Supercharge Scientific Computing in Python with Numba
summary
Numba JIT compiles a subset of Python and NumPy into machine code and shows concrete wins: a 500×500 Mandelbrot goes from ~7 s to <1 s to <100 ms after applying the njit decorator and the parallel option with prange, a diffusion PDE loop runs ~12 s in pure Python, vectorizing yields ~200× speedup but hurts readability, while the stencil decorator gives >700×, and an open source blurhash port reports 33–63× encode and 34–70× decode speedups with ~20× on Raspberry Pi.
AI-written summary · about
chapters (11)
- 0:00Numba introduction
- 3:14What Numba is
- 5:20Mandelbrot set example
- 8:49Numba options for performance
- 10:50Numba with NumPy arrays
- 12:57Vectorizing functions with ufuncs
- 14:20Diffusion equation efficiently
- 17:32When to use Numba
- 18:13Real-world Numba project
- 20:42Numba on edge devices
- 21:13Q&A
related talks
37mAccelerating Scientific Computing in Python with Numba: A Real-World Example
37mUsing Numba Effectively Today
01h 33mPower up your work with compiling and profiling
30mNumba, a JIT compiler for fast numerical code
33mLet's Run Python on a Supercomputer!
42mMaking your code faster: Cython and parallel processing in the Jupyter Notebook
30mHighLoad Python SIMD, GPU, and Horizontal Scaling Looking into Silicon
44mPython's Bright Future in Science