How Python’s Next Release Plans to Eliminate the GIL and Boost Speed

The next CPython release, slated for October, promises major performance gains and the ability to run inside browsers.

At the first Python language summit since 2019 in Salt Lake City, the core development team outlined upcoming features for Python 3.12, making its roadmap clearer.

While Python has several VM implementations (JVM, .NET CLR), the official interpreter remains CPython, which still suffers from the Global Interpreter Lock (GIL) that limits multi‑core utilization.

Historically, multiple interpreters in a single process had to share the same GIL, causing interference; developer Eric Snow proposes giving each interpreter its own GIL.

A broader effort aims to remove the GIL entirely. An earlier “GIL slicing” project stalled, but Meta’s Sam Gross has introduced a new “nogil” experiment that may become viable.

Meta’s Instagram runs Python on an internal fork called Cinder, designed for Facebook’s needs; the new work seeks broader applicability.

Python’s recent resurgence includes Guido van Rossum’s return from retirement to join Microsoft as chief engineer, collaborating with Eric Snow and Mark Shannon to extend existing ideas.

Plans include using inline caches to speed up the bytecode interpreter—a feature already in Python 3.10—and future enhancements to memory management, allocation, and the performance of binary operators and integers.

Backed by Microsoft, Bloomberg and others, the internal “HotPy” project is making progress, and CPython 3.11 beta already shows a 25 % benchmark improvement over the previous version.