Arm’s Scalable Vector Extension (SVE) has arrived, but what does SVE mean for typical high performance computing workloads and how do you use it? This hands-on, non-NDA event will present SVE and show how common HPC codes can be recompiled to take advantage of large vector lengths without code modifications.
Schedule
9:00am – Introduction and Fundamentals
Short presentations from community members, SVE tool developers, application teams, and hardware developers to introduce core concepts and lay the roadmap for future SVE applications.
Minutes | Materials | Presenter |
---|---|---|
5 | Welcome and Overview of Arm in HPC | John Linford |
25 | Introduction to Programming for Arm SVE | John Linford |
25 | Timing-agnostic SVE optimization with ArmIE | Roxana Rusitoru |
25 | RIKEN Fugaku Processor Simulator | Yuetsu Kodama |
30 | A Closer Look at Arm SVE | John Linford |
5 | Training Cluster First-touch | Everyone! |
11:00am – Hands-on with SVE
The entirely hands-on afternoon session will introduce the Arm’s tools for SVE developers and show how to generate SVE-enabled binaries targeting popular vector lengths. The session will include contributions from people who are running and porting applications for SVE.
- Training Cluster:
ssh student@cluster.arm-hpc.org
- Password:
Tr@ining!
- Password:
- Materials: https://gitlab.com/arm-hpc/training/arm-sve-tools/
Registration
- Register now to guarantee your seat.
- https://pages.arm.com/sve-hackathon-sc19
Contact
- John Linford john.linford@arm.com
Acknowledgements
Many thanks to all our partners and teams within Arm who contributed content including, but not limited to, Dani Ruiz-Munoz, Miguel Tairum-Cruz, Olly Perks, Francesco Petrogalli, Alex Rico, Roxana Rusitoru, and Jelena Milanovic.