After much discussion, the Python Steering Council plans to adopt PEP 703, “Making the Global Interpreter Lock Optional in CPython.”
The removal of Python’s Global Interpter Lock, or GIL, has undergone numerous attempts over the years, culminating in this proposal. The removal of the GIL makes Python a true multi-core language and significantly boosts its performance for workloads that benefit from parallelism by removing a significant barrier to multi-threading. This proposal brings first-class multithreading and concurrency support in Python one step closer to reality.
The current suggestions for a no-GIL version of Python employ a variety of methods to make reference counting thread-safe, with little to no impact on the speed of single-threaded programs.