sort:

This example is work in progress. The goal is to speedup sorting a 
large array of strings by using several cores.

sort1: This version fails completely. I left it here because it has
       an interesting message-passing interface.

sort2: This version works, but the multi-core program is only faster
       for giant arrays. It is using shared memory for fast data
       transfer between processes.

When I find time, I am going to develop a third version that
uses camlboxes for data transfer. This is way more systematic than
what is done in sort2.