As to those two tests, BenchJS I’ve seen before and I like, the Browser Speed Test page I’ve also seen before but I don’t like. It seems like the author really doesn’t know much about the browsers or how they work—if you want to test “CSS speed” you should have the page already loaded and change the stylesheet, not completely reload the page (which requires it to be reparsed, etc.), if you’re loading many small images you should at least mention pipelining, max-persistent-connections-per-server, etc., if you’re testing table layout, you shouldn’t have form controls or other junk on the page… all the things like that. And if you want a real-world test, you shouldn’t restrict yourself to a small, hand-picked sample size, so it doesn’t qualify as that, either.