world performance, and the say-so of the people who write the cross-platform applications. When that crazy Mathematica guy says that the G5 is not only trouncing the Wintels but also beating up on high-end UNIX systems that cost twice as much, we see no reason to doubt that-- especially after that onstage bake-off in which Mathematica on the G5 flew past the Dell so fast we're pretty sure we heard a little sonic boom. We can hardly wait until August when people can run their own independent head-to-head tests to see what the real deal is.

Meanwhile, to get a sense of the sort of axe-grinding going on back at Haxial, check out this complaint: "Apple gives the price of the low-end G5 as '$1999,' and the high-end G5 as '$2999.' In other words, they have subtracted $1 from a $3000 computer to make it seem cheaper, which is absolutely ridiculous. This demonstrates that both Apple and Dell are willing to mislead people when stating their prices." Wait a minute, you mean that $2999 top-of-the-line Power Mac G5 is actually almost $3000?! But how could we possibly have been expected to know that without being told? Why, we're going to write our congressman about this vile attempt to bilk the general public; if we don't nip this nefarious practice in the bud now, eventually everyone's going to be pulling this "one dollar less" scam!