How-to f**k Snow Leopard 10.6 benchmark result…
MacWorld “Lab” tested 10.6 Snow Leopard speed boost over 10.5 Leopard. But they did it the wrong way, and ended-up with expected wrong results: Snow Leopard hard-drives performance is slower than Leopard.
How they did it: they use 2 partitions on each computer with half-drive space for each, first with 10.5 Leopard, second with 10.6 Snow Leopard. A beginner error that clearly impair every test where hard-drive is involved, data copy, boot time, etc…
As any IT guy know, or must know, the average speed of a hard-drive decrease from the beginning of the disk to the end, usually loosing 40% to 50% performance at the end! That’s why the boot partition must be the first one on Mac OS X, followed by data partitions, the most used data partition first, the last at the end. Diglloyd explained it better than me
When you measure the performance of two OS on partitions located at the beginning of the hard-drive versus one located in the middle or end of the hard-drive, any disk access is faster for two parameters: sequential read and writes are faster because there’s more data per disk rotation, and random accesses are faster for the same reason : to put same amount of data you need less tracks, so the head will move less at the beginning of the disk when going from one data to the other!
Anyway, putting a test of the same Leopard OS in the beginning and end of a hard-drive will ALWAYS make the first win the race for any hard-drive related test!
So MacWorld just f**ked up the firsts tests of Snow Leopard, being incompetent!
), the new MacBook Pro is twice SLOWER than my old one, only due to DVD reader being way way too slow on DVD-Video content!