Temperature-Aware Dynamic Voltage Scaling to Improve Energy Efficiency of Near-Threshold Computing

Abstract

Power and energy reduction is of uttermost importance for applications with stringent power/energy budget such as ultralow power and energy-harvested systems. Aggressive voltage scaling and in particular near-threshold computing is a promising approach to reduce the power and energy consumption. However, reducing the supply voltage leads to drastic performance variation induced by process and runtime variation. Temperature variation is one of the major sources of performance variation. In this paper, we study the impact of temperature variation on the circuit behavior in the near-threshold voltage region and show that the ambient temperature has a huge impact on the metrics such as circuit delay, power, and energy consumption. We also propose a low-cost, ambient temperature-aware voltage scaling technique to reduce the unnecessary energy overhead caused by temperature variation.

Existing System

Adaptive Body-Bias and Ultra-Low-Power Techniques.

Proposed System

Power and energy are the major design constraints, especially for applications with stringent power and energy budget such as ultralow power systems. To reduce the power and energy consumption, aggressive supply voltage scaling such as NTC is a promising approach. However, the behavior of an NTC circuit is very sensitive to process and runtime variation and in particular temperature variation. In this paper, we comprehensively investigated the effect of temperature variation on circuit power, performance, and energy. We showed that the circuit temperature is very close to the ambient temperature in NTC. Based on this observation, we proposed a low cost, ambient temperature-aware voltage scaling to reduce the unnecessary energy overhead caused by the ambient temperature variation.

Architecture

BLOCK DIAGRAM