This workshop will introduce participants to high performance Python on GPUs using tools to provide “simplified” GPU programming, as well as offer a brief look into creating custom kernels by hand.

Learning objectives: Attendees will learn a collection of methods to accelerate Python code if they have access to a GPU, as well as have an idea of how to migrate existing Python code to a GPU.

Knowledge prerequisites: Intermediate Python knowledge (including Numpy). High-Performance Python: CPU recommended but not required.  if you have not used conda before, Intro to Conda is recommended.


Hardware/software prerequisites: (1) Bring a laptop which can connect to the eduroam wireless network. You will also need to be able to Duo authenticate to use campus resources. (2) Have an SSH client installed on your laptop. (3) Register for an account on Adroit. This is the cluster we will use for demonstration purposes. Make sure you can SSH to Adroit before the workshop by following this guide. (4) Optional: Have the Anaconda Distribution installed on your laptop to run things locally if you have a laptop with an NVIDIA GPU.

Workshop format: Lecture and hands-on