[MLton-user] Some numbers on SML/NJ & MLton runs and also a few questions

shivers@cc.gatech.edu shivers@cc.gatech.edu
Tue, 14 Oct 2003 11:16:00 -0400


Stephen-

I'm trying to get some MLton runs on a SPARC, and deal with this
problem relating to the illegal -mno-epilogue flag.

Your suggestions were:
    > 1. Recompile MLton, taking out the bad switch.  It should be easy to
    > find in main.sml.
    > 
    > 2. Build your own package or ask me to build an experimental one from
    > the MLton CVS.  Then you can modify the mlton script.
    >
    > 3. Switch to a gcc that supports -mno-epilogue.  I used
    > binutils-2.11.2-sol7-sparc-local and gcc-2.95.3-sol8-sparc-local from
    > sunfreeware.com.

I started down path #1. But having read up on cross-compilation, it looks
pretty hairy, in that I have to install a mess of gcc cross-comp tools, 
with which I have no experience.

Could I ask one of you guys to build a SPARC rpm or tgz or just executable
with the bogus flag excised? This would be a fine thing to put on your web
page for people other than myself, as well...

BTW, an update on my code measurements. The reason the side-effecting sort
timings were so fast (1000x) is that I had a bug in the timer code for
those runs. (It's necessarily different from the timer code for the
pure algos.) I wrote #sys where I needed #usr, and was measuring kernel
time, which is nothing! Moral: the type systems doesn't *always* find *all*
your bugs...

The side effecting code is still really fast, but it's more like 2x, instead
of 1000x. I'll send along more times to the MLton list for SMLNJ/MLton
comparison purposes when I have put it all together.
    -Olin