Power Consumption of Instruction Encodings on Cortex-M4

Abstract

Energy efficiency is increasingly important with wider use of batterypowered devices. There are many factors involved in the power consumption of instructions, such as encodings. The energy implication of these factors is not clear, and needs to be evaluated. The main contribution of this work is the characterization of power consumption of some instruction encodings of ARM v7-M on Cortex-M4 MCU STM32L476VGT6. I also designed and implemented a self-contained power measurement infrastructure on 32L476GDISCOVERY. I found that the encodings of instructions could affect the energy consumption of the MCU. The results of this work can make application programmers and hardware vendors be aware of the energy characteristics of programs and devices. Future work of a more detailed model could allow us to predict the energy consumption when designing new systems. Show more