Many thanks 
to @babreu at NCSA for sharing these links and opportunities to learn about GPU programming and how to use the NCSA Delta HPC system.
Getting Started on NCSA’s Delta Supercomputer
- Event Type: Conference/Workshop
- Date: Oct 27, 2022 1:00 - 2:30 pm
- Sponsor: NCSA
- Speaker: NCSA Staff
- Registration: Complete this form to register.
- Contact: Sandra Kappes
In this webinar, you will learn about NCSA’s Delta supercomputer and how to use it for your scientific research. The following topics will be covered:
- Introduction to Delta
- Node Architecture and Interconnect
- System Services and Features
- Project and Account Management
- Accessing
- Storage
- Running Jobs
- Data Management
- Programming Environment
- Installed Software
- Containers
- Open OnDemand, Science Gateway
- Q & A
Presenters will highlight important features and current issues researchers may encounter when using Delta. This webinar is suitable for current and potential users of Delta.
Resources for learning GPU programming
Architecture
Introduction to GPU Architecture (Cornell)
CUDA
OpenACC
Intro to OpenACC (NERSC + NVIDIA)
Loop Optimization with OpenACC (NERSC + NVIDIA)
General GPU Programming
Accelerated Computing for C/C++ (NVIDIA, not free, certificates available)
GPUs with Python
Using GPUs with Python using Numba and CuPy
NVIDIA Deep Learning Self-Paced Courses
OpenACC and CUDA are good frameworks for folks coming from C/C++ programming, the first being a higher-level approach.