Great to see Unigine pretty much the same between the two OS's, especially considering Linux has no real quality development put into the graphics side of things.
These tests show that Linux is maturing in the gaming and graphics. We are in a great period where opportunity for game makers opened up by mobile Linux devices prominence. Linux has a special placing in mobile area that Microsoft can't meet any time soon. People simply don't want Windows Mobile 7, or other operating system that are limited and suppressive. Apple is the only exceptions to this rule as they are attracting people who are buying an fashion image, not so much computing.
All Linux fields should be dedicating all their emphasis on graphics drivers, 3d standards, and quality of use, or the opportunity will be lost.
Value (Steam) know that Linux is the next big money area. If they get in now they dominate professional mobile gaming.
Excellent job. I imagine you had the pain of actually using windows 7 during development and testing, so kudos to surviving it too.
Benchmarking vs windows is always good to put things in perspective.
You should think about including some subjective tests every now and then. Many 3D sites do this, capturing particular frames using different cards, since FPS doesn't always equate to quality.
PTS already does this, just the Windows profiles for them are not yet implemented.
Can't help but notice the P4X system seems to have a severe CPU-bottleneck, at least in those benchmarks. The extra core doesn't do much good there, does it?
One thing I miss though, is maybe a short summary of the in-game settings used for those who have never used PTS.
@Michael, on page 3:
Shouldn't it read "through 1280x1024 with Windows"?Quote:
The Clarkdale system was playable at 800 x 600 through 1280 x 1024 with Nexuiz, but even at 800 x 600, the open-source Intel graphics driver used by Ubuntu 10.04 LTS put out just six frames per second.
And page 1, I think the "it's already running" link should be placed after the "Steam client" one, really ;)
One thing that is definitely missing is a DX run of the tests. OGL vs OGL is fine to have as well but to give a clear x vs y comparison a DX run should have been ran since that is what most windows gamers utilize.