Research

The current research focus of the Computer Engineering Group is Reconfigurable Computing (Custom Computing). Classically, computers consist of software and hardware. The hardware is fixed when the computer is manufactured and the software is loaded and removed when the computer is in field usage. Reconfigurable computing structures change this classical view of computer systems. Computers built from reconfigurable structures do not rely on a fixed hardware, but adapt their architecture to the application under execution. Making hardware soft leads to a paradigm shift in computing and is believed to radically change the way we construct and use computing systems.

Reconfigurable hardware is a relatively young and exciting technology opening up new opportunities for both embedded and high-performance systems. Being massively parallel programmable architectures, reconfigurable hardware devices share many research challenges with emerging homogeneous and heterogeneous many-core chips. Interestingly, the first commercial reconfigurable hardware device introduced in the 80s featured 64 logic blocks which is about the number of processors found in current many-cores. While programming reconfigurable devices requires computing in time and space, spatial aspects such as placement and routing have not yet played a role for chips with multiple processor cores. We are interested in all aspects of "programmable concurrency", including the design and evaluation of innovative architectures, the development of novel design methods and tools, and mapping demanding applications to reconfigurable and parallel systems. Our current research activities are grouped into the following three areas:

Operating Systems for Reconfigurable Computers

Keywords: operating system kernels, hardware multithreading, dynamic and partial reconfiguration, real-time and online scheduling, verification and security

  • Programming of Hybrid CPU/FPGA Systems (funded by the German Academic Exchange Service)
  • XFORCES: Executables for Reconfigurable Embedded Systems (completed)

Self-Adaptive Components and Systems

Keywords: evolvable hardware, intrinsic and online evolution, self-reconfiguration

  • Adaptive System-on-Chip (completed)

Custom Computing and Parallel Architectures

Keywords: FPGA accelerators, reconfigurable processors, many-cores

  • MM-RPU: A Multimode Reconfigurable Processor Unit (funded by the Intel Microprocessor Technology Lab)
  • Just in Time Hardware Compilation
  • IMORC: An Infrastructure for Performance Monitoring and Optimization of Reconfigurable Computers
  • Innovative Architectures for Medical Image Processing

      • Advanced Computing Frameworks using Heterogeneous and Reconfigurable Multicore Architectures (completed)
      • ZIPPY: A Novel Reconfigurable Embedded Processor (completed)
      • Acceleration of Hard Combinatorial Problems (completed)


      Our research is conducted in close cooperation with the Paderborn Center for Parallel Computing, supported by the DFG (German Research Foundation) within the Priority Programs 1148, "Rekonfigurierbare Rechensysteme" and 1183, "Organic Computing", by the DAAD (German Academic Exchange Service), by grants awarded by the Intel Advanced Computing Center, the Intel Microprocessor Technology Lab, Microsoft Research, the German Ministry of Economics and Technology, and the University Programs of Altera-AMD-Sun-XtremeData and Xilinx.

      Imprint | Webmaster | Recent changes: 07.02.2012