Let the application run parallel / distributed

[frbl]

Currently we can run the program locally, but it can only scale to max 8 cores (since we only have 8 cores).

[frbl]

Alternative 1
If we would have an easy to use docker image, we could run the entire project on one of the 32 core machines at amazon.

[frbl]

Alternative 2
Alternatively, we could make the code runnable on the RUG supercomputer, but this needs the application to run massively parallel. One option for doing so would be to write all model configurations to disk, and run several workers that each pick up one of these configurations to proces.. This requires a lot of programming