pyquickbench 0.2.1

A pure Pyton tool to perform time and accuracy benchmarks

Pyquickbench

Pyquickbench is an open source, easy to use benchmarking tool written in pure Python. Checkout the example gallery to get an idea of what pyquickbench is capable.

• Tutorial: https://gabrielfougeron.github.io/pyquickbench/tutorial.html
• Documentation: https://gabrielfougeron.github.io/pyquickbench/
• Source code: https://github.com/gabrielfougeron/pyquickbench
• Bug reports: https://github.com/gabrielfougeron/pyquickbench/issues
• Changelog: https://github.com/gabrielfougeron/pyquickbench/releases/

Main features

• Timings / repeatability / output benchmarks
• Error handling
• Benchmark results caching
• Multithreaded / multiprocessed benchmarks
• Benchmark timeout
• Multidimensional benchmarks
• Transformed data plotting (relative values, convergence order, ...)
• Intelligent plots
• Sensible defaults

Installation

Pyquickbench is available on the Python Package Index. To install using pip, simply type:

pip install pyquickbench

Pyquickbench is available on conda-forge. To install using conda, simply type:

conda install pyquickbench -c conda-forge

To install the current development version of pyquickbench from the github repository, you can type:

pip install git+ssh://[e-mail]/gabrielfougeron/pyquickbench.git 

Tests

To run tests locally on your machine, first checkout this reposity and install dependencies using pip:

git clone [e-mail]:gabrielfougeron/pyquickbench.git
cd pyquickbench
pip install .[tests]

Then, run tests using pytest:

pytest

License

This software is published under the BSD 2-Clause License.

Other open source alternatives

You might like perfplot, one of the inspirations for this work.