Performance understanding

[willscott]

there should be a basic unit-test benchmark for the library to let us have confident that the current rw-lock structure isn't leading to bad lock contentions

[aarshkshah1992]

@tchardin is looking into writing benchmarks and metrics collection.