iTunes benchmark for CPU's

[Bainne]

Hey all,
I have noticed that Anand, in their CPU and other benchmarks, uses iTunes in order to compare efficiency and speed. I have noted that other review websites do this as well.
I encode quite a bit of music from my massive collection and I am trying to find a way to compare my current system's preformance to that of a potential upgrade and between systems I am currently running

However I am slightly confused as to how all these review websites use iTunes and get a time for MP3 encoding as I cannot find any built in timer? Only the speed multiplier.

While I realize that using a manual stopwatch would be the easiest, would that also not give pretty inaccurate results? I would like to be as close as I can also many of these choices differ by mere seconds so how can I be sure my results are bang on?

Is there a command line script that could be used to log the start and stop time of the transcode?

If manual timing is the only way it can be done, why is Anand, toms, neoseeker and all these other websites using a method like this, when you look at tom's CPU charts, there are quite a few ties, how can they be sure those are accurate results?

Any suggestions or revealing of the method behind the madness would be appreciated.

 

[v8envy]

A bit of a reality check. If you can't tell the difference between two CPUs using a stopwatch (assuming a task which takes > 1 second) then is the difference in performance significant? I'd say no. And if the task typically takes < 1 second, are you going to notice a CPU that's even 30% slower in normal usage?

Differences in performance of 5%+ will be easily detectible and even a .5 second margin of error will not skew the results if your tasks are lengthy enough.

That said, I would imagine the creation and modification timestamps of the resulting encoded file in the filesystem give you the start and end time of the task to within 1/100th of a second. =)