Concurrency
using Python
Understanding the basic concepts
of
$whoami: Shashank
Writes software for Nykaa.com (Nykaa Display Network)
Open-source enthusiast, works mostly on Python, JS & related technologies
https://shashankaryan.tech | aryanshashank31@gmail.com
Contact Details
Contents:
- Concurrency vs Parallelism - thread, task and process
- How to use Concurrency - I/O Bound vs CPU Bound Programs - Demo
- Conclusion - When to use Concurrency?
Note: This talk doesn't include any specificity about the libraries used
Concurrency v/s Parallelism
clever switching of tasks on a single core
performing computations using multiple cores
Concept of thread, task and process
Thread - threading| Pre-emptive multitasking | Single core of CPU
Task - asyncIO | Cooperative Multitasking | Single core of CPU
Process - multiprocessing - Multiple cores of CPU
How to use concurrency?
Use Case 1: I/O Bound Programs
- spends most of its time interacting with a I/O endpoint, like a network connection, a hard drive, etc.
- speeding involves overlapping of waiting time, spent for such interaction
DEMO - A program that gets multiple URLs and prints their length.
Synchronous version
threading Version
asyncIO Version
multiprocessing Version
Use Case 2: CPU Bound Programs
- speeding involves more computation with the same time duration
- spends most of its time performing CPU computations.
DEMO - A program that computes the sum of items in a very large list.
Synchronous version
What will be the result if we apply threading or asyncIO versions for the CPU bound process?
Question!
multiprocessing Version
Conclusion: When to use Concurrency?
"Premature optimization is the root of all evil." - Donald Knuth
Questions/Suggestions/Discussion
Thank You :)
Concurrency 101
By Shashank Aryan
