Allow generating graphs from benchmarks

[Konrad `ktoso` Malawski]

Ass seen in scala/scala in scala/collection-strawman@848e8b6 by @julienrf

[Julien Richard-Foy]

As explained in the collection-strawman discussion, I’m afraid that the charts are too specific to our benchmarks. I’m not sure how we could generalize them. Examples of decisions that are specific to our project:

• by convention, benchmarks for different collection implementations (e.g. List vs Array) but exercising the same feature (e.g. adding an element) must have the same name ;
• benchmarks must all be parameterized by a size parameter ;
• the x axis uses a log scale.

[Konrad `ktoso` Malawski]

Sounds perfectly doable via 2 or 3 params in config :-)
I'd love to have this available for everyone, esp since that's exactly the style of benchmark that is often out there - how perf changes when n changes.