This is my favorite algorithm ever, and the first serious one that I tried (and failed) to implement in my life when I was a teen. I remember dearly implementing the first three algorithms (A, S, M) in the nasm assembler on my debian "hamm". Then I tried for many weeks to get D right, but never managed to be sure all the cases worked.

These algorithms are probably easier to implement in assembly than in a higher-level language.