9 Levels of Profiling Python Apps in 2026: From cProfile to Tachyon | by Yang Zhou | TechToFreedom | Jun, 2026 | MediumSitemapOpen in appSign up<br>Sign in
Medium Logo
Get app<br>Write
Search
Sign up<br>Sign in
Mastodon
TechToFreedom
Technology gives us more and more freedom. Start learning today.
Member-only story
Python<br>9 Levels of Profiling Python Apps in 2026: From cProfile to Tachyon
A practical guide to finding real Python bottlenecks with timing, tracing, flame graphs, and Python 3.15’s Tachyon.
Yang Zhou
9 min read·<br>Just now
Listen
Share
Press enter or click to view image in full size
Image from WallhavenAt some point in their career, almost every software engineer hears the same request: “Make it faster!”<br>The problem is that most programs are not slow for the reasons you initially suspect.<br>Before optimizing a single line of code, you need to know where the program is actually spending its time.<br>That’s exactly what profiling is for.<br>Fortunately, Python developers have a better profiling toolbox than ever before. While cProfile remains a reliable option, Python 3.15 introduces a new profiling package and a statistical sampling profiler called Tachyon. (According to PEP 790, the final release is expected on October 1, 2026.)<br>In this article, we’ll start with simple timing techniques, move on to cProfile, and then explore Python 3.15’s new techniques.<br>1. Measure Runtime Before You Profile<br>Timing is not really profiling, but it is still useful.
Published in TechToFreedom<br>1.6K followers<br>·Last published just now
Technology gives us more and more freedom. Start learning today.
Written by Yang Zhou<br>14.6K followers<br>·209 following
Full-Stack Engineer
Help
Status
About
Careers
Press
Blog
Privacy
Rules
Terms
Text to speech