Refactor the plotting part to enable parallelization

[OmerRosler]

If instead of a loop on the generator we can turn this into a transform call, then we would be able to parallelize the algorithm.

[OmerRosler]

Plotting is now basically a for_each, unfortunately due to the algorithm's manual memory management C++17's parallelism won't do (see SO question)
Instead we will use classical concurrency.