nwstephens/meetup-uofa

These notebooks introduce RAPIDS and CUDA-X for drop-in GPU-accelerated libraries; CuPy and Numba for GPU accelerated Python code; and torch and cuda.ml for running GPU's with R.

GPU Accelerated Libraries and Python with NVIDIA GPUs

In this session we will cover some of the most popular and effective GPU accelerated libraries that give high performance without the requirement of writing your own custom GPU code. We will cover CUDA-X which has libraries for math, image/video processing, deep learning, and GPU tailored partner libraries. On top of CUDA-X we will cover RAPIDS which will target data science and data analytics workloads. We will conclude the session with interactive coverage of NVIDIAs profiling tools. We will conclude with a brief coverage of Python specific tools we have like CuPy and Numba for customizable GPU accelerated code. By the end of the workshop, you'll have the skills to utilize existing GPU accelerated libraries and write your own Python codes with NVIDIA GPUs!

Topics

1. RAPIDS and CUDA-X for drop-in GPU-accelerated libraries
2. CuPy and Numba for GPU accelerated Python code
3. torch, reticulate, and cuda.ml for running GPU's with R

Tools, libraries, and frameworks used

• RAPIDS
• CUDA-X
• CuPy
• Numba
• Torch for R