Pythran stories

Unified builds (also know as Jumbo Builds) is a build techniques that aims at improving build time through the concatenation of several sources as a single unified source before compilation.

This is the story of a bug hunt that lasted much longer than expected, but ended with the dearest of all treasures: knowledge.

-ftrivial-auto-var-init=[pattern|zero|uninitialized] is a compiler flag that controls how the stack is initialized before any value is initialized to it.

#include <iostream> #include <unordered_set> #include <string> const std::unordered_set<std::string> primary_colors = {"red", "green", "blue"}; int main(int argc, char const ** argv) { char const * color = argc > 1 ?

Recently, I've submitted a patch that shaves ~2.5MB on one of the core Firefox library, libxul.so.

Unity Build Unity builds, a.k.a. unified build or jumbo builds, is a technique that consists in concatenating several C or C++ files in one before invoking the compiler.

Project Overview There are a lot of algorithms in SciPy that use Cython to improve the performance of code that would be too slow as pure Python, e.g. algorithms in scipy.spatial, scipy.stats and scipy.optimize.

Note: I'm publishing this article the 26th of March, 2023, because… I forgot to do this when I wrote it some two years ago :-) xsimd is a header-only C++11 library that provides an efficient abstraction for Single Instruction Multiple Data (SIMD) programm…

Fedora rawhide has been moving to Python 3.9, which broke the Pythran package.

Not every story begins with an issue, but this one does.

Pythran 0.9.4 just got released, and it has an unusual number of unexpected features.

In spite of its dynamic nature, the Python language has been granted a rich ecosystem of Abstract Syntax Tree(AST) analysis and transformations.

If you're curious or genuinely interested into how Pythran transforms your code, but not brave enough to dive into the generated C++ code, Pythran provides a compilation switch to dump the refined Python code, after optimization and before it gets transla…

Hey folks, it's been a long time since I last did a post to celebrate a release.

Introduction As a researcher/engineer (really, an algorithm developer) in the general area of audio and speech processing, I've always run into the same difficulty at all the companies I've worked at: How to quickly prototype and develop algorithms, and s…

Once upon a time, on IRC, Serge Guelton asked me whether I wanted to work on having Cython using Pythran for Numpy-related computation.

boost.simd provides a C++ abstraction of vector type, allowing for efficient vectorization of array computations.

As many projects, Pythran has its own test suite that runs on travis-ci and AppVeyor.

This blogpost originally was a Jupyter Notebook.

This blogpost originally was a Jupyter Notebook.

Testbed So the question was: do Pythran-generated binaries use more disk space than Cython ones?

This blogpost originally was a Jupyter Notebook.

This post is not about the famous Hoi-Poi Capsule but about a feature I recently discovered from Python: PyCapsule.

This blogpost originally was a Jupyter Notebook.

Over the last six months, I've been working on improving Pythran for the OpenDreamKit project.

Pythran is currently part of OpenDreamKit, a project that aims at improving the open source computational mathematics ecosystem.

Pythran presentation hold on Université Lyon1 on Monday 27th June together with a Julia presentation, and later at Compas'2016.

Pythran presentation hold on Université Lyon1 on Monday 27th June together with a Julia presentation.

The Pythran team (still 2 active developers) is delighted to announce the release of Pythran 0.7.5, available on the traditional channels: pypi: https://pypi.python.org/pypi/pythran conda: https://anaconda.org/serge-sans-paille/pythran github: https://git…

This blogpost originally was a Jupyter Notebook.

This blogpost originally was a Jupyter Notebook.

Foreword This is not a Jupyter notebook, but it could have been.

The benchmark First there was a paper [0], in the paper there was a github [1] and in the github [1] some benchmarks.

When Size Matters Everything started a few days ago with a Pythran user complaining about the size of the binaries generated by Pythran.

While hanging on Stackoverflow (everybody does this, no?)

The pythran team (a great total of 2 active developers) is delighted to announce the release of Pythran 0.7.4, available on the traditional channels: pypi: https://pypi.python.org/pypi/pythran conda: https://anaconda.org/serge-sans-paille/pythran github:…