How to speed up your CI pipelines

This repo implements six different strategies for setting up a Python environment in your CI pipeline in order to compare how fast these are for installing your dependencies.

The different strategies are:

• Use pip as package manager
• Use pip but cache dependencies between runs
• Use conda as package manager
• Use conda but cache dependencies between runs
• use mamba as package manager
• Use mamba but cache dependencies between runs

You can find the results of the CI pipelines here.

Resources

• Blog post by epassaro for caching mamba
• official documentation for conda strategies
• Github actions documentation for pip caching

Contact

Alexander Fottner a.fottner@googlemail.com