I'd been hearing that windows 2000 is faster than XP in parrallels vm on the mac. Is this really true? I was gonna rebuild my image from scratch (allocated way to little space for the initial image and even a second image does not work as I need some apps to be on C... Damass) and was wandering if W2K makes more sense.
It's widely claimed but I've never found it to be noticeable - assuming that you're actually comparing the same thing. If you configure XP with the classic windows theme, disable the gratuitous eye-candy effects, and trim down all of the unnecessary background crap a default install includes I strongly doubt you'd notice any difference between the two.
I would agree under those conditions that 2000 is only slightly faster mostly because it doesn't look to service multimedia crap first, then the eye-candy next, then the real application. I think that still there is an element of stability in Win 2000 that XP doesn't have. I have many VMs and the only ones that cause problems are the XP ones, but that's designed in not Parallels fault. Makes it obvious that a monoply can get by with turning out nothing and still rake in the bucks from the unknowledgable.
When using emulators (QEMU, VirtualPC, etc.), WinXP is noticeably slower than Win2k, even with the eye-candies and extraneous services turned off. Apparently, with virtualized / native environment the performance delta is easily masked, but I think there are lots of extra services happening in the background that count towards bogging down the performance. With the emulators it has always been said that you should use the oldest OS that the apps you need to run can run on. I think the same can be said when using Parallels. Plus, Win2k takes up about 1GB less space than WinXP after installation and patching is done.
I notice that my Win2k VM is faster than the XP one. XP boots faster. The Win2k has 1518 fonts installed and runs really nice. CorelDraw 12 works just great. Ordered another GIG of RAM for my iMac. Will have 2GB RAM. The system should be able to run 2 VMs then.
CPU load differs a lot between Win98 and Win2K Generally speaking older OSs require lower CPU power. However, I found a large difference in the CPU load of the Host CPU when I compared two guests, Windows 98 and Windows 2000. Windows 98 guest always consumes more than 100% of the CPU load even if no applications are running. While Windows 2000 consumes a few percent of the CPU load. I think this large difference would be due to the design of the kernel of Windows OSs, not to Parallels WS.
IIRC, not all of the Parallels Tools components can be used on Win98. Maybe that has something to do with it, as well, i.e. some tasks that can be off-loaded and accelerated aren't getting the needed benefit. Anyways, that was primarily the reason why I went for Win2k instead of Win98, despite the fact that the stuff I need can be all run on Win98 and raw Win98 installation only takes about 500MB.
