Our research focus is on three innovative paradigms for building computing systems: reconfigurable computing, self-* computing, and approximate computing. The research mission is to study and investigate these paradigms, to develop novel architectures as well as design methods and tools for them, and to demonstrate their feasibility and usefulness by creating prototypical applications in the domains of embedded and high-performance computing.
Reconfigurable computing systems are built from reprogrammable hardware structures and can adapt their hardware architecture to the application under execution. This hardware-on-demand paradigm dramatically increases system flexibility and is beneficial for acceleration, for reducing energy consumption, and sometimes even for reducing cost.
Approximate computing trades off computational accuracy against reductions in energy consumption, runtime, or hardware cost. This trade-off becomes possible since for many applications approximate, i.e., inaccurate, results are indeed good enough and often hard to distinguish from accurate results.
We also value research into novel technologies driven by demanding applications. In the past, we have worked on the classification of EMG signals for prosthesis control, on the use of advanced Monte-Carlo tree search methods for computer Go, and on the optimization of energy distribution networks.
Over the last years our research has been supported by grants from the Deutsche Forschungsgemeinschaft (DFG), for example through the collaborative research center On-The-Fly Computing (CRC 901), the European Commission (EC), the German Federal Ministry for Education and Research (BMBF), the German Federal Ministry for Economic Affairs and Energy (BMWI), Microsoft, and Intel.
Prof. Dr. Marco Platzner