The theme of the first edition was GPU acceleration. This event gathered the people working on applications running on GPUs but also the people who are still looking to port their applications to GPUs. By engaging both groups - those who have ported their applications and those who have not - we hope to identify common obstacles and opportunities for future support. Moreover, the event was a place where we exchange practical experiences, lessons learned, uncover technical challenges, and better understand the needs of the scientific community regarding GPU acceleration. During the meeting there were site presentations, plenty of time to talk to each other, and we wrapped up with a discussion about opportunities for future events and collaboration.
We had presentations from Dutch application oweners from UvA, UU and TU Delft where they provided their insights on accelerated computing. In particular the applications that they discussed were X-PS, AGILE (AMRVAC and BHAC frameworks), IgANet, lattice Boltzmann (LB) and CaNS. We found that scientific applications need more attention as they are the key and play an important role in knowledge discovery and innovations. From one side, we need to invest into design, development and maintenance of scientific applications, and from the other side we need to make effort to adopt them to new technologies. In addition to to investing into research software engineering capabilities, workflow engineering and data and storage management integration, abstractions are of the utmost importance. Abstractions can help scientific applications to readily harness of various accelerated computing platforms.
You can find out the presentation slides and recordings along with photos taken in the event via this link: https://surfdrive.surf.nl/s/4mWXsAPBn7Czscx
0 Praat mee