afaik, it's the only cpu miner in assembly (and one of the only which is not a rip of Wolf0) so it's to be the fastest. i however got reported being slower on some Intel i7 or Xeon. i don't own one so i cannot optimize well yet.
That would BTC the key to many systems or applications, optimizations are quite heavy when tuned right && having spent 10,000 hours over the years on just playing around with the basic hardware on all algorithms (configurations files). Still love to setup & test hardware, even if it's not going to earn many coins Kinda of a refresher course every week, so funny that people still think only sha256 or Scrypt are the only ones