The Glasgow ScotGrid facility is now a truly heterogeneous site, with over 4,000 ARM cores representing about 20% of our compute capacity, which has enabled large-scale testing by the experiments and more detailed investigations of performance in a production environment. We present here a number of updates and new results related to our efforts to optimise power efficiency for High Energy Physics (HEP) computing. We will show updated benchmark results, including a new Figure of Merit designed to characterise the power usage during the execution of the HEPScore benchmark. We expand our HEPScore/Watt comparison to include machines with processors from the Ampere Altra and Altra Max series, NVIDIA Grace, and the most recent AMD EPYC chips. We also introduce a Frequency Scan methodology to better characterize performance/Watt trade-offs, potentially informing strategies like frequency scaling during peak hours to optimize power efficiency. Our findings contribute to advancing heterogeneous computing strategies and power efficiency optimizations in HEP, paving the way toward more sustainable hardware solutions.
Simili et al. (Wed,) studied this question.