Provide a detailed description of each benchmark

[lbulej]

We should provide a detailed high-level description of each benchmark, so that a reader can get a quick understanding of what a benchmark is doing with what data. For benchmarks that we actually implement locally, there is always at least the code, but a high-level description would be nice anyway.

However, for some benchmarks, we just have a bit of Scala code that calls into a big library with some data and that's it. This occurred to me when looking at the code of the Alternating Least Squares benchmark, which is exactly the case (it calls into Spark Machine Learning library).

In addition to what the benchmark is doing, we should also document what kind of parallelism it uses (if any).

[axel22]

Should we add a detailedDescription method to the RenaissanceBenchmark class?

[lbulej]

Well, being able to plug the detailed description into a generated document should get us extra points :-)

[lbulej]

Just to keep track of progress, #135 introduces the @Description annotation which should be used for this purpose. Note that the one-line summary should go to @Summary.