More Longhorn Reloaded and Virtual Vista Follow-Up

The “Longhorn Reloaded” folks are in the news again. I blogged about these guys a few weeks back. They reminded me of the good old days of OS/2 zealots and similar “denial” cults. Well, it looks like they finally received a cease and desist letter from Microsoft (big surprise). Perhaps now they’ll get the message and move on to something more productive – like cleaning the windows on my house.

Speaking of Windows, I got quite a bit of feedback on my Vista virtualization entry. Looks like some of you simply wouldn’t accept the results without data from Vista + Office 2003. And since I’m loathe to disappoint my readership, I decided to re-run the test scenarios using exactly that combination.

The result: The performance aberration I observed when comparing virtualized Vista running Office 2007 is even *more* pronounced when I switch the tests to Office 2003. Whereas the delta between virtualized Vista + Office 2007 scenarios was 33% higher than it should have been (when compared as a ratio to the delta on bare iron), the delta for the virtualized Vista + Office 2003 was over 40% higher than the equivalent delta on raw hardware.

Bottom Line: Vista is slower (as measured by comparing it to Windows XP) under virtualization than it *should* be when compared to its performance vs. Windows XP when running the same scenarios natively.

Clearly, something’s going on here. I still don’t have any solid leads, however, I’m leaning towards the relative immaturity of the VMware Tools (and equivalent Virtual Machine Additions under Virtual PC) as a possible culprit. Hopefully some smart people will take a look at this and get it fixed soon.