Python Multi Processing

less than 1 minute read


I have developed a basic Python Library to provide abstraction to parallel programming. This library is based on in built-in library multiprocessing in python and 3rd party library ray. In this blog post I will explain in details about the implementations in addition to existing documentations.

CustomMP is the abstraction of multiprocessing library with a SharedList. SharedList is a generalization of commonly used Manager, pair of Queue to keep track of task and results. SharedList implementation take care of this, thus avoid redundant implementation of same multiprocessing structure each and every time of implementation. Motivation for this implementation is based on the discussion I had in stackoverflow regarding proper way to implement SharedList.

Sample Use Case of CustomMP

from pyparallel.CustomMP import CMPSystem

limit = 4
args = (1,)
cmp_sys = CMPSystem(limit)
cmp_sys.add_proc(func=child_func, args=(args,))
contents =