HPC on carbon emission trends based on grey forecasting and Topsis method

Abstract

This study investigates the carbon footprint of high-performance computing (HPC) systems by developing integrated models to estimate energy consumption and associated carbon emissions. Using data from the 2023 World Energy Statistical Yearbook, we assess the global energy consumption of HPC systems under different utilization scenarios and predict future trends through grey forecasting. Our results show that HPC systems currently contribute to significant carbon emissions, with Asia being a major contributor. By incorporating the TOPSIS method with entropy weighting, we evaluate the environmental impact of transitioning to renewable energy sources and propose actionable recommendations to mitigate HPC's carbon footprint. This research highlights the urgent need for sustainable practices in HPC to align with global climate goals and provides valuable insights for policymakers and industry stakeholders.