Energy-Efficient Algorithms in High Performance Computing: Approaches, Applications, and Innovations

HPC systems are crucial to scientific computation and industry processes however, their power consumption is a major economic and environmental concern. Efficient algorithms are important to decrease the power consumption of HPC systems, and make such systems more efficient and feasible. To that end, this survey focuses on the main algorithmic techniques aimed at enhancing energy consumption throughout HPC components, such as processors, memory, and networks. We classify these approaches into classes depending on the dynamic and static power management, workload based techniques, and domains specific optimizations. In this paper, we compare energy efficient algorithms and their trade off with the performance and the issues related to integration and scalability of the algorithms are also presented. Last, we present the directions for future work, including the development of enhanced programming models, standardized policies, and quantum computing. This survey gives a clear picture and presents the future direction of the energy-aware solutions in HPC.