Quantifying the impact of GPUs on performance and energy efficiency in HPC clusters

Author

Enos, Jeremy ; Steffen, Craig ; Fullop, Joshi ; Showerman, Michael ; Shi, Guochun ; Esler, Kenneth ; Kindratenko, Volodymyr ; Stone, John E. ; Phillips, James C.

Author_Institution

Center for Supercomput. Applic., Univ. of Illinois at Urbana-Champaign, Urbana, IL, USA

fYear

2010

fDate

15-18 Aug. 2010

Firstpage

317

Lastpage

324

Abstract

We present an inexpensive hardware system for monitoring power usage of individual CPU hosts and externally attached GPUs in HPC clusters and the software stack for integrating the power usage data streamed in real-time by the power monitoring hardware with the cluster management software tools. We introduce a measure for quantifying the overall improvement in performance-per-watt for applications that have been ported to work on the GPUs. We use the developed hardware/software infrastructure to demonstrate the overall improvement in performance-per-watt for several HPC applications implemented to work on GPUs.

Keywords

computer graphic equipment; coprocessors; power consumption; workstation clusters; CPU hosts; GPU; HPC clusters; cluster management software tools; hardware-software infrastructure; performance-per-watt; power monitoring hardware; power usage data; power usage monitoring; software stack; Benchmark testing; Graphics processing unit; Monitoring; Power demand; Power measurement; Sensors; GPU; cluster; power efficiency; power monitoring;

fLanguage

English

Publisher

ieee

Conference_Titel

Green Computing Conference, 2010 International

Conference_Location

Chicago, IL

Print_ISBN

978-1-4244-7612-1

Type

conf

DOI

10.1109/GREENCOMP.2010.5598297

Filename

5598297