I've noticed Firefox and other application loading times are much lower in Windows XP compared to Windows 2000. Both machines are similar (actually, the Windows 2000 one is a bit faster has a SATA HD in UltraDMA mode, while the other has an IDE HD in UltraDMA mode), and both of them have all the unused services disabled, no crapware/system tray bullshit, very few processes running, no antiviruses or anything, and fresh boots.
What's Windows XP doing/not-doing? Is there a way to have Windows 2000 do/not-do the same?
Name:
Anonymous2006-03-29 11:01
Well, firstly the interface type of a drive isn't necessarily a good indicator of overall drive performance: there are factors like seek time, buffer size and spindle speed that affect the rate at which data can be read from a drive (you could easily have a slow-ass SATA drive and a rock-solid fast PATA drive).
However, what I really suspect is causing the performance difference is XP's "prefetch" system at work. Every now and then, during idle system time, XP reorganises files that are used when booting and launching common applications into a contiguous block of space on the disk, making the files much faster to access. So when you launch an application, XP can just pick up all its files in one area of the disk, while 2000 has to go seeking back and forth across the disk to find the same files. Hard disk seek times are in the order of milliseconds, a hell of a long time by comparison to CPU/memory operations (nanoseconds), so anything that can be done to shorten the wait by preventing that slow-ass servo arm lumbering lethargically back and forth across the platters looking for stuff is going to make a massive difference to performance (this is also why defragging is a good thing). Sadly, 2000 does not, and will never, support prefetching. :(
Name:
Anonymous2006-03-29 12:55
I see, I was thinking it had to be something like that but I didn't know it did it in the background.
Aren't there third party defragmenters that will do this for you? Intel had something like this a long time ago.
Well, firstly the interface type of a drive isn't necessarily a good indicator of overall drive performance
Very true, I just forgot to say both are Seagate Barracudas.