Articles

DRAFT: From PyParallel to Python Free-Threading

Optimally Exploiting Multiple Cores with Python

Python

PyParallel

Free-threading

No-GIL

PyTorch

asyncio

This article takes a look at the new no-GIL, free-threading functionality introduced in Python 3.13, as well as how it compares and performs against PyParallel’s attempt at explointing multiple CPU cores over ten years ago. Using a demo project named Parallelopedia, we demonstrate the benefits afforded by the new functionality, whereby large data structures (tens to hundreds of gigabytes) are loaded in main memory and then accessed in parallel from Python—a feat not previously possible with the contemporary multiprocessing approach. Additionally, we investigate multi-threaded, parallel inference (generation) of transformer-based Deep Neural Networks via PyTorch, and explorer the limits of what can and can’t currently be done, as well as a roadmap for future PyTorch support.

Is Prefix Of String In Table?

A Journey Into SIMD String Processing

AVX2

SIMD

C

Assembly

MASM

This article details an approach for efficiently determining if a given string prefix-matches a set of known strings. That is, do any of the known strings represent the prefix of a given string? A custom data structure is employed with successive implementations benchmarked to find the fastest possible solution.