DelftBlue Supercomputer Course

To give you a head start on this new and exciting facility, the Delft High-Performance Computing center (DHPC) organizes two-day workshops. The two days have separate themes and (learning) objectives, and you may register for either of them, or both.

On both days, there will be plenty of time to get hands-on experience on DelftBlue, and experienced HPC users and developers will be available for questions and discussions.

This day targets people who have never worked on an HPC cluster and want to get an understanding of how such a machine works and can be used in practice. The contents of the course do not go far beyond what is presented in the online documentation but should enable you to;

  • Understand the basic concepts of DelftBlue (shared vs. distributed memory, node types, job scheduling, managing your data, and accounting),
  • Login, create a job script, start a batch or interactive job, and transfer data between DelftBlue and your local machine or network storage.
  • Effectively use the documentation to familiarize yourself with the machine,
  • Have a rough overview of software and programming models suitable for use on DelftBlue.

Programme

09:30 - 09:45 Arrival with coffee and tea
09:45 - 10:00 Welcome to DHPC. (C. Vuik)
10:00 - 10:30 What is a supercomputer?
10:30 - 11:00 What does DHPC offer, and how?
11:00 - 12:30 Practical use of DelftBlue

12:30 - 13:30 Lunch

13:30 - 14:15 Programming Techniques for supercomputers
14:15 - 15:00 Hands-on session (bring your own work, or try one of our examples)
15:00 - 16:00 Wrap-up and open discussion

This day targets both experienced and new HPC users. Its aim is to develop a model of DelftBlue in your mind which you can then use to determine how to optimally map your applications onto the real machine. Furthermore, we will recap the essential usage of DelftBlue from day 1, and discuss some advanced topics like using the data transfer and visualization nodes, and assessing program performance.

At the end of the day, you should:

  • have a good overview of the complete system and how to use it for your applications,
  • understand the system description of DelftBlue in more detail (e.g. network topology, basic performance characteristics),
  • know some useful tools and workflows for debugging, profiling and performance analysis, pre- and post-processing.

Programme

09:30 - 09:45 Arrival with coffee and tea
09:45 - 10:15 Quick recap of Day 1: DHPC, policies, login to DelftBlue
10:15 - 10:30 Overview of DelftBlue
10:30 - 12:30 Workflow demos

12:30 - 13:30 Lunch

13:30 - 14:30 Performance Analysis

14:30 - 14:45 Coffee Break

14:45 - 15:30 Hands-on session (bring your own work, or try one of our examples)
15:30 - 16:30 Open discussion

More information

Dr. J. (Jonas) Thies

Planned Courses

/* */