The Python Compiler
summary
Nuitka presents itself as a statically optimizing Python compiler that compiles Python to C++ and aims for full language parity. It implements variable value propagation at local and module scope, function in-lining (recently completed on the development branch), constant propagation and exception-focused optimizations, and it generates C++ files for g++/MSVC. It passes the CPython test suite for 3.4/3.5, handles extension modules (PyQt just works), shows ~2.5× on a Python benchmark and up to ~20× on tight call-heavy microcases, and exposes a plugin framework for standalone builds and type-hinting.
AI-written summary · about
chapters (13)
- 0:00Start
- 0:58The Python Compiler
- 9:49Constant Propagation
- 10:30Interfacing with C Code
- 12:36Reformulations
- 13:30Optimization
- 14:21Demo
- 17:47Xml Representation of the Node Tree
- 20:28Generated Code
- 21:53C Code
- 23:17Parser Optimization
- 24:47Testing for Correctness
- 25:39Questions
related talks
01h 02mWriting Python Bindings for C++ Libraries: Easy-to-use Performance
29mBuilding a JIT compiler for CPython
29mThe Art of Extending Python with other languages
33mPythonizing Qt
01h 33mPython and C++: the best of both worlds
32mNim for Pythonistas
25mHow to bring your C++ project to Python land
29mWhat they don't tell you about building a JIT compiler for CPython