Add support for running only specified subbenchmarks #54
Description
Currently the run wrapper only knows how to kick the whole pyperformance test suite going, which should remain the default behaviour, but in some instances (such as testing changes) this is unnecessary.
We should support supplying benchmark names to be passed to pyperformance (using their -b flag). This way we can quickly test to make sure everything works (at least targeting the most problematic benchmarks).