Join us and get access to thousands of tutorials and a community of expert Pythonistas.

Unlock This Lesson

This lesson is for members only. Join us and get access to thousands of tutorials and a community of expert Pythonistas.

Unlock This Lesson

Hint: You can adjust the default video playback speed in your account settings.
Hint: You can set your subtitle preferences in your account settings.
Sorry! Looks like there’s an issue with video playback 🙁 This might be due to a temporary outage or because of a configuration issue with your browser. Please refer to our video player troubleshooting guide for assistance.

The Global Interpreter Lock (GIL)

Austin Cepalia
How Python Manages Memory Austin Cepalia 02:01
Recommended Tutorial Course Slides (.pdf) Ask a Question

00:00 CPython’s approach to thread safety is to use what is called the Global Interpreter Lock, or GIL. Rather than use a bunch of locks on individual shared resources, which would be very hard for compatibility and readability reasons, the GIL locks the entire Python interpreter.

00:20 Remember, that’s the part of CPython that reads in the bytecode and carries out the associated operation. Because the entire interpreter is locked, instead of a specific shared resource, each thread must obtain the interpreter lock to interpret some bytecode.

00:37 And because only one thread can have the lock at any given time, this essentially makes CPU-bound code—trying to perform a computation with multiple threads—singly-threaded. However, threading does have uses in other places of Python, like user interfaces and IO, but that’s not the focus here.

00:58 This performance limitation on CPU-bound code makes the GIL particularly infamous in the Python community. It’s a simple approach that allows Python to be both readable and compatible with old C extension modules, but at the cost of performance.

01:16 Lots of work has been done to come up with a better method of ensuring thread safety, but none have been able to outperform the GIL as of the time of this recording.

01:28 The bottom line is this: memory is a shared resource. We can only have one process or one thread accessing that block of memory at any given time or else unpredictable behavior might occur.

01:42 Protecting shared resources with mutexes is not easy, but it’s an essential part of memory management. Luckily for us, the GIL takes care of all of that.

01:52 If you’d like to learn more about the GIL, take a look at the article I’ve linked down in the video notes below.

antchal on March 12, 2021

Where is the article link for GIL that you mentioned in the video?

Bartosz Zaczyński RP Team on March 12, 2021

@antchal It’s in the dropdown labeled “Supporting Material” just below the course title.

antchal on March 14, 2021

I was aware of the possibility of finding it in the corresponding article. But I thought, since I’m not reading an article here maybe RP attaches the URL at a more readily accesible place.

Thanks for clearing that doubt tough.

Become a Member to join the conversation.