PyCon X

firenze

2-5 maggio 2019

Sharing is Caring! - how to free the CPU from message passing hell - shared memory saved the day

When dealing with a real time application in which you are required to process a lot of data from multiple sources in parallel, system resources disappear quickly. Add some CPU intensive processing and you are set to fail.

This talk will illustrate a solution to solve a CPU bound problem with a data stream of tens of megabytes. I will talk about how easy it is to fail and how to obtain a working solution.

The following arguments will be discussed:

  • threading and its limitations
  • multiprocessing and its limitations
  • serialization
  • shared memory
  • a solution: multiprocessing + shared memory

Code will be provided and shown during the talk.

in on domenica 5 maggio at 11:45 See schedule

Do you have some questions on this talk?

Nuovo commento