What kind of computations were you doing on MRI which resulted in such performance boost by switching to JRuby? I have tried Rails under JRuby many times, the performance is the same as 1.9.3 and memory consumption usually higher.
I didn't mean 2.5Mb per copy of the JVM. I meant per runtime within the JVM. These are run as threads, rather than processes.
So we're really comparing
n instances of MRI * 50Mb
1 JVM + (n * 2.5Mb)
Sorry, what I said originally is easily misleading.