SPUR is built on the earlier Bartok compiler, built for the Singularity project, which provides a lot of the awesome optimizations, e.g. guard implication/strengthening. MSR has been working on some really cool projects; would be nice to see them rolled into the main product line.
Please, someone bring me back down to earth before I start reading papers again instead of studying for my exams...
by all means, read more papers ;) study just enough to get a decent grade on your exams. if you're into reading academic papers and find them enriching, they can be a far more rewarding experience than studying for exams (and they might actually indirectly help you with future exams, heh)
Probably not future exams but you learn more, exam study is really about going over the same stuff you have already been over to ensure you remember it down to the minute details. Usually very uninspiring, unless of course you haven't touched the course before exam time then it can be interesting.
LuaJIT is great and it's very interesting. (In fact, I've been reading through its source code trying to understand the magic.)
My impression is that the developer is reading the important papers and putting the theory into practice. It's innovative in the sense of "new stuff that is made useful".
Guys like the SPUR team at MSR on the other hand are trying new ideas and writing these papers. That's innovation in the sense of "a change in the thought process for doing something".
Both are important and each one would be useless without the other.
On both reddit and lambda-the-ultimate; it has all the important stuff. Also there's the shootout stuff already posted as a reply to you. And you can download LuaJIT2 and experiment yourself.
