Green Energy HPC Data Centers to Improve Processing Cost Efficiency

Abstract

AbstractThe cost of processing in an HPC data center is one of the determining variables for its implementation and operation, with energy consumption being one of the most significant operating variables due to the high energy demand required by the different elements that make up a HPC data center. This research proposes the use of clean energy to operate HPC data centers, to allow optimization of the efficiency of the processing operation in these spaces, considering their service availability needs and the technology installed in a High performance computer data center whit a medium capacity that seams to a regular equipment installed in Latin America. The implementation of renewable energies, such as solar energy, represents an option to make the effectiveness of energy consumption more efficient in a data center, but since its availability is not stable, it is necessary to implement it alongside other energy sources that allow an uninterrupted power supply, to ensure constant data center operation. Determining the cost of HPC processing is a metric that the different HPC centers of the world seek to make more efficient in order to take advantage of the installed capacities to the maximum. This cost has different variables that largely concern the operation of the data center where the HPC equipment is housed. In this article we propose a model that projects the cost of HPC processing based on capex implementation costs and Opex operations.KeywordsHPC data centerSolar energyHPC processing cost