Firefox is still a poor man's Opera. IE is still a poor man's Firefox.
They all copy from Opera, anyway. Most innovations start there, and the lesser browsers pick them up later.
But, let's be real. When you're measuring milliseconds, it's not that important. Who notices? The look and feel, and personal preference will always matter more than that.
After trying to like Firefox for a week, I finally gave up. It's not bad, but Opera is just easier, and fits better. Some will prefer FF4 too. It's really personal preference.
Where Opera falls on its face is memory, as indicated. Even after you close Opera, it seems to take longer to give up all the memory. Where this would help is reloading prior pages, probably, since they probably hold onto this data to speed that up. But, it's probably more bad than good.
I haven't used IE9, but IE8 was so bad, I don't know how anyone could enjoy using that awkward, horrendous browser. It's just painful to use, in the true Microsoft tradition.
The summary about how it's unavoidable that IE9 is the fastest is pure stupidity. It's beaten in a lot of tests, and how one decides to interpret that data, based on their own criteria of usage and importance could easily come up with a different conclusion. For example, I don't think it's too far fetched to say that something differing by less than 1/2 a second is irrelevant since it won't impact the human experience.
The whole concept that all rankings are equal is wrong on two accounts; it ignores magnitude, and it assumes each test is equally important. Both are obviously fallacious.
If TH wants to crown IE fastest on this crude basis, that's fine, because there's a basis, and everyone will have a different one. But, to say this conclusion is inescapable is ignorant and uninformed, and ignores even the slightest bit of statistical understanding.