Experiment with ways of efficient parallel processing of a text stream in perl.
The goal is to replace zcat ... | sort .. | .. bla-bla .. | cpu-intensive-slow-code.pl | .. bla-bla > out with zcat ... | sort .. | .. bla-bla .. | cpu-intensive-but-parallel-so-N-times-faster-code.pl | .. bla-bla > out (of course Im talking about wallclock time speedup, not cpu time) by wrapping payload of cpu-intensive-slow-code.pl with parallelization wrapper with small overhead.
perl-sysread.pl works and perl-piped.pl is broken