Power consumption of physical and virtual machines is a major challenge for small to larg-scale cloud computing data centers. Hundreds of thousands of physical and virtual machines run in cloud data center which consume huge amount of electricity. The energy efficiency and power optimization is an open research question. \\ This project aims to present a study on the energy efficiency and power consumption of physical and virtual machine. The problem statement is addressed by running different workloads while scaling the frequencies of physical and virtual machines. The tools used in the project are Stress, NUMActl and CPUfrequtils. The power is measured through an external APC power distribution unit used in data centers. The results and experiments show that running different workloads on physical and virtual machine has different effect on the energy efficiency and power consumption. The study provides a good understanding of how different workloads effects the energy efficiency and power consumption of different servers.