A CMP Energy Consumption Estimate Model for Computer Systems
Junyan Yu · Dianzi Ke-ji Daxue xuebao · 2015
As the development of cloud computing and the emergence of large data centers, computer system energy consumption and efficiency increasingly become a focus of research. This paper aims at solving the problem of computer's energy estimation. The energy-consumption contributions of the all components in a computer are analyzed. A CMP energy consumption model which reflects the dominant monitored parameters of system energy consumption is constructed. This model could choose different dominant parameters according to the state of the computer's major components. Experiments show that during a variety of workloads and different task types, the CMP model can achieve a higher accuracy on power estimation than FAN and Cubic model, especially when computers doing data intensive tasks.