/ website / blog

attempts to make python fast

October 20, 2020

Posted on Hacker News there was an Implementation Plan for making CPython (the official Python implementation) faster. The author claims a 5x speedup is possible for the low cost of $2 million USD.

The four step plan includes

We have witnessed other attempts at making Python fast, each achieving their own degree of success in terms of performance and compatibility. For posterity I started keeping a list of them here, in no particular order.

Seq

Seq’s performance is usually comparable to that of C or C++, and can often be even better once domain-specific compiler optimizations are applied.

Pyston 2

[Pyston] version 2 is 20% faster than stock Python 3.8 on our macrobenchmarks.

Pyston

Pyston is a performance-oriented Python implementation built using LLVM and modern JIT techniques.

Unladen Swallow

An optimization branch of CPython, intended to be fully compatible and significantly faster.

Stackless Python

Stackless Python is an enhanced version of the Python programming language. It allows programmers to reap the benefits of thread-based programming without the performance and complexity problems associated with conventional threads.

PyPy

A fast, compliant alternative implementation of Python.

Jython

Jython is approximately as fast as CPython–sometimes faster, sometimes slower. Because most JVMs–certainly the fastest ones–do long running, hot code will run faster over time.

HotPy

The HotPy virtual machine is a high-performance virtual machine for Python.

Iron Python

Performance is comparable to CPython - much faster for some things … but slower for other things.

Psyco

Psyco is a Python extension module which can greatly speed up the execution of any Python code.

2c-python

Using the generated binary code gives a speed boost from 2 to 4.5 times.

Cython

Easily tune readable Python code into plain C performance by adding static type declarations.

Nuitka

Nuitka is more than 2 times faster than CPython …

Pyc

Pyc is a python compiler intended for high performance computing and programming-in-the-large

Shedskin

For a set of 75 non-trivial programs …, measurements show a typical speedup of 2-200 times over CPython.

Numba

Numba makes Python code fast

Parakeet

Parakeet was a runtime accelerator for an array-oriented subset of Python.

Cannoli

Cannoli is a compiler for a subset of Python 3.6.5 and is designed to evaluate the language features of Python that negatively impact performance.

gl;hf!

➡️ related posts in the observations series ...

No follow up posts yet. Check back soon!