About the article comparing Windows XP/Vista and Linux w/wine/Cedga
I have a few comments about the article... I'm so terribly sorry if there's already a thread with a discussion about it, I must admit my search skills are "teh suckz", if that is the case, mods feel free to delete/merge this thread.
At any rate, my comments are in regards of the following:
Cedega and Wine both implement the Win32 API and DirectX API on top of Linux so that native applications can run, said it in a simple way, Cedega's been tuned for gaming, while (for instance) CrossOverOffice has been tuned for productivity desktop applications, though both are based on Wine (at least partially). With this in mind, I think that the tests should have to include at least one DirectX title ideally a title that uses DirectX in Windows and OpenGL in Linux; too bad you had troubles with UT2004, it'd been the app to test, as you get a native port with OpenGL.
Failing the above (UT2K4), another way to really test both Wine and Cedega performance would have been to run the most used Windows game on Linux (with either the OpenGL or Direct3D renderer): WoW. However I don't know how would Vista had affected these results.
The point of both Wine and Cedega is to try and implement those APIs used in Windows, and while installing some games may be challenging, if the games are OpenGL, they virtually will skip the whole "Windows" management and API, especially if they use either GLUT or SDL for Window management, for which Linux has native libs, that Wine or Cedega can make use of, skipping the most demanding task of these "compatibility layers": Converting D3D calls into OpenGL calls, or D3D shaders into OpenGL shaders. That would make a much more interesting article.
I have to admit that I was a bit surprised by the selection of hardware for this test, though... Maybe for the next time you could choose a more "mainstream" video card (something along the lines of GF 6600/7600 which seem to be fairly common in Linux desktops) and maybe a plain A64 3200 single core... Of course there's quite a bit of room for a LOT more tests that could be performed constrasting the four (WinXP/Vista LinWine/Cedega). At any rate, kudos for the article and I'm eager to see the next round!
This is no comparison. Take a 7900 or better a 8800 and we'll talk again.
The test was done using a card that most Linux gamers and the public in general would more commonly use. Testing on a geforce 8800 would be trivial and would not help any out. Plus if a low end card can be shown to have reasonable framerates how could you argue that wine/cedega is not a great alternative to gaming on windows.
Testing doom 3 on Windows Wine/Cedega really is the best choice as both Windows and Linux have a native OpenGL client and when compared to Wine/Cedega you get a real benchmark to go off of.