This report is a summary of part of my internship work at TIMA laboratory, Grenoble, France. Energy management is becoming more and more important in today’s IT field, especially when dealing with battery-powered devices. Internet of Things devices in particular, require an extreme power consumption optimization and surface area reduction, while retaining the performances acceptable. One way to achieve this target is the sacrifice of computing precision thus leveraging the approximate computing paradigm. This work is therefore based on a new Multiply and Accumulation unit, designed by TIMA laboratory members, that aims to reach the above mentioned target by exploiting redundant arithmetic and on-line operators.