SPECrate2006: Alternatives Considered, Lessons Learned

  • John L. Henning
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 5419)

Abstract

Since 1992, SPEC has used multiple identical benchmarks to measure multi-processor performance. This “Homogeneous Capacity Method” (aka “SPECrate”) has been criticized on the grounds that real workloads are not homogeneous. Nevertheless, SPECrate provides a useful window into how systems perform when stressed by multiple requests for similar resources. This paper reviews SPECrate’s history, and several performance lessons learned using it: (1) a 4:1 performance gain for startup of a benchmark when I/O was reconfigured; (2) a benchmark that improved up to 2:1 when a TLB data structure was re-sized; and (3) a benchmark that improved by 52% after a change to NUMA page allocation. The SPEC CPU workloads usefully exposed several opportunities for performance improvement.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Carlton, A.: CINT92 and CFP 92 Homogeneous Capacity Method Offers Fair Measure of Processing Capacity, http://www.spec.org/cpu92/specrate.txt
  2. 2.
    Chew, J.: Memory Placement Optimization (MPO), http://opensolaris.org/os/community/performance/mpo_overview.pdf
  3. 3.
    Gove, D.: CPU2006 Working Set Size. ACM SIGARCH Computer Architecture News 35(1), 90–96 (2007), http://www.spec.org/cpu2006/publications/ CrossRefGoogle Scholar
  4. 4.
    Henning, J.L.: SPEC CPU Suite Growth: An Historical Perspective. ACM SIGARCH Computer Architecture News 35(1), 65–68 (2007), http://www.spec.org/cpu2006/publications/ CrossRefMathSciNetGoogle Scholar
  5. 5.
    McGhan, H.: Niagara 2 Opens the Floodgates. Microprocessor Report (November 6, 2006), http://www.sun.com/processors/niagara/M45_MPFNiagara2_reprint.pdf
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
    Sun Microsystems, UltraSPARC T2 Processor, http://www.sun.com/processors/UltraSPARC-T2/datasheet.pdf
  13. 13.
    Sun Microsystems, UltraSPARCT2 Supplement to the UltraSPARC Architecture 2007, section 12.2 (2007), http://opensparc-t2.sunsource.net/specs/UST2-UASuppl-current-draft-P-EXT.pdf
  14. 14.
    Sun Microsystems, Solaris 10 Reference Manual Collection, http://docs.sun.com/app/docs/doc/816-5166/trapstat-1m?a=view
  15. 15.
  16. 16.
    Sun Microsystems, Solaris Tunable Parameters Reference Manual, http://docs.sun.com/app/docs/doc/817-0404
  17. 17.
    STREAM: Sustainable Memory Bandwidth in High Performance Computers, http://www.cs.virginia.edu/stream/
  18. 18.
    Tufte, E.R.: The Visual Display of Quantitative Information, pp. 107–121. Graphics Press, Chesire (1983)Google Scholar
  19. 19.
    Weicker, R.P., Henning, J.L.: Subroutine Profiling Results for the CPU2006 Benchmarks. ACM SIGARCH Computer Architecture News 35(1), 102–111 (2007), http://www.spec.org/cpu2006/publications/ CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2009

Authors and Affiliations

  • John L. Henning
    • 1
  1. 1.Sun Microsystems 

Personalised recommendations