[Tuna-Fish]

Dude, your code got:

1. Turned from recursive form into iterative form 2. Unrolled heavily 3. Autovectorized (!)

The throughput will be substantially more than the simple version.

[user51442]

Well, the optimizations are pretty smart to be sure, the stupid thing is that anything over 12! overflows (as GlitchMr points out).

I'd like to know how the recursion to iteration step was done.

[eutectic]

Presumably there is a pass to turn f(recur(x), y) into recur(f(y, acc), x) and then tail-call optimization can be applied. This works for any associative f.