Completed HPC Projects
Showing 1 - 8 of 8 Results
-
Bull Deutschland GmbH
Bull Germany and the IT Center have joined forces to optimize HPC standard applications, such as OpenFOAM, for hybrid cluster architectures. Furthermore, the collaboration seeks to investigate methods and approaches to make high-performance computers more energy-efficient. In this context, the researchers focus on methods such as the application-dependent control of the CPU frequency, but also optimizations on the level of applications. more
The intensification of agriculture causes a high loss of species-rich grasslands. In Central Europe, these semi-natural grasslands belong to the landscapes with the highest biodiversity. Only a well-directed management will maintain these valuable areas. The GraS-Model (Grassland-Succession-Model) was developed as decision support system for the grassland areas of the Eifel National Park. This dynamic, process based, spatially-explicit (raster-based) simulation model predicts the development of grassland areas over 100 years.
A highly detailed spatial resolution is necessary to ensure a realistic simulation, as even small landscape elements are often crucial for the development. Due to its high complexity and the large-scale landscape modeled, the GraS-Model already exceeds the available computer main memory when applied to the Eifel National Park. The Parallelization of the model allows running simulations on a cluster. This high-performance-computing is essential to ensure realistic simulations with a high predictive power, thus allowing the application and refinement of the GraS-Model as decision support system for the landscape management. more.