The models are all embedded in the SWF, so the results don't include download times.
On Feb 26, 9:37 am, "Joshua Granick" <[email protected]> wrote: > Is this a benchmark of parse times, or download + parse times? > > If it combines both, it would be interesting to benchmark separately, > depending on whether you need to prefer download performance or runtime > performance. > > Of course, whether your model is parsed correctly also matters. I've > certainly had trouble with that before. So far using Maya + OpenCollada has > been the most reliable (only reliable?) combination I've used. > > > > On Thu, 25 Feb 2010 14:59:05 -0800, Matt <[email protected]> wrote: > > It looks like the results change quite significantly with the less > > complex sphere mode. Interesting... > > > On Feb 26, 8:50 am, Matt <[email protected]> wrote: > >> The app has been updated with an MD2 file supplied by Peter. I have > >> also randomized the order in which the models are loaded. > > >> The complexity of the sphere model has been reduced, so loading times > >> will probably drop, by the relative load times should still be valid.
