Add benchmarks to compare performance against other JVM tabular processing libraries

[twilco]

It would be interesting and useful to have benchmarks of Tabitha's performance versus other JVM-based tabular processing libraries. We should measure both memory consumption (average, peak, more?) and processing speed. Could this be automated somehow?

Possible libraries to compare against might include...

CSV/TSV:

fastcsv
flatpack
commons-csv
supercsv

XLS/XLSX:

POD
excel-streaming-reader (only works with XLSX files)
zerocell

There are lots of benchmark tools out there, such as JMH. We should explore the options.