More benchmarks for each library

[jonahsnider]

Right now benchmarks only test a single conversion operation of x amount of y converted to z. This is a common use case which is why I chose it, but if possible there should be lots of other common use cases to get a broader collection of data from libraries (ex. converting bigints).

Doing this would require rewriting a lot of code to expect several test functions for each library that are all run as their own benchmarks and then turning that into reasonable markdown output.

[jonahsnider]

Closed in 15ab9c5