July 18, 2019

High Performance Computing (HPC) for neuroscience: Hands-on introduction to supercomputing usage, tools and applications

Neuroscience research has become increasingly interdisciplinary in recent years. New imaging technologies deliver ultra-high resolution images, and new simulator technology enables scientists to simulate larger and more detailed neural networks. Such data can no longer be analysed and such simulations can no longer be run solely on a user’s computer in the office: clusters, supercomputers and good data management strategies have become indispensable.

The Education Programme of the Human Brain Project (HBP) offers innovative learning packages, online and with on-site events, for early career researchers working in and across the fields of neuroscience, information & communications technology and medicine. The SimLab Neuroscience of the Jülich Supercomputing Centre (JSC) organised an HBP Education Workshop that set the grounds for students to get started with HPC-based research and laid the foundation for them to advance the state of the art in their fields. This event took place from July 9-11 at Forschungszentrum Jülich. It was well-attended and well-received by students from all across Europe.

The workshop taught the basics of supercomputing needed in order to start using HPC systems for (neuroscience) research. This included, on the one hand, introductory lectures with hands-on sessions about scientific computing in Python and an introduction to the usage of HPC systems and (big) data management. In addition, the students received hands-on training for tools and applications that can be used both on a supercomputer and on the user’s local computer; these included the simulators NESTand Arbor, as well as visualisation tools that can handle large imaging or simulation data, as generated on a supercomputer. Finally, the students learned how to get access to the Fenix Infrastructure, which offers HPC and data resources for their research projects.

The tools and applications that were presented in the workshop are currently being developed in the HBP’s High Performance Analytics and Computing (HPAC) Platform. The introductory lectures enabled the students to also make efficient use of the other HBP Platforms in the future, in particular the Neuroinformatics, Brain Simulation and Neurorobotics Platforms that use the HPAC Platform as a basis.

Text: Meredith Peyser, Images: HBP Education Team

 

 

ENDS