NCI and NVIDIA will be holding a pair of bootcamps in November focused on utilising Gadi's GPU capabilities. GPU Bootcamps are an exciting and unique way for scientists and researchers to learn the skills needed to start quickly accelerating codes on GPUs.
N-Ways (GPU Bootcamp)
1-3 November 2021, 11.30am to 2.30pm AEDT
Click here to register (via GPU Hackathon)
Summary
Participants will learn about multiple GPU programming models and can choose the one that best fits their needs to run their scientific codes on GPUs.
This Bootcamp will cover an introduction to GPU programming using OpenACC, OpenMP, stdpar, Fortran Do-Concurrent and CUDA, and provides hands-on opportunities to learn how to analyse GPU-enabled applications using NVIDIA® Nsight™ Systems.
1st Nov 2021 →
- Welcome (Moderator): 11:30 AM - 11:45 AM
- Connecting to a Cluster: 11:45 AM - 12:00 PM
- Introduction to GPU Computing: 12:00 PM - 12:45 PM (Lecture)
- Introduction to Profiling Tools: 12:45 PM - 1:15 PM (Lecture)
- Accelerating Standard C++ and Fortran with GPUs: 1:15 PM - 2:15 PM (Lecture + Lab)
- Q&A: 2:15 PM - 2:30 PM
2nd Nov 2021 →
- Directive Based Programming with OpenACC on GPU : 11:30 AM - 12:30 PM (Lecture and Lab)
- CUDA C/Fortran Programming 12:30 PM - 2:15 PM (Lecture + Lab)
- Q&A 1:30 PM - 2:15 PM
3rd Nov 2021 →
- Mini Application Challenge : 11:30 AM - 1:30 PM ( Lab)
- Wrap-up and announcing winners 1:30 PM - 2:15 PM
Prerequisites
- Basic experience with C/C++ or Fortran. No GPU programming knowledge is required.
- Presented by NCI Australia, the bootcamp is open to all current and prospective NCI users.
- This event has limited capacity, so please make sure that prerequisites are met before applying. You will be receiving an acceptance email with details on how to participate by 27th Oct 2021.
AI For Science (GPU Bootcamp)
29-30 November 2021, 11.30am to 2.30pm AEDT
Click here to register (via GPU Hackathon)
Summary
Participants will learn the basics of GPU computing and AI concepts, particularly as it relates to applying Deep Learning (DL) for scientific problems.
DL terminology and concepts will be instructed, and participants will start by creating their first “Hello World’ DL network. Concepts of AI for Science will be discussed to draw a distinction between how scientific AI deployments differ from traditional AI. We’ll then explore Convolutional Neural Networks and apply them to a Tropical Cyclone Detection case where you will learn the impact of modifying hyperparameters, network layers, and data augmentation to improve your model. Finally, we’ll explore how more complicated networks (U-Nets) work and you’ll gain experience applying them to a steady flow example problem.
29th Nov 2021
- Welcome (Moderator): 11:30 AM - 11:45 AM
- Connecting to a Cluster: 11:45 AM - 12:00 PM
- Introduction to GPU Computing: 12:00 PM - 12:30 PM (Lecture)
- Introduction to AI: 12:30 PM - 1:00 PM (Lecture)
- Break 1:00 PM - 1:15 PM
- CNN Primer and Keras 101: 1:15 PM - 2:15 PM (Lecture + Lab)
- Q&A: 2:15 PM - 2:30 PM
30th Nov 2021
- AI for Tropical Cyclone Detection : 11:30 AM - 1:00 PM (Lecture and Lab)
- Mini Application Challenge : 1:00 PM - 2:00 PM ( Lab)
- Introduction to Physics Informed Neural Network : 2:00 PM - 2:30 PM ( Lecture)
Prerequisites
- Basic experience with Python. No GPU programming or AI knowledge is required.
- Presented by NCI Australia, the bootcamp is open to all current and prospective NCI users.
- This event has limited capacity, so please make sure that prerequisites are met before applying. You will be receiving an acceptance email with details on how to participate by 24th Nov 2021.