Add beginner info how to extend fit jobs to parallel and distributed resources

[cdeil]

I'm new to mystic and mpi, but I do have fit jobs that I currently run on a single core and that I'd like to run faster.

Could you write some beginner-level info how to get started with running optimizations using multi-core and CPU clusters? (This is one of the major features of mystic listed here, but as far as I can see there's no tutorials on how to do this here.)

What I'm looking for are instructions what to install, a simple example to run and maybe how to measure how it scales with n_cores or n_cpu for my setup.

[mmckerns]

Correct, I haven't added a "tutorial" how to do that, and there should be one. I do have a collection of examples that extend the tutorial examples, however… and with a little bit of cleaning, a selection from these will be the tutorial.

For now, see: https://github.com/uqfoundation/mystic/blob/master/tests/test_solver_constraints.py https://github.com/uqfoundation/mystic/blob/master/examples/mpmap_desolve_rosen.py https://github.com/uqfoundation/mystic/blob/master/examples/TEST_ffitPP.py and similar.

Also see: https://github.com/uqfoundation/mystic/tree/master/examples_UQ https://github.com/uqfoundation/pathos/tree/master/examples2