What is the Global Interpreter Lock (GIL)?

Question

Accepted Answer

The **GIL** is a mutex in **CPython** (the reference interpreter) that lets **only one thread execute Python bytecode at a time**. It exists because CPython's memory management (reference counting) isn't thread-safe; the GIL makes object access safe without per-object locks, and keeps single-threaded code fast.

Consequences:

- **CPU-bound** multithreading gets **no speedup** — threads take turns, so two threads finish about as fast as one (plus switching overhead). Use **`multiprocessing`** (separate processes, each with its own interpreter and GIL) for real parallelism.
- **I/O-bound** multithreading **does** help: the GIL is released during blocking I/O (sockets, disk, `time.sleep`), so other threads run while one waits.
- C extensions (NumPy, etc.) can release the GIL around heavy native work, regaining parallelism.

The GIL is a CPython implementation detail — Jython/IronPython don't have one, and Python 3.13 ships an experimental **free-threaded (no-GIL)** build.

**Interview ti

What is the Global Interpreter Lock (GIL)?

Understand the problem

Attempt it yourself

Stuck? AI Nudge Available

Study the solution

The solution is waiting

Run the code

Join the discussion

Discussion (0)