在多处理中与Pool结合使用时，map如何划分数据？

I have a function f that I want to compute on a certain large data in parallel. The data can be divided in many ways and I am trying to make a decision on how to divide it. I am trying to understand how "map" in multiprocessing.Pool distribute/divide the data exactly so that I make the right decision about splitting my data as well as choosing the number of processors. My input data is not simply a list, as in the below example, but rather list of dictionaries and a list of lists so understanding how Pool.map divides the data seems critical. I have a function f that I want to compute on a