Topic ComputingCopyright: FZJ
The processing of data on the computer can be symbolical or numerical. Both concepts are based on fundamentally different methodical approaches, even though symbolical (pre-) computations are used in some cases to guarantee an accelerated numerical computation. The term computing is generally related to numerical computations or to pre- and postprocessing approaches such as mesh generation or visualization methods. Besides advanced solution methods, computing, i.e., the efficient numerical implementation using software and hardware, is for future research efforts of central significance to deal with the other topics such as data and modeling. High performance computing (HPC), which enables to address large-scale problems in numerous scientific and socially relevant issues by connecting intelligent computer architecture and numerics, is only a very demanding component in the field of scientific computing.
Topic Computing in the Profile Area
Together with the Research Center Jülich, the RWTH has a unique computing infrastructure, which can be regarded as an excellent hardware basis for the profile area CompSE. Via JARA-HPC, the profile area already has privileged access to the Tier-0 computer architecture JUWELS and JURECA and can therefore prioritize the planned research projects in the four thematic areas. The further expansion with regard to an Exascale computer architecture is currently the subject of several projects such as DEEP. The renewal of the Tier-1 computer architecture at RWTH Aachen University is also in the application phase. The application, which was designed by the IT center, provides for a computing power of approx. 900 TFlops from 2017. The unique Jülich-Aachen-Research-Alliance thus ensures that the latest Tier 1 computer technology is available for research work at the Tier-1 level every 3 years, since the upgrading of the computer capacity in Jülich and at the RWTH Aachen University in coordination with the state and the federal government takes place alternately after 6 years. Based on this excellent computing infrastructure, the following research topics will be dealt with as part of the future scientific challenges:
- Exascale Computing
- Highly efficient parallelization of program modules
- program verification
- Scientific Visualization
- grid generation
- computer architectures
In the context of these integratively interdisciplinary research topics, the existing Tier-1 computer architecture at the RWTH Aachen University is an ideal training field to introduce the multidisciplinary networked young scientists in the profile area of CompSE to future Tier-0 HPC structures.