updated to allow for fine-grained parallelism via OpenMP.

[quattro]

The total number of threads can be controlled by a new program argument "-n". The default value is the
total number of threads available to the system. This can easily be changed by altering the assignment
at the top of the main.cpp file.

For comparison, when running on the sample data with 1 thread:

real 0m23.472s
user 0m19.635s
sys 0m0.024s

and with 8 threads:

real 0m6.474s
user 0m26.259s
sys 0m0.025s

The important metric is the "real" time. The user time should be about the same, as it is the summation
across all threads. The final estimates are exactly the same (we should do some more complicated
examples; however, this looks very promising for now).

[theboocock]

This looks great, I will give it a test.