> I am pretty confident (and hope to test, but it sounds like you might > beat me to it) that for either the classical (non-Rembrandt) render, > doing a depth-write *only* pass of the entire scene will be a net win. > (Again assuming the fragment shader is the bottleneck). Since this will > reduce overdraw to zero - solving both the 'cockpit covers terrain' > optimization, but also where random trees or buildings do. (Which is > presumably where your 30% gain comes from)
I guess there's only trees, random buildings, cockpit and clouds which can cover substantial parts of the scene (all else is rare and exotic...). Trees I've accounted for, random buildings most likely cover urban landclass which has no overlay texturing and is hence pretty cheap to render if no urban effect is on (in which case you can't have buildings), so the expected gain is less. Clouds are a bit tricky if done directly because their rotations sit pretty heavy on the vertex shader, so doing a second pass might create a new vertex shader bottleneck (at least in one try it did, but I may have done something fishy there). Hence the idea to slip the obscuring rectangles in... The cockpit is the biggest potential gain, but due to the near camera - far camera thingy, I don't see how this can be done on the level of editing effect files - maybe a suitable edit of the camera group code can pull that off, but I have no idea where to start there. If you figure that one out, it'd be great. * Thorsten ------------------------------------------------------------------------------ Don't let slow site performance ruin your business. Deploy New Relic APM Deploy New Relic app performance management and know exactly what is happening inside your Ruby, Python, PHP, Java, and .NET app Try New Relic at no cost today and get our sweet Data Nerd shirt too! http://p.sf.net/sfu/newrelic-dev2dev _______________________________________________ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
