GPU Accelerated Python
summary
Shows concrete wins and code paths: hands‑on notebooks use Numba kernels, CuPy arrays and cuDF/pandas plugin under CUDA on cloud VMs. Example: a 30M-element double runs in Python loops ~8 s, NumPy vectorized ~30 ms, GPU kernel on device at tens of milliseconds — overall orders‑of‑magnitude speedups when memory stays on the GPU. Also demos zero‑change pandas acceleration (groupby reduced from ~9 s to ~2 s) and multi‑GPU scaling via Dask + UCX; kernels, memory movement, and fallbacks are shown and validated.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 1:26Tutorial overview
- 3:37Rapids and GPU Python
- 8:30CPU vs GPU models
- 19:47Numba and CUDA kernels
- 45:42FizzBuzz exercise
- 1:01:58Image processing with CUDA
- 1:11:33Coffee break exercise
- 1:22:46Monitoring GPU performance
- 1:35:23GPU library interoperability
- 1:50:09GPU pandas with cuDF
- 2:01:04Distributed GPU computing
- 2:22:06Q&A
related talks
30mCUDA in Your Python: Effective Parallel Programming on the GPU
35mCUDA in Python- A New Era for GPU Acceleration
29mA Beginners Guide to GPUs for Pythonistas
35mPython at the Speed of Light: A Complete Tour of the CUDA Ecosystem
02h 14mGPU development in Python 101
33mWhen GPUs Make Python Slower: Patterns and Pitfalls
01h 02mExploit your GPU power with PyCUDA (and friends)
01h 02mExploit your GPU power with PyCUDA (and friends)