Faster Python Code

Learn tips to help optimize your Python code. Discover how to pick the right data structures, use caching, integrate performance in your process, and more.

Introduction

• Welcome
• What you should know
• Use Codespaces with this course

1. Tools of the Trade

• Always profile first
• General tips
• Measuring time
• CPU profiling
• line_profiler
• Tracing memory allocations
• memory_profiler

2. Picking the Right Data Structure

• Big-O notation
• bisect
• deque
• heapq
• Beyond the standard library

3. Tricks of the Trade

• Local caching of names
• Remove function calls
• Using __slots__
• Built-ins
• Allocate

4. Caching

• Overview
• Pre-calculating
• lru_cache
• Joblib

5. Cheating

• When approximation is good enough
• Cheating example

6. Parallel Computing

• Amdahl's Law
• Threads
• Processes
• asyncio

7. Beyond Python

• NumPy
• Numba
• Cython
• PyPy
• C extensions

8. Adding Optimization to Your Process

• Why do we need a process?
• Design and code reviews
• Benchmarks
• Monitoring and alerting

Conclusion

• Next steps