Explore the intricacies of measuring CPython's performance in this 50-minute PyCon US talk by Michael Droettboom. Delve into the history of performance measurement in Python and understand the unique challenges of evaluating an interpreter's efficiency. Gain insights into CPython's pyperformance benchmark suite, learning what makes a benchmark valuable and why defining a "typical Python workload" is complex. Discover the difficulties in accurately measuring time and the innovative solutions involving software, hardware, and statistical methods. Get an overview of the continuous benchmarking system based on GitHub Actions. Apply these performance measurement techniques to your own software projects, enhancing your ability to evaluate and optimize code efficiency across various programming contexts.
Overview
Syllabus
Talks - Michael Droettboom: Measuring the performance of CPython
Taught by
PyCon US