Examples and teaching material for the July 2018 "Introduction to GPU programming" summer school at the University of Warwick.
To run the notebooks you will need a CUDA capable GPU card. The examples use double precision arithmetic and assume a GPU with a compute capability of 2.0 or higher. The school was conducted using Kepler K20c and K80 hardware.
- CUDA Toolkit (tested with version 9.1-11.1)
- Python 3
- Numpy
- Numba
- CuPy
- Pillow
- Matplotlib
- Jupyter
This software is available on all Linux hosts managed by the Scientific Computing Research Technology Plaform at the University of Warwick via environment modules,
On the (old) CentOS 7 desktop use:
$ module load GCC/10.2.0 CUDA/11.1.1 OpenMPI/4.0.5 Python OpenBLAS numba CuPy Pillow IPython matplotlib
One new (new) Rocky 9 desktop use:
$ module load GCC/11.3.0 OpenMPI/4.1.4 CUDA/11.7.0 Python numba CuPy Pillow IPython matplotlib
Porgs are cool.