Performance Optimization

Profile first, then use better algorithms, built-ins, vectorisation or compiled code.

Intermediate

Measure with timeit, cProfile, Scalene or py-spy.
Fix the algorithm — $O(n^2) \to O(n \log n)$ beats any micro-tweak.
Use built-ins and the standard library (implemented in C).
Vectorise with NumPy instead of Python loops.
Compile hot spots with Numba, Cython or mypyc; cache results with functools.lru_cache.

Spotted a mistake or want to add detail? Suggest an improvement — Content Contributors' edits are reviewed by editors before they go live.

Discussion (0)