If you're curious, here is an optimizing assembler.

https://code.google.com/p/mao/