With the advent of heterogeneous computing systems con- sisting of multi-core CPUs and many-core GPUs, robust methods are needed to facilitate fair benchmark compar- isons between different systems. In this paper we present a benchmarking methodology for measuring a number of per- formance metrics for heterogeneous systems. Methods for comparing performance and energy efficiency are included. Consideration is given to further metrics, such as associ- ated runnings costs and even carbon emissions. We give a case study for these metrics applied to BUDE, a molecular mechanics-based docking application that has been ported to OpenCL at the University of Bristol.
|Translated title of the contribution||Energy-aware metrics for benchmarking heterogeneous systems|
|Title of host publication||ACM SIGMETRICS 1st International Workshop on Performance Modeling, Benchmarking and Simulation of High Performance Computing Systems (PMBS 10)|
|Publication status||Published - 15 Nov 2010|
Bibliographical noteName and Venue of Event: ACM / IEEE Supercomputing 2010 (SC10), New Orleans, November 13-19, 2010
Conference Organiser: ACM / IEEE