rml, 
"Using the compiled #scheme code is around 470x faster than our current interpreter, at least for this simple arithmetic loop."
The #Unison project on replacing their compiler backend with #Chez scheme
https://www.unison-lang.org/whats-new/jit-announce/
