0

I have been reading (a lot) about recommendations to use Multithreading (GIL-restricted) for I/O bound tasks and use Multiprocessing for CPU bound tasks. I understand that CPU bound tasks are which involve heavy computations, whereas Network R/W or Disk R/W would fall under I/O bound tasks.

I am not completely convinced as to which technique would suite for task, where I have to create multiple raw socket connections and make multiple clients perform exact sequence of R/W operations with the server, and then eventually map all the data from multiple R/W operations onto a common structure to later just read it.

In order to parallelize the task above, I guess I CAN use either techniques? But I would like to understand which one I SHOULD use in terms of parallelizing the network R/W operation.

Thanks in advance.

eecs
  • 141
  • 2
  • 13
  • 1
    Possible duplicate of [Multiprocessing vs Threading Python](http://stackoverflow.com/questions/3044580/multiprocessing-vs-threading-python) – Ozgur Vatansever Sep 12 '16 at 04:56
  • Thanks, that was the 1st SO post I read (re-read) to understand the concept. But it is unclear to me in terms of Parallel Socket programming which technique is suitable. – eecs Sep 12 '16 at 05:06

0 Answers0