Long story short, I grabbed the "Hello world!" examples from both Node.js and Rhino running on Jetty and ran 'ab' on them.
I was not able to do high-concurrency tests on my Mac, because Jetty gave up around 100, and Node somewhere around 200. I think I'm having the issue described here.
So here is the graph. I don't have much experience with benchmarking, but again, it is obvious that Node.js is in fact quite a lot faster than the Rhino/Jetty combo.
