The power of heterogeneity in Near-Threshold Computing

One potential method to attain more energy-efficient circuits with the current technology is Near-Threshold Computing. However, this energy savings comes at a cost of increased functional failure, which necessitates that circuits must be more fault-tolerant, and thus contain more gates. Thus, achieving energy savings with Near-Threshold Computing involves properly balancing the energy used per gate with the number of gates used. We consider both the setting where the supply voltages must be homogeneous and the setting where they may be heterogeneous. We show that, for small circuit error bounds, there are many natural functions that can be computed with a log factor less energy by heterogeneous circuits than is possible with homogeneous circuits, and that this result is tight for many functions. In contrast, we show that there are relations that can be computed with a log-squared factor less energy by heterogeneous circuits than is possible with homogeneous circuits.