In this project we propose to research and develop (runtime) techniques to dynamically adapt both software and hardware components according to the hardware resources available during runtime. We will research schemes for dynamic reconfiguration of computing engines coupled to a microprocessor. Specifically, we will use a regular expression and a sorting computing hardware engine to analyse and evaluate the potential of software/hardware dynamic adaptation in order to use more hardware resources whenever higher performance is required and hardware space is available. We will research and develop new dynamic, partial, reconfiguration schemes to specialise domain-specific architectures in order to improve resource usage and maximum operating clock frequencies.
The major scientific goals of this project are:
- To research methods for both software and hardware runtime adaptations. This will be the first time to be addressed, as far as we know, and will consists on novel scientific achievements;
- To continue researching dynamic reconfiguration techniques to complement automatic generation of strategies based on a set of programmable, hardware templates;
- To understand better the impact of dynamic reconfiguration when dealing with hardware/software systems. The use of a sorting machine and a machine to accelerate evaluation of large sets of regular expressions will be used to test and apply the main techniques.
- Institut für Technik der Informationsverarbeitung (ITIV), Karlsruhe Institute of Technology (KIT), Karlsruhe, Germany
- Faculty of Engineering, University of Porto, Porto, Portugal
Bilateral cooperation: FCT/DAAD 2010/2011, under Ref. daad124412622223295
João Manuel Paiva Cardoso
Department of Informatics Engineering (DEI), Faculty of Engineering,
University of Porto,
Phone: +351 916629046
Faculdade de Engenharia da Universidade do Porto (FEUP)
Rua Dr. Roberto Frias
4200-465 Porto Portugal