[MLton] cvs commit: sped up output1 a lot

[Stephen Weeks]

> Then why is the C program beating us by so much? 

Just to be clear, we are talking about comparable speed on Linux and a
roughly 1.5X slowdown vs C on FreeBSD and NetBSD, right?  And this is
a very small hot loop.  That kind of difference, while not noise, is
easily attributable to an extra test or different code generation.  If
someone really wants to try to cut things down more, the place to
start is to look at the SSA and then on to the assembly.  To my eyes,
though, we're already close enough that there are better places to
spend time.