Effective logic self repair based on extracted logic clusters

Due to predictions and first experiences with large-scale integrated systems based on nano-size features, fault tolerance and eventually even self repair have become important technologies for the implementation of long-time dependable systems. While memory self test and self-repair is state-of-the-art in system design, logic in-system repair seems to be a much more challenging problem, due to the irregular structure of logic. The paper describes a concept of logic self repair, which gains effectiveness by extracting regular sub-circuits from arbitrary logic net lists.