OEHI started the new year 2023 with a bang in the annual Open Edge and HPC Initiative (OEHI) training workshop at GWDG, Göttingen from 17-19th January 2023. Hosted by GWDG, this event also marks a successful start of the collaboration/partnership with DECICE (https://decice.eu), a Horizon Europe project that aims to develop an AI-based, open and portable cloud management framework.

The workshop was very well received with 75 registered participants, of which 18 registered to attend in person. Participants came from a wide range of backgrounds including scaleups, SMEs (EIT Digital community), members of OEHI, DECICE, EuroCC and other academic institutes.

The hybrid workshop was well attended, with almost 50 participants on one day. A hybrid workshop with on-site and online participants presented many challenges and was also a learning exercise for the organisers and training providers.
The lectures and hands-on sessions were all recorded so that those interested could follow along at their own pace. The 3 days are independent of each other and it is easy to follow a particular day. Details of the videos and slides can be found below.
OEHI would like to thank GWDG for hosting and helping us with the logistics, all the presenters and the participants of this event.

Topic | Speakers | Video | Slides |
Day 1. Implement, parallelise and optimise simulation applications on Arm-based clusters | |||
Welcome and Introduction to OEHI | Karthee Sivalingam | Video | Slides |
Arm-based architectures and performance analysis | Dirk Pleiter | Video | Slides |
Porting annd optimisation; Performance tools ; Performance analysis | Dirk Pleiter | Slides | |
Introduction to HAICGU cluster Hands-on : Introduction to cluster | Stepan Nassyr and Marcus Richter | Video | Slides |
Performance optimization | Dirk Pleiter | Video | Slides |
Hands-on: Porting and Optimisation Implement a 2d Poisson solver for heat equation | Dirk Pleiter | Video | Slides/Code |
Day2. Implement and test a training and inference workflows using Arm CPUs and (in a second step) ML and inference accelerators | |||
Building a Versatile, User friendly, All-scenario AI Platform | ShangPengFei | Video | Slides |
Introduction to AI Basics, Da Vinci architecture; Atlas products ; CANN | Kubilay Tuna | Video | Slides |
ACL, pyACL | Kubilay Tuna and team | Video | Slides |
Hands-on – AI inference offline and online | Kubilay Tuna and team | Video | |
Mindspore, MindX; Hands-on – AI training using TensorFlow | Serkan Celik and team | Video | Slides |
Hands-on – AI Training using PyTorch and Mindspore | Alper Balmumcu and team | Video | |
Day 3. Implement and test a micro-service for inference running on Arm CPUs and deployed on an Edge using Kubernetes/KubeEdge instance | |||
KubeEdge Introduction | Lin Guohui | Video | Slides |
KubeEdge Demo / hands-on | Bao Yue | Video | |
Huawei cloud introduction | Atul Athavale | Video | Slides |
SEDNA introduction | Vittorio Cozzolino | Video | Slides |
SEDNA – Demo/ hands- on | Vittorio Cozzolino | Video |