Benchmarking suite

[mmatous]

While everyone can cook up their own files, it would be nice to have a uniform way of generating data for proper comparison in repo. Crates I found that could help: test-data-generation or regex_generate.

[dufferzafar]

Whenver you do this, please also keep https://github.com/koraa/huniq/ in mind too!

ref: koraa/huniq#13

[taggart]

I like this idea, also a couple things to add:

• provide data sets with varying percentages/numbers of duplicates
• test the various tools with/without locales (supposedly normal sort is faster(but not always correct?) with LC_ALL=C)