High Performance Ruby in RPython
docs.topazruby.com
docs.topazruby.com
http://tratt.net/laurie/blog/entries/fast_enough_vms_in_fast...
* Most of the standard library is missing.
I don't think "high performance ruby" is very useful without the standard library, personally.
If it's not so great, of course, it doesn't matter if it's active.
Every time I tried it out so far it turned out slower than MRI for me. Perhaps I was always hitting on pathological cases, though (tight data processing loops).
Separately, the start time makes JRuby awful for TDD/Dev, but once it has time to warm up in production it is vastly faster. Scan back through some Ruby Rogues episodes, they discuss it a few times in more detail.
To see Topaz at its best, scroll down to Benchmarks and click Synthetic.
Truffle is great, but strangely its documentation is spread all over the Web, one piece on java.net, another on github/jruby's wiki, yet another on uni-linz.ac.at... How does the team work?
The documentation is a bit all over the place at the moment, reflecting the different aspects of the project - part academic/industrial research collaboration, part open source product development. We're not trying to get anyone to use it yet, which is why there isn't a single coherent flashy landing page, but we'll get to that in time.
Off topic now but Truffle looks slightly more interesting than Topaz.