colesbury/pyperformance

This is a collection of performance tests designed to measure how fast Python runs. Instead of artificial tests that only measure one thing in isolation, pyperformance focuses on real-world scenarios, like running actual applications or common programming tasks, to see how different versions or implementations of Python perform in practice. Think of it like a standardized speedway where you can test different cars under the same conditions. A developer or team maintaining Python (or an alternative Python implementation) can run these benchmarks to see if changes they made actually made the language faster or slower. It's a fair, reproducible way to measure performance across different setups and Python versions. The tool is meant to be the go-to source that the entire Python community trusts. Because Python has multiple implementations (CPython is the most common, but there's also PyPy, Jython, and others), having one agreed-upon set of benchmarks helps everyone compare fairly. You can install pyperformance as a package and run it on your machine to get a baseline of how your Python setup performs, or use it as part of continuous testing to catch performance regressions before they ship. Someone working on the Python runtime itself, a performance engineer optimizing Python for their company, or a researcher comparing Python implementations would all use this suite. The README doesn't detail what specific benchmarks are included, but the philosophy is clear: real applications matter more than toy problems when it comes to understanding real-world speed.