LFBS Home Lehrstuhl für Betriebssysteme RWTH
Home > Research  

Research & Projects

The Chair for Operating Systems and its follow-up research groups work on research and development projects concentrating on the areas of parallel and real-time computing. Main goals were to enhance and to develop operating systems and system software for use on parallel and real-time capable hardware.

Parallel Computing
For decades, parallel computing has been used in order to gain greater efficiency and availability especially in the area of high-performance computing. Nowadays, multicore systems and computing clusters based on commodity hard- and software pose new challenges for research in parallel computing. For that reason, we work on system software for efficient programming and use of heterogeneous cluster networks, as well as on multicore-related improvements for modern operating systems. In particular, broadening and challenging technologies like NUMA architectures, heterogeneous multicore CPUs and Cluster-on-a-Chip approaches, as for example the Intel Single-Chip Cloud Computer (SCC), are in the focus of our research interests.

Real-time Computing
In the future, time critical tasks will most probably be solved by a group of machines. These tasks, which are constrained by strict deadlines, are mostly run on embedded systems that are severely limited in computing power, a fact that further complicates the solution. Since standard PC-based systems and corresponding components are becoming more and more important for embedded applications, one of our major research topic is to develop distributed and/or parallel real-time environments with off-the-shelf hard- and software.

Current Research Projects

MetalSVM: A Bare-Metal Hypervisor for Non-Coherent Memory-Coupled Cores
Development of a shared virtual memory management system that will be implemented in terms of a bare-metal hypervisor on the Intel Single-Chip Cloud Computer (SCC).

SCC-MPICH: A Power-Aware, Low-Latency MPI Implementation on the SCC
Development of a high-performance Message-Passing Interface (MPI) library that is able to minimize the CPU-related energy consumption of parallel applications on the Intel Single-Chip Cloud Computer (SCC).

Real-Time Computing on Multicore Computers
Mixed hard real-time and high-performance computing on multicore processors.

GPGPU Programming in heterogeneous systems
Task scheduling on GPUs in heterogeneous HPC systems.

Please do not hesitate to Contact us if you have further questions or inquiries of any kind.