Right, but the suggestion was of using JRuby for the Ruby implementation (including Rails), and using "real Java" for the performance bottlenecks. I responded to a comment making fun of that idea.
Now, ensure that you also score memory performance.
"Memory performance" is not a goal in itself - it only matters to the extent that it affects speed (though GC, cache misses, and so on). (And of course if you don't have enough memory to run your application at all or without paging, but that's seldom a problem these days.)
Note I'm not saying that memory usage is unimportant - today calculation is free and accessing non-cache memory is expensive. But nobody cares about memory usage except to the extent that it affects speed.
As you point out, these are short artificial benchmarks, so they may not tell us much about real-world performance. My point remains: Don't dismiss the performance of Java. A lot of smart people have worked hard on making Java run fast, and that is one reason it is widely used in "the enterprise".
The article poster hasn't been paying much attention
Hint: Google for "android g1".
Remember, UNIX spelled backwards is XINU. -- Mt.