Authors: Ghazanfar Ali (Texas Tech University), Sridutt Bhalachandra and Nicholas Wright (Lawrence Berkeley National Laboratory), and Alan Sill and Yong Chen (Texas Tech University)
Abstract: General-purpose graphic processing units (GPUs) are becoming increasingly important in high-performance computing (HPC) systems due to their massive computational performance. Although GPUs are high performant, modern GPU architectures consume a lot of power, making it imperative to improve their energy efficiency. The current work focuses on understanding the power consumption and performance of several GPU benchmarks under different operating conditions and on architectures. The analysis of these benchmarks that represent diverse computational patterns provides insights for future predictive models and informs procurement designs/decisions.
Best Poster Finalist (BP): no
Poster summary: PDF
Back to Poster Archive Listing