Big ideas shaping scientific Python: the quest for performance and usability
summary
NumPy 2.0 lands with a June 16 rollout that requires recompiling downstream wheels, tracks ~100 packages and takes roughly a year and a half for ecosystem adoption while delivering cleaner Python/C APIs and array API support. The array API standard lets SciPy/Scikit‑learn accept PyTorch/JAX arrays with measured wins like ~2× on CPU and ~30× on GPU in Scikit‑learn examples and up to ~1000× for JAX+JIT on GPU. Early free-threaded Python builds (3.13/3.14 pipeline) show near-linear core scaling, ~6× web throughput and much better data‑loader throughput versus multiprocessing; big-company and foundation funding (Meta, Quans, NVIDIA, NASA, Chan Zuckerberg Initiative) accelerates the work.
AI-written summary · about
chapters (6)
- 0:00Introduction
- 10:41NumPy 2.0 release
- 21:22GPU array support
- 30:17Free-threaded Python
- 43:30Summary
- 44:11Q&A
related talks
45mPython for Speed, Scale, and Science
44mPython's Bright Future in Science
47mPanel: Fostering better collaboration between the scientific Python community and core Python...
43mKEYNOTE: Scaling Out PyData
22mSupercharge Scientific Computing in Python with Numba
30mBreaking the Speed Limit: Fast Statistical Models with Python 3.14...
27mSpeed up Python with hardware accelerators
55mAccelerated Python: The Community and Ecosystem