[JohnM15141]

I ordered one shortly after it was announced.

If I were to be so lucky as to get one early, how should I benchmark it so that it is credible and nobody cries foul like what has happened with the Barefeats benchmark?

I'm just curious...

[icruise]

I'm not sure it is possible to get a set of benchmarks that everyone will agree one, but I for one would be more interested in everyday type of benchmarks. How long it takes to rip a CD using iTunes, how many frames per second you get in Quake 3, whatever. These seem a little bit easier to identify with.

[Dan-the-Man]

Yep I agree. Benchmark programs reveal theory only. It is practical tasks that reveal the overall performance of a system (can any one say "megahertz myth"?).
3D games tend to give a good indication of overall system performance as they push heaps of data through memory/CPU/GPU, and to a lesser extent the HD bus.
Quake III 1.32 for OSX can be downloaded from http://www.versiontracker.com/macosx (assuming you have the original boxed copy) and you can run the four.dm_68 timedemo test.
You need to include what resolution, texture and colour depth you use, and what detail settings you used in order to make this mean anything.
I will do this when we eventually get one here for ourselves in the store, and post the results. we may have a wait though.