GPU Accelerated Python w: CUDA
summary
Claims practical demos: a transparent pandas plugin that dispatches to cuDF and a scikit‑learn plugin, hands‑on use of Numba CUDA kernels and CuPy within the Rapids stack. Measured orders of magnitude: a simple Python loop on 30M elements goes from ~9 s on CPU to ~200 ms with naive GPU copies and down to ~40 ms/µs with pinned device data; joins and groupbys report up to 100× speedup. Builds include CSV readers and UDFs that run on the GPU and they work, but memory movement and bounds checks matter.
AI-written summary · about
chapters (4)
related talks
01h 29mGPU development with Python 101
29mA Beginners Guide to GPUs for Pythonistas
01h 12mGPU-Accelerated Data Analytics in Python
33mWhen GPUs Make Python Slower: Patterns and Pitfalls
31mcuDF: RAPIDS GPU-Accelerated Dataframe Library
35mPython at the Speed of Light: A Complete Tour of the CUDA Ecosystem
41mLuxuries, necessities, and the challenges that remain
01h 36mRAPIDS: GPU-Accelerated Data Science for PyData Users