https://github.com/nicholasjng/mew

A microbenchmarking CLI for Python.
https://github.com/nicholasjng/mew

benchmark google-benchmark nanobind python

Last synced: about 1 month ago
JSON representation

A microbenchmarking CLI for Python.

• Host: GitHub
• URL: https://github.com/nicholasjng/mew
• Owner: nicholasjng
• License: apache-2.0
• Created: 2026-05-13T11:10:58.000Z (2 months ago)
• Default Branch: master
• Last Pushed: 2026-06-16T12:07:14.000Z (about 1 month ago)
• Last Synced: 2026-06-16T13:26:26.494Z (about 1 month ago)
• Topics: benchmark, google-benchmark, nanobind, python
• Language: Python
• Homepage: https://mew.readthedocs.io
• Size: 600 KB
• Stars: 0
• Watchers: 0
• Forks: 0
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE
  • Roadmap: ROADMAP.md

Awesome Lists containing this project

README

          
# mew

[![Documentation Status](https://readthedocs.org/projects/mew/badge/?version=latest)](https://mew.readthedocs.io/en/latest/)

[![CI](https://github.com/nicholasjng/mew/actions/workflows/ci.yml/badge.svg)](https://github.com/nicholasjng/mew/actions/workflows/ci.yml)

A tool for microbenchmarking Python code, powered by [Google Benchmark](https://github.com/google/benchmark).

```python

import mew

@mew.benchmark

def bench_sorted(state: mew.State) -> None:

    data = list(range(1000, 0, -1))

    for _ in state:

        sorted(data)

```

```console

$ mew run

mew · host=laptop cpus=10 @ 3200MHz scaling=enabled

Benchmark                              │     Iters │       Real │        CPU

────────────────────────────────────────────────────────────────────────────

benchmarks/bench_sort.py::bench_sorted │ 1,000,000 │   32.10 ns │   32.05 ns

```

- **Decorate**: `@mew.benchmark`, `@mew.parametrize`, `@mew.product` register one benchmark or a benchmark family.

- **Run**: `mew run` discovers `bench_*.py` files, streams results to a formatted table, JSON, or Parquet.

- **Profiling**: `mew run --sample` (in-process pyinstrument) and `--profile-memory` (memray allocations) in the same run; `mew profile` for native C frames via xctrace / py-spy / perf.

- **Compare**: `mew compare baseline.json head.json --fail-on-regression 5` for CI jobs.

## Installation

With a `[tool.uv.sources]` entry in your `pyproject.toml`:

```toml

[tool.uv.sources]

mew = { git = "https://github.com/nicholasjng/mew" }

```

```console

$ uv add mew                 # or: pip install mew

$ uv add 'mew[cpu,memory]'   # opt-in profilers

```

Requires Python 3.11+. A pre-built wheel is installed where available;

otherwise the C++ extension is compiled via CMake + nanobind.

A PyPI release is planned for the future.

## Docs

- Quickstart, concepts, user guide: 

- API reference: 

- CLI reference: 

## License

This project is licensed under the Apache-2.0 license.