What about http://docs.racket-lang.org/math/ ?
In contrast, SBCL uses tagless float arrays, so that array operations tend to be cache friendly. With adequate declarations, numerical code in SBCL tends to very fast (~1-3x C speed). I assume one can do the same with Bigloo and Chicken scheme as well.
http://lists.racket-lang.org/users/archive/2013-April/057487...
This is also the case in Racket.
http://docs.racket-lang.org/reference/flonums.html#%28part._...