Paul Duncan writes:
> > Yeah, I think you're right. So, suppose we say start > off with Cessna 172 at KSFO *without* the Nimitz > scenario loaded. > You need a given time as well, to count the lighting effects out. Suggest --timeofday=noon. You also need consistency in FSAA and anisotropic filtering settings - both of these can hit frame rate on pixel-bound setups. > Regarding the Linux/Windows issue, my plan was to use > frames/second as the key field, and then it any > operating system issue *might* become obvious... The > results will hopefully speak for themselves. > > I can see this survey process is probably something > that will be refined over time, but it would be nice > to begin with a few data points :-) > You also need to get consistency in geometry - eg 800x600, 1024x768 etc. Measuring FPS at 2 pixel resolutions and 2 scene complexities (eg. Downtown SF vs. somewhere flat and featureless) is a good way of finding if you are pixel or polygon limited on a given setup. Cheers - Dave _______________________________________________ Flightgear-users mailing list Flightgear-users@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-users 2f585eeea02e2c79d7b1d8c4963bae2d
