Debugging and profiling techniques
summary
Shows concrete steps to find and fix a real Python hotspot: uses a sampling profiler (cProfile/pyinstrument) to spot a function taking ~120ms per call, replaces an O(n^2) loop with an O(n) algorithm and drops it to ~15ms, and verifies with instrumentation timers and timeit. Compares wallclock and CPU counters, explains memory tracking on an 8GB test workload, and demos using perf/Visual Studio and simple flamegraphs to prove the change actually lowers latency and CPU use.
AI-written summary · about
chapters (8)
- 0:00Introducere în debugging
- 10:56Primele pași în depanare
- 20:32Instrumentare și profilare
- 30:36Analiza erorilor comune
- 43:02Tehnici avansate de debugging
- 50:25Profilarea performanței
- 1:06:01Optimizarea sistemului
- 1:11:10Concluzii și resurse
related talks
37mTuning Python applications can dramatically increase performance
32mProfiling at the speed of light
30mWrite faster Python! Common performance anti patterns
30mHow to Tune your Python Analysis Pipeline: A Profiler Guide
24mPractical Profiling
02h 33mMeasure, don't Guess - How to find out if and where to optimize
28mKnocking On A Process’s Door: Introspecting Live Python Processes
24mEngage the Hyper-Python