Leveraging workload diversity through OS scheduling to maximize performance on single-ISA heterogeneous multicore systems

  • Authors:

  • Juan Carlos Saez;Daniel Shelepov;Alexandra Fedorova;Manuel Prieto

  • Affiliations:

  • Complutense University of Madrid, Facultad de Ciencias Físicas, Ciudad Universitaria s/n, Madrid 28040, Spain;Microsoft Corporation, One Microsoft Way Redmond, WA 98052-7329, USA;Simon Fraser University, School of Computing Science, 8888 University Drive, Burnaby, BC, Canada V5A 1S6;Complutense University of Madrid, Facultad de Ciencias Físicas, Ciudad Universitaria s/n, Madrid 28040, Spain

  • Venue:
  • Journal of Parallel and Distributed Computing
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Recent research has highlighted the potential benefits of single-ISA heterogeneous multicore processors over cost-equivalent homogeneous ones, and it is likely that future processors will integrate cores that have the same instruction set architecture (ISA) but offer different performance and power characteristics. To fully tap into the potential of these processors, the operating system must be aware of the hardware asymmetry when making scheduling decisions and map applications to cores in consideration of their performance characteristics. We propose a Heterogeneity-Aware Signature-Supported (HASS) scheduling algorithm that performs this mapping using per-thread architectural signatures, which are compact summaries of threads' architectural properties. We implemented HASS in OpenSolaris, and demonstrated that it always outperforms a heterogeneity-agnostic scheduler (by as much as 12.5%) for workloads exhibiting sufficient diversity. Our evaluation also includes an extensive comparison with other heterogeneity-aware schedulers to provide a more clear understanding of the pros and cons behind HASS.