add gcd proc example

[proppy]

As suggested by @grebe in #935

I think gcd is a good candidate for proc implementations, which would let you do the "early return" I think you're describing. Each proc tick would do one division (or gcd_binary_match), and you send the output when you reach the stop condition.

This would also be useful to show comparison and tradeoff between the proc and the automatically pipelined version.