Minimum-Energy Operation Via Error Resiliency

Error resiliency has demonstrated significant robustness and energy benefits in superthreshold performance-constrained applications (Shanbhag, Proc. Des. Autom. Conf., Jun. 2010). In this letter, we study the impact of error resiliency, in particular algorithmic-noise tolerance (ANT) (Hedge and Shanbhag, IEEE Trans. VLSI Syst., vol. 17, no. 8, pp. 813-823, Dec. 2001), in subthreshold energy-constrained applications where designs are operated at their minimum-energy operating point (MEOP) and error resiliency is still under-explored. We show that the MEOP in subthreshold designs can be further lowered by employing frequency overscaling (FOS) or voltage overscaling (VOS) and ANT to correct for intermittent timing errors. We demonstrate a 26% reduction in the total energy of an ANT-based filter in a commercial 130-nm CMOS process along with increased robustness to voltage variations.