Research

High performance computing solves problems that are too complex for classical methods. First, the descriptive equations are modified (discretised) so that a computer algorithm can find a solution by iteration. The CCSS members work in algorithm development, software implementation, and use of the programs.

These programs are used to investigate phenomena that often occur on different scales (e.g. from a crystallite to a rotor blade), and the impact at small scales on the full simulation must then be modelled using closure models or scale transition conditions. Such modelling theories are similarly developed by members of the CCSS and tested in complex simulations.

CCSS members from Biology, Chemistry, Computer Science, Engineering, Mathematics and Physics often deal with problems that require huge amounts of computing capacity to solve – executing the corresponding programs sequentially would often take decades(!). For that reason, different computations must be executed simultaneously, which is referred to as parallel computing. Parallel computing is only possible with efficient, often complex parallelisation strategies and suitable parallel computers. Our supercomputer, magnitUDE, is an NEC system with close on 15,000 cores.

This supercomputer is instrumental in attracting external funding and outstanding researchers and conducting research projects with an international profile. Equally, it also allows the researchers at the UDE to develop their algorithms so that they can use even more powerful supercomputers (e.g. in Jülich, Munich or Stuttgart). This means that various research groups now have the capability to undertake top-level research that requires many times the computing power available at the UDE. It is precisely in enabling the UDE researchers to do this that the joint CCSS and ZIM support team plays such an important role (https://www.uni-due. de/ccss/sc_support.php).

Nevertheless, groups that are able to work on individual projects in Jülich, Munich or Stuttgart also rely on magnitUDE for much of their work, as it can be used without the lengthy project application process and scientific review that would be required by those facilities.