Will wonders never cease? Apple said that the iMac G5 would ship in mid-September; well, spin our vowels and call us Vanna, because here we are smack-dab in the middle of the month, and darned if people aren't indeed receiving stock-config iMacs that they preordered shortly after the product's Philnote debut! Could it be that when Phil Schiller mentions a ship date, he means that actual date, and not a later one parsed through a Jobsian verbal time dilation filter? Apparently there are benefits to getting a keynote delivered by a native Earth life form now and again.

So, yeah, iMac G5s are already dotting the landscape, and one of the cool side effects of that happy scenario is the appearance of some quick 'n' dirty benchmarks of the actual shipping systems. MacRumors has posted the Cinebench and Xbench scores submitted by a couple of lucky readers whose models had just arrived, and reportedly the stock 1.8 GHz units score a 243 and a 134.71, respectively. So should you get excited over the new iMac's performance? Well, uh, that depends on which numbers you look at.

See, according to Bare Feats, a dual-processor 1.42 MHz Power Mac G4 scores a 247 in Cinebench, so considering that Apple's new single-processor consumer system matches the performance of the company's fastest pro-grade desktop from just over a year ago at a fraction of the price, we think that's pretty spiffy. (Remember, a 20-inch 1.8 GHz iMac G5 costs $1,899; last year a dual 1.42 GHz Power Mac G4 plus a 20-inch Cinema Display cost a hair under four grand.) Better still, a single-processor 1.8 GHz Power Mac G5 scored a 251, which means that the iMac is certainly comparable in Cinebench performance to its aluminum-clad big brother.

But the Xbench score of only 134 is perhaps a little more troubling. Apple's G4-based PowerBooks have scored in that range; shouldn't a 1.8 GHz G5 beat out a 1.5 GHz-or-less G4? Well, here's the thing: benchmarks are... um... unreliable, we suppose might be one way to put it. Another way would be to say that they suck-- some more than others, and Xbench probably belongs in the "more" category. A quick glance at an Xbench comparison site shows a couple of 1.8 GHz Power Mac G5s-- you know, with the fast bus and high-speed RAM-- also scoring between 125 and 146, while some aluminum PowerBooks score in the 130s. If Xbench is trying to tell us that a Power Mac G5, an iMac G5, and a PowerBook G4 are all neck and neck (and neck) performance-wise, we're just not buying it.

As it turns out, Xbench scores can fluctuate wildly between successive benchmarking runs, and they're also very sensitive to available RAM. (The iMac in question was Xbenched with its stock 256 MB of RAM, so it was almost certainly gagging for more.) So we wouldn't put much stock in the Xbench numbers-- or even the Cinebench ones, for that matter. Because what really matters is real-world performance, and thankfully, the customer who submitted the original set of scores posted a glowing early performance review.

It seems that this guy has his own real-world benchmarking tools that measure how a given Mac will perform at his business, doing the actual tasks they'll be expected to do on the job; the iMac G5 repeatably scores a .948, with 1.0 being a single-processor 1.8 GHz Power Mac G5. Combine that fact with words and phrases like "ecstatic," "amazing," "beautiful," and "quite snappy even with 256 MB" and we think Apple's got at least one very happy customer.

It'll be a while before enough iMac G5s are in the field for a clear picture of their relative performance to emerge, but based on this very early testing, it sounds like cramming a G5 into a two-inch-thick screen (without nine fans and four independent thermal zones) didn't cramp its style much. So if you've been drooling at the new iMacs but concerned about the performance, so far it looks like you don't have much to worry about. Well, other than trying to figure out where you're going to get $1,799 for a 20-incher...