python Python PEPs ·

CPython experimental JIT compiler matures for Python 3.15

infra
feature patch

CPython's experimental Just-in-Time (JIT) compiler, initially designed in PEP 744 and present since Python 3.13, has been re-architected and significantly improved. For Python 3.15, it offers a 4-12% performance increase, improved native debugger support, and reduced memory usage for generated code. These advancements provide valuable insights into developing a production-ready JIT for CPython.

  • CPython JIT compiler delivers performance gains and improved debugging in Python 3.15
  • JIT compiler's development continues based on CPython implementation learnings
Enhancements (1)
  • CPython JIT compiler delivers performance gains and improved debugging in Python 3.15

    The experimental JIT compiler, now part of CPython's main branch since version 3.13, has been re-architected and matured. In Python 3.15, it provides a 4-12% geometric mean performance improvement and enables native debuggers to unwind frames, while also reducing the memory footprint of generated code.

Notes (1)
  • JIT compiler's development continues based on CPython implementation learnings

    The JIT compiler's design and implementation have benefited from significant learning and re-architecture since its initial conception. These ongoing efforts aim to establish a viable and supported JIT compiler for CPython.

Read the original announcement →

https://peps.python.org/pep-0836/