Hagen Gädke-Lütjens

Learn More
We present a low-level infrastructure for use by high-level language to hardware compiler back-ends. It consists of the highly parameterizable, technology-independent module library Modlib and the LMEM framework for localizing variables in fast on-chip memories. Modlib not only supports all high-level language operators (including memory accesses), but also(More)
We present an improved method for scheduling speculative data paths which relies on cancel tokens to undo computations in mis-speculated paths. Performancewise, this method is considerably faster than lenient execution, and faster than any other known approach applicable for general (including non-pipelined) computation structures. We present experimental(More)
Control-memory-data flow graphs (CMDFGs) are a unified intermediate representation for compiling high-level languages onto reconfigurable adaptive computing systems. We present both their initial construction as well as transformations for parallel memory accesses. The impact on a number of applications is examined, also considering the effect of caches on(More)
Our concept of a virtual transaction layer (VTL) architecture allows to directly map transaction-level communication channels onto a synthesizable multiprocessor SoC implementation. The VTL is above the physical MPSoC communication architecture, acting as a hardware abstraction layer for both HW and SW components. TLM channels are represented by virtual(More)
Wavelet-based image compression has been suggested previously as a means to evaluate and compare both traditional and reconfigurable computers in terms of performance and resource requirements. We present a reconfigurable implementation of such an application that not only achieves a performance comparable to that of recent CPUs, but does so at a fraction(More)
  • 1