serge-sans-paille/pythran

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

2141 stars · 202 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Pythran is an ahead-of-time compiler that translates a subset of Python, focused on scientific computing with NumPy, into fast native C++ extension modules. It exploits multi-core parallelism and SIMD instructions to accelerate numeric kernels while preserving the original Python module interface.

Use cases

• speed up numpy-heavy python code
• compile python numeric kernels to native extensions
• accelerate scientific computing functions
• vectorize python loops with SIMD
• parallelize python computations across cores
• build fast python modules without writing C++

When to choose

• you have numeric/scientific Python code that is too slow
• you want native speed without hand-writing C++ or Cython
• your code uses NumPy expressions and pure-Python numeric kernels
• you want automatic SIMD and multi-core optimization

When to avoid

• your code relies on arbitrary Python objects, dynamic features, or non-numeric libraries
• you need JIT compilation or interactive acceleration of full Python programs
• you target platforms with weak C++ toolchain support

Facets

cli-tool · maturity active

compiler machine-learning math compilers performance python windows cpp ahead-of-time-compilation numpy simd python-to-c++ numeric-kernels compiler scientific-computing linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="serge-sans-paille/pythran")