Multiprocessing and saving intermediate results with multiple, large object arguments

Question

I am trying to utilize multiprocessing to speed up a project. I have one large dataframe that gets produced at the parent process level, then sent into to one function (usin