PRACE and VI-SEEM Spring School

From PUBLIC-WIKI
Jump to navigation Jump to search

The PRACE 2017 Spring School will be a joint training event with the VI-SEEM project and will be organised by CaSToRC at The Cyprus Institute.

The school will be a 3-day event starting Tuesday 25 April and ending Thursday 27 April 2017.

The Seasonal School will provide insights into topics of high interest to people of the PRACE, VI-SEEM and European communities and will have two parallel sessions - a Developer Track and aSystem Administration Track.

The Developer Track includes (among others) the following topics:

  • Introduction to Parallel and Scalable Machine Learning: Basics and Parallelization Benefits
  • Introduction to Deep Learning with Convolutional Neural Networks & Applications
  • Accelerator Programming: GPU and Xeon Phi
  • Data Visualisation: Information visualization, visual analytics, VisIt

Speakers from the EUDAT, OpenSESAME and VI-SEEM projects will also present these project and relevant case studies.

For more details on the Developer Track please visit the link below: https://events.prace-ri.eu/event/601/program#developerTrack

The System Administration Track includes (among others) the following topics:

  • System Administration Tools: Easybuild and other tools
  • Security Aspects of HPC Centers
  • PRACE Network and Monitoring Issues* System Monitoring
  • Cloud Services - OpenStack
    • OpenStack and HPC, the crossroads of scientific research
    • How to build your first OpenStack Cloud
    • Considerations for maximising the performance of OpenStack
    • RDMA enabled fabrics in an OpenStack environment
    • IOPS in the cloud – impossible! Or is it?
    • Case studies from HPC centres using OpenStack for HPC workloads.

For more details on the System Administration Track please visit the link below: https://events.prace-ri.eu/event/601/program#systemAdministrationTrack

A Keynote Address will also be given by the Winner of the 1st PRACE Ada Lovelace Award for HPC – Dr Zoe Cournia.

Useful Links: