Python GC Settings - Change This and Go 20% Faster!

Python's default memory settings work well for many applications. But if your app pulls in a lot of data, such as a database heavy website, API, or data science explorer, they may be doing a lot of harm to your app's performance with nearly zero benefit.

Python has two memory models: reference counting and garbage collection (GC). Ref counting handles well into the 99% of your memory cleanup needs. But GC is omnipresent too. You'll see some tips to adjust that second level of memory management that can really benefit apps that create many objects and container types.

▬ Full Memory Course ▬▬▬▬▬

If you found this content helpful, then please check out my full Python Memory Tips and Tricks course at Talk Python. It has over 4 hours of tips and techniques like these show here:

https://talkpython.fm/mem

▬ Chapters ▬▬▬▬▬▬▬▬▬▬▬▬

0:00 - Why mess with the GC?
0:54 - How does Python's GC work?
1:59 - What triggers a GC cycle?
3:06 - Check out our memory using app
4:44 - Run the app without modifications
5:38 - Can we do better? Change the GC
7:35 - Is it faster?
9:04 - Using GC debugging to report every GC cycle
10:40 - Locking out start up memory with gc.freeze()
12:09 - Should you tweak the GC in your app?
13:01 - Wrap up

▬ Code from this video ▬▬▬▬▬▬▬▬▬▬▬▬

The code shown in the video is hosted on GitHub here:

https://github.com/mikeckennedy/pytho...

▬ More Videos ▬▬▬▬▬▬▬▬▬▬▬▬

Subscribe and get more videos by me at @mikeckennedy to see all the topics I've covered.

▬ Talk Python Courses ▬▬▬▬▬▬▬▬▬▬▬▬

If you're looking to take any of our full length courses, just visit https://training.talkpython.fm/courses You can even play with the final search over there on the live website.

▬ Dive deeper ▬▬▬▬▬▬▬▬▬▬▬▬

Listen to the Talk Python To Me podcast at https://talkpython.fm
Check out our 100's of hours of Python courses at https://training.talkpython.fm/courses
Follow Michael on Twitter at   / mkennedy