CPU-Memory Co-optimisation for Battery Lifetime and Age Enhancement

Abstract

Smartphones are increasingly prevalent, and they have a significant influence on user activity. Nevertheless, many smartphones on the market suffer from high power consumption and high temperature. As a result, they are losing their competitive edge in the market. Current smartphones run applications that interact with the software and hardware in various ways, depending on the type of workload. These cross-layer interactions usually cause significant changes in the chip temperature profile. Heat is considered a crucial parameter since it impacts reliability. This requires us to consider a method to improve performance while reducing the peak power level. Power control has a direct impact on on-chip temperature, which is the primary cause of ageing results. Consequently, the power and temperature management method should be implemented to permanently improve performance as well as the overall system lifespan. For this report, a study was conducted on the topic of CPU memory co-optimisation for battery lifetime and age enhancement because smartphones need to be more efficient and more powerful, and we found that many companies, such as Samsung, focus on this topic. Approximately fifty books and journals were considered, and then we conducted our experiment. In this study, up-to-date and trusted resources were chosen. The research concerned finding a method that achieves the best performance and controls the voltage/frequency balance for energy efficiency and thermal management for the CPU and memory. The experiments were conducted on several pieces of software that run on the Ubuntu system, which simulates the embedded system environment. The experiment can decrease power loss by around 30.47% and lower the temperature by approximately 0.1% for the CPU. Additionally, it lowers the memory power dissipation by 7%