Add a page with some notes on reproducibility

[PGijsbers]

At a later point:

• add concrete examples that people can follow to reproduce a specific experiment
• provide an example script that parses information from the results.csv file to queue benchmark jobs (or, even better, properly integrate it in runbenchmark.py)