Well, I have
another one that proves reverse: Parallels Desktop is faster than VMWare... Don't know how old it is but it looks rather convincing...
I actually don't like that article. It's not based on the actual differences between the speed of Windows applications themselves under Parallels Desktop vs. VMware Fusion. Instead, it's based on "productivity" differences where Windows applications are launched from OS X or vice versa (like opening Mac Excel from a Windows Outlook attachment, or opening Windows Excel from a Mac Entourage attachment). In this case, Parallels has a nifty feature that allows associating Windows applications with OS X files and vice-versa, VMware does not, and in the test they count the time in VMware to save the attachment, manually launch whatever version of Excel, manually find the file, and then open it to the actual application results. Boo.
For that reason, I think it's actually more a marketing tool to sell that Parallels feature (which is nifty) then a real comparison of the two. It'd be like a benchmark that relied on VMware's dual core support (where Parallels only supports one) and then built a suite of tests that relied on that dual core support to beat Parallels... it'd not really be a good comparison, but it'd be a great marketing tool. Or worse, a set of 64-bit Vista tests... oops! Parallels can't run that! And it wouldn't really be a test, would it?
It's not what people expect in a benchmark. They want to see stuff like here's a nasty Excel macro, how long does it take to execute under Parallels vs. VMware. Or the same game on both. Or the same exact task with the same exact clicks on both. Or the impact on CPU use with the same exact task, or memory, or response time.
I also HATE how they present charts in that article, it's like they were purposely trying to make it difficult to compare the products and draw conclusions.
