Distribute v1.1

This is a simple perl script which I wrote for use on our network, and
thought other people might be interested in as well. It is used to
dispatch jobs to multiple machines, when the order of the jobs is not
important but you want to make efficient use of the computing
resources.

The program requires ssh, rsh, or some equivalent permission on the
remote machines. It configures itself as a server, and then spawns a
client on each remote machine (possibly including the machine the
server is running on). Each remote client is configured to run a set
number of parallel jobs, and when the number of jobs running is lower
than that number, it contacts the server for another task.

The program is useful in situations where a large number of tasks must
be performed, but it is difficult to predict how many tasks any one
machine can perform in a set time, so dividing up the tasks ahead of
time may result in several machines sitting idle while others are only
part-way through their list of assigned jobs.

As of version 1.1, a client program, dist_submit, for submitting
processes to the distribute script allows easy distribution of
compilations via "make" across multiple networked machines.

See the manual page for more details.

The script requires Perl 5 with the IO::Select package, available
separetely from CPAN sites, or bundled with Perl starting with
v5.0 patchlevel 4 subversion 3.
