On Thursday, July 26, 2001, at 10:47 PM, Chris Nandor wrote:

> Benchmarks on Mac OS are notoriously unreliable.  If your email app 
> decides
> to look for email, it can throw off your benchmark, etc.  I wouldn't 
> worry
> about such a small number.  If you saw a difference of 1 minute to 2
> minutes instead of 1 second to 2 seconds, that could possibly be 
> worrisome.
> You need to run benchmarks a few times and look at the results, and even
> then be skeptical.

I don't see why you should single on the Mac OS here :-).

You should always run more than one benchmark and apply statistics to 
the results.

If you are running on a pre-emptive OS, and the kernel decides to give 
some other task a big chunk of CPU time during your operation, you'll 
get different results too.


--
Jim Correia
[EMAIL PROTECTED]

Reply via email to