Team PLClub ICFP entry -- comparing the performance of OCAML and SML

[John H. Reppy]

I'm dropping off comp.lang.ml, since this is getting pretty narrow in
its interest.

The chess input is probably the best test case to judge overall performance,
since it stresses most of the GML specification.  Fractal, on the other
hand, is a Tier-1 test that has a single plane and lots of spheres, so
it stresses the sphere intersection code (and not much else).  Thus
the sqrt bottleneck could be very important.  In the actual application,
there may be I-cache effects that make the SML version of sin slower;
also the range of inputs might require more computation.  How did you
determine your calls per second number?

	- John