Developing Services for Scientific Communities: Experiences from the National High Performance Computing Program in Germany
DescriptionThe academic HPC landscape in Germany has become much more service-oriented over the last decade. Based on the insight that scientists require more than purely technical support to successfully and efficiently work on HPC clusters, state and federal programs have been instituted that provide consulting, training activities, and more. In particular, the launch of the initially 10-year National HPC project (Nationales Hochleistungsrechnen, NHR) in 2021 has provided more opportunities to address the needs of researchers in a fashion targeted to their respective scientific communities, thus making these services more focussed and effective. However, identifying and approaching different communities regarding HPC support, and finally providing the right services for them are not self-evident tasks at all. We find these groups to be amorphous, heterogeneous, and often quite inexplicit in their requirements. Using several communities in the computational engineering sciences as example, we reflect on measures that have and can be taken to facilitate the development of services for scientific communities in HPC. Examples for such measures are tailored events, communication strategies and information offerings. We stress the importance of the continuity and sustainability of these efforts, and the availability and training of a dedicated and service-oriented scientific staff.
TimeMonday, June 2615:00 - 15:30 CEST
Computer Science, Machine Learning, and Applied Mathematics