python/pyperformance

Pyperformance is a benchmark suite for Python, a standardized set of tests that measures how fast Python code runs. It exists so that anyone working on Python itself (or building alternative versions of Python) can check whether their changes make the language faster, slower, or about the same. The suite focuses on real-world workloads rather than synthetic tests. That means it runs actual applications and common tasks to see how Python handles them, instead of using artificial loops designed just to stress the system. The goal is to give a trustworthy, apples-to-apples comparison across different Python implementations, so a team can say with confidence whether their new version is genuinely faster in practice. The people who would use this are mostly Python core developers and anyone building or maintaining an alternative Python runtime (like CPython, PyPy, or others). For example, if a developer rewrites part of Python's internals to speed up function calls, they can run the benchmark before and after to see if real-world programs actually got faster. It is less relevant to everyday app developers, who typically care about their own code's performance rather than the language engine itself. One notable caveat: the project states it is not yet tuned for PyPy, an alternative Python implementation known for speed. If you want to benchmark PyPy specifically, the README points you to a separate benchmarks project for that. Beyond that, the README is fairly sparse, it does not list which specific benchmarks are included or how long a typical run takes, so you would need to dig into the documentation for those details.