The cost of energy is a substantial part of the overall costs of simulations in high- performance computing (HPC). Reducing energy consumption leads to reduced costs and correlates to lower greenhouse emissions. With climate change and increasing costs for energy as pressing problems right now, many HPC centers and companies strongly emphasize reducing energy consumption. Those energy optimizations are carried out at di↵erent levels, like switching to a sustainable energy source for computing and minimizing heat loss. This study focuses on node-level energy optimization by selecting the most energy-ecient algorithm. Achieving this requires integrating a portable energy measurement method into the molecular dynamics simulation framework. Therefore, we introduce the software library ”PMT: Power measurement toolkit” and how it can be used in the node-level auto- tuned particle simulation library AutoPas to ensure an accurate way of measuring energy consumption on di↵erent hardware platforms. Additionally, we will show that in most cases, a reduced runtime comes hand in hand with a reduction in energy consumption. Finally, we present new energy metrics that can be used to compare performance across di↵erent hardware and highlight a case study between AMD and INTEL-based clusters.