When the G5 first came out, the mac community went orgasmic. For the first time in years we had a system that could blow the panties off any PC on the planet. The PC advocates screamed bloody murder. No fair, no fair! The fact that the tests were performed as fairly as possible didn't even enter their head. They were waiting for the tests that proved they were right and that the G5 blowed the big goat in the sky. Finally, many months later, this article arrived (in PC World, no less), with these results.

Now I can and have pointed out why this test is a complete and utter waste of time, energy and paper clips. But those PC users have very thick skulls and I cut no ice, so I've decided to take an alternative route to proving the point I've been making all along...

Apple Mac's are better than PC's.
37,853,081 times better infact!

Since I don't have a G5 (donations welcome, will do "favors" in return) I used what I had at hand. One iBook (G3 800MHz, combo, 12", 640mb, 30gb, airport) and one PC (PIII 500MHz, cd-rw, 17", 256mb, 30Gb). They both cost about the same, though the PC is considerably older at 3 and a half years, while my iBook is merely a youngster at 6 months (she can't yet crawl, but can say her name. The PC is still working on that one, but can throw killer tantrums!).


Since PC World went to all the trouble to document their tests in extreme detail, I think I'll do the same. The test was: Photoshop, 9 actions, 6.2mb image.

Now thankfully you'll be able to replicate this test exactly at home, cos I told you all the nitty gritty about it. Heck, I even mentioned what actions I did on the image and what version of Photoshop. Oh... I didn't? I'm certain I did, I'm following PC World to the letter. Oh alright... here's what I did:

• Open hands.psd
• Resize to 150 dpi (constrain, bicubic interpolation)
• Rotate 3° CW
• Unsharp mask (50%, 5.0px, 0 levels)
• Smart blur (5.0px radius, 25.0 threshold, low quality, normal mode)
• Merge visible layers
• RGB » CYMK
• Gaussian blur (5.0px radius)
• CYMK » RGB

Happy now? No? What! Okay, I used Photoshop v6.01, under OS 9 and windows 98 (Winders had virus and firewall software running, since you need to have them running these days.) Right, that's all I'm telling you!

The actions above were repeated three times (I was feeling impatient) and then averaged, giving the following...

• Mac (OS 9) average: 31.5s
• PC (Win98) average: 40.8s

But we can't stop there. We're comparing (as someone pointed out) oranges to hats here, so I've applied some adjustments to try and even things out. Why? Well to be fair on the PC, it is older and has less ram, but then again it is bigger too, and has a faster hard drive and it doesn't look as pretty. So behold the following adjustments...

I took the iBook's results as a baseline, then adjusted the PC's in the following manner...

• Ram: 3 times less, so divide result by 3
• Age: 2 generations out (by Moores law), so ÷ result by 22
• Hard drive speed: 4200 vs 7200, so multiply PC's result by 1.7
• OS: assume that win98 and OS 9 are roughly equal, so no change
• Cost: both cost around the same, so no change
• Size: the PC is 39 times bigger, so multiply its score by 39
• Weight: the PC is 12 times heavier, so multiply by 12
• Power consumption: The PC consumes approx 3 times more power than my mac (probably much more)
• Screen size: The PC has 1.4 times as many inches of screen, so handicap it by that much
• Noise: I can hear my PC upstairs, but have to actually concentrate to hear my iBook, so I'm assuming that the PC is 5 times louder, so multiply it's score by 10⁵ since sound is logarithmic.

Yeah, so I'm assuming that a lot of things increase linearly, but hey, assumptions are all part of the game! Sides, this isn't really serious. I mean, wait... oh never mind.

Now that we've equalised things up a bit, how do the final results stack up?

• Mac (OS 9): 31.5s
• PC (Win98), adjusted: 1,192,372,060.6s

Which is another way of saying that my Mac (when tested "fairly") is approximately 38 million times better than my PC. Case closed, PC's lose.

What can we draw from this? (apart from Mac's being 37,853,081 times better than PC's).

Well I think it's safe to say that benchmarks always have been and always will be bullshit. They prove one thing only and that's that they're only too easy to use to prove your point. It's also impossible to compare oranges to hats.

Anyway, enough of the serious talk, I hope that you enjoyed this lighthearted look into the world of benchmarking!
Cheers,
Michael Adams.