Closed > Open
NVIDIA > AMD
AMD Open > NVIDIA Open
Ohh Jesus who would have thought that...
Originally Posted by blackout23
In throughput, which is what was tested here. So yeap, the results to me were pretty obvious without even looking. Although I'm sure that owners of one of those particular cards appreciate the test.
Originally Posted by brosis
No, it isn't, because a horse has no relation to a car aside from the fact that it CAN be used for transportation - the nvidia 8000 series has some resemblance to the 7000 series, and the 9000 has some resemblance to the 8000, and so on. Your comparison in computer terms would be like learning how to develop software for a keyboard vs a mouse. What I'm saying is results will be better if everything is taken 1 step at a time. Why do you think intel is working on openGL 3 first when their real attention is toward 4? You can't just jump ahead and expect success.
Originally Posted by calim
What OpenCL applications are you interested in using?
Originally Posted by wargames
Fine. But that approach is not practical, because then you'd not be able to use nouveau on modern cards at all; also, for example, one of the components extremely vital for performance, the shaders, are completely different between GF7 and GF8+ (from whence on the compiler code and a lot of other code, too, is shared), and that's just the beginning. Yes, cards between GF8 and Kepler didn't change much, so we're starting there. It's just easier to work on the more modern ones first because they're less quirky, and they happen to be plugged in because I'm using them for other things.
Originally Posted by schmidtbag
And the MOST important thing to note: It's NOT that we understand the cards poorly, it's that we are staffed poorly.
Well, just keep up the good work any way that you can. Personally I think you guys are doing a fantastic job without nVidia's official support.
Question? How many contributers are too many before toes start getting stepped on? The only reason I ask is that I've heard on this very forum board that having more developers wouldnt help because then they would just start overlapping efforts and step on each others toes.
Well to be honest if you wanna take gaming on linux serious and you do benchmarks start including games that have some higher standards, the quake games are nice and cool but run an ancient game with just some improvements.
You should really get with some developers from Steam/3rd party to get some real game benchmarks suites running with games like TF2, SSam 3, Oil Rush. Those would stress the cards a lot more and will use more features that modern/semi modern cards offer.
You missed that the game is based on ancient engines that are CPU limited, not GPU limited. when your system is kicking out frames faster then your screen can display them you are best off turning on Vsync since you are never going to see those frames save for maybe a small piece of them in a partial screen chop.
Originally Posted by Vim_User
Originally Posted by duby229
The number of contributors is difficult to express in a meaningful measure because most of them (in this case, all but me and the guy hired by Red Hat) only very rarely find the time to do some nouveau-related work. Some are just occasionally submitting fixes when they run into issues. Another one does only REing (but at miraculous speed) and not writing any driver code at all.
Originally Posted by duby229
Suffice it to say that there's no danger of anyone stepping on another's toes by a long shot.