Description
Summary
This GPU Bootcamp is an exciting and unique way for scientists and researchers to learn the skills needed to start quickly accelerating codes on GPUs. Held as a virtual event across three days (three three-hour sessions), participants will learn how to write intermediate to advanced CUDA optimisation techniques for code acceleration.
Presented by NCI, Pawsey Australia, NeSI New Zealand, the Bootcamp is open to all current and prospective NCI, Pawsey, and New Zealand HPC users.
Agenda
Sep 7, 2022 – 11.30am–2.30pm AEST
- Welcome (Moderator): 11.30am-11.45am
- Connecting to a Cluster: 11.45am-12.00pm
- CUDA 101: 12.00pm-1.00pm (Lecture)
- GPU Memory Hierarchy (1.00pm-2.30pm)
Sep 8, 2022 – 11.30am–2.30pm AEST
- GPU Memory Hierarchy - Shared Memory: 11.30am-12.30pm (Lecture + Lab)
- CUDA Thread Programming: GPU Occupancy, Atomic Operation: 12.30pm-2.30pm
Sep 9, 2022 – 11.30am–2.30pm AEST
- CUDA Thread Programming: 11.30am-12.30pm (Lecture + Lab)
- Warp Divergence
- Debugging and profiling: 12.30pm-2.30pm
Prerequisites
- Basic experience with C/C++ or Fortran.
- Knowledge of parallel computing fundamentals or basics of CUDA.
- This event has limited capacity, so please make sure that prerequisites are met before applying.
Event Format
The 3-day Bootcamp will be hosted online in the Aust Eastern Standard Time (AEST) zone. All communication will be done through Zoom, Slack and email.
