YOU! JAVA! Stop losing at microbenchmarks I specifically designed for you to win!

Python is coming out much faster than java in a trivial tight memory allocation loop. I must be doing something wrong, right?

If you care to look at the code, go here: https://github.com/apenwarr/avebench/tree/master/memtests

It's for my presentation at http://py.codeconf.com. I was sort of hoping my theory that, you know, JITted languages are faster than interpreted ones, would work out.