On the Usefulness of Detecting Soft Errors in Parallel Pipelines for High-Speed Machine Vision Based on Organic Computing
Konferenz: ARCS 2009 - 22th International Conference on Architecture of Computing Systems
11.03.2009 in Delft, The Netherlands
Tagungsband: ARCS 2009
Seiten: 6Sprache: EnglischTyp: PDFPersönliche VDE-Mitglieder erhalten auf diesen Artikel 10% Rabatt
Komann, Marcus; Taubert, Frank; Fey, Dietmar (Friedrich-Schiller-University Jena, Germany)
Computer hardware is always in danger of being hit by ionized particles which cause different kinds of faults. In this paper, the effect of transient soft-errors on the stages of parallel pipelines is investigated by the example of emergent algorithms and machine vision. A fault injection model is introduced and the corresponding protection scheme is presented. During simulation, the number of relevant faults showed to be relatively small and the effect on the pipelines was mostly negligible. However, the costs for protection are low possibly making use of protection worthwhile.