In this paper, we present an adaptive voltage scaling (AVS) scheme to tune the supply voltage of digital circuits according to variations. Compared to worst-case designs, which produce fixed and excessively large safety margins, a considerable amount of energy can be saved by this approach. The AVS technique is based on in-situ delay monitoring, i.e. observing the timing in critical paths. For this task, we propose a PreError flip-flop, that is capable of detecting late data transitions -so-called pre-errors. We provide an in-depth analysis, that is based on a Markov model, to describe the closed loop voltage regulation. We simulated the power saving potential compared to the worst-case design and obtained a reduction of 13.5% in active energy for a negligible error rate of 1E-15. Moreover, we illustrate the opportunity to further reduce the power consumption when tolerating higher error rates. This way, our approach can gain the optimal power saving for a given allowed failure probability.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.