The benchmarks page was updated with fresh tests that include RightJS 1.5.5 and jQuery 1.4.2, and here some thoughts and notes about that.
Just to be clear. I don't really use the TaskSpeed framework anymore and therefore didn't update the benchmarks page for a while. The reason is that tests in TaskSpeed are not clear, don't really do exactly what they say and their results are heavily affected by utility operations, like CSS-search, iterations and so one.
More of that in TaskSpeed frameworks like jQuery, Dojo and YUI, because of luck of some functionality were cheating. Dojo and YUI did the things in pure DOM, jQuery didn't build the things at all, simply using HTML strings here and there.
But as RightJS was beating them anyways and TaskSpeed is still kinda standard tool for performance comparison, I used it to show off.
Now with jQuery 1.4 release, jQuery folks was saying that it's all not true anymore and stuff. So here we go again. You asked for that.
I've updated my branch of TaskSpeed to the latest updates in the upstream. And then, last time I was easy on jQuery 1.3.2 allowing it to cheat with HTML strings instead of proper elements building as RightJS, Prototype and MooTools do. This time we have this functionality in jQuery 1.4, so to get even and I've updated its test so all of us did the same thing.
You also might notice that in some browsers, jQuery is faster in operations like add/remove classes and html updates. Don't worry, that's because jQuery has a collections processing interface only, and because the test is mainly designed to process large collections it has an advantage. In reality you mostly handle nodes one by one and in this case it will be working much slower than RightJS. But I'll leave it like that for now, in case they release jQuery 1.5 with another 10x speedup ;)
This is about it. If you interested in more clear, feature-to-feature performance comparison, you should check my shakker project it has clear benchmarks for RightJS and all jQuery versions.