- De
- En
High Performance Computing (HPC) is becoming ever more present in everyday life, but has affected the same for many decades already. HPC applications have superseded conventional methods in numerous scientific fields, as it enables researchers and engineers to gain more insights in less time. And literally everyone benefits from the achievements in the HPC community.
Applications run on supercomputers are traditionally complex equation systems. The level of detail strived for, requires the parallel execution of multiple complex calculations or a high amount of memory, or both.
One of the most impressing examples is weather prediction: the prediction needs to be done before the weather hits. I.e. tomorrow’s forecast needs to be faster than 24 hours. The more detail can be considered, the more accurate the forecast can be. This also hold true for crisis management and climate research. To accurately calculate, where the flood hits, billions of computations need to be made in literally no time.
Thorough knowledge of the modelled problem, the computational methods and the underlying hardware is necessary to find more efficient solutions.
To efficiently solve these computationally intensive problems, we need:
In addition to hardware, we also need suitable software and a lot of technical and specialist knowledge. The GWDG AG Computing is therefore composed of specialists from very different fields: computer scientists, physicists, mathematicians, chemists, economists and of course HPC experts.
GWDG has been providing researchers with solutions for their computational problems for many years. As such, Göttingen has become a National High Performance Computing and AI Center. We offer:
Operation and maintenance of these systems, user support and user trainings are main tasks of the AG Computing at the GWDG. In addition, we do research to help users to make the computations more efficient. This research comprises the optimization of user applications or code as well as more general HPC-related research to enhance the user experience and make HPC accessible for a broader audience. We also encourage students to write their theses in an HPC-related field or to join the team for a project.
Research by our users spans a wide range. The different research fields and projects of our HLRN users are listed on the HLRN pages and reports thereof are published every four years.
Current projects involve, amongst others, great threats to our society, like the dying of forests. Forests, as providers of oxygen and balancers of humidity and temperature, are fundamental for our climate. The protection of our forests is thus a responsibility of society as a whole. Newest artificial intelligence technology serves us in developing strategies for healthier forests in the future.
One of the main research interests of the DLR is the computation of aerodynamic flows, where the computational capacities of HPC systems enable researchers to solve complex fluid dynamical equations in a time frame, that allows for several runs on different designs, before the best candidate is modeled and analyzed in a wind tunnel. The more computational capacities are available, the higher the level of detail.
Other examples of compute intensive problems, massively benefitting from HPC resources are drug design, where bioloical processes are modeled and simulated before going into clinical studies, making these safer for the probands, or genomics. All of the above are examples are use cases from users on GWDG hosted systems. For our users from the University of Göttingen, the Max-Planck-Society, the DLR or any other public research institution in northern Germany, we operate three different systems: the Scientific Compute Cluster, the NHR-system Emmy and Grete and the DLR-system CARO.