The Journal of Supercomputing

, Volume 55, Issue 1, pp 28–50 | Cite as

Leakage-efficient design of value predictors through state and non-state preserving techniques

  • Juan M. Cebrián
  • Juan L. Aragón
  • José M. García
  • Stefanos Kaxiras
Article
  • 32 Downloads

Abstract

In the last decade computer engineers have faced changes in the way microprocessors are designed. New microprocessors do not only need to be faster than the previous generation, but also be feasible in terms of energy consumption and thermal dissipation. Recently, a new challenge appeared for computer engineers, the static power consumption. As process technology advances toward deep submicron, the static power component becomes a serious problem, especially for large on-chip array structures such as caches or prediction tables, and it must be taken into consideration. We can fight to reduce leakage power in two different ways: we can switch off the structure, reducing its leakage to zero but losing its contents (non-state preserving techniques), or we can lower its voltage (state preserving techniques), obtaining less savings but being able to restore the state of the structure in a reasonable time.

Data dependences are one of the key factors that limit performance in modern microprocessors. Value Prediction (VP) is a paradigm that exploits value locality in order to predict the output of an instruction, overcoming data dependences. The more accurate the predictor, the more performance is obtained, at the expense of becoming a potential source of power consumption and a thermal hot spot.

In this work we propose a leakage-efficient design of traditional Value Predictors (Stride, FCM, and DFCM) based on the fact that many VP entries remain unused during long periods of time before being eventually evicted. By applying both state and non-state preserving techniques, the unused entries are disabled obtaining substantial leakage energy reductions (50–80% depending on the configuration and predictor type).

Keywords

Power consumption Value prediction Low power Leakage Static power 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Borkar S (1999) Design challenges of technology scaling. IEEE Micro 19(4) Google Scholar
  2. 2.
    Butts JA, Sohi G (2000) A static power model for architects. In: Proc of the 33rd int symp on microarchitecture Google Scholar
  3. 3.
    Calder B, Reinman G, Tullsen DM (1999) Selective value prediction. In: Proc of the 26th int symp on computer architecture Google Scholar
  4. 4.
    Cebrián JM, Aragón JL, García JM (2007) Leakage energy reduction in value prediction through static decay. In: Proc of the int workshop on high-performance, power-aware computing HP-PAC’07 (in conjunction with IPDPS’07) Google Scholar
  5. 5.
    Cebrián JM, Aragón JL, García JM, Kaxiras S (2007) Adaptive VP decay: making value predictors leakage-efficient designs for high performance processors. In: Proc of the ACM int conference on computing frontiers (CF) Google Scholar
  6. 6.
    Flautner K et al (2002) Drowsy caches: simple techniques for reducing leakage power. In: Proc of the 29th int symp on computer architecture Google Scholar
  7. 7.
    Flynn MJ, Hung P (2005) Microprocessor design issues: thoughts on the road ahead. IEEE Micro 25(3):16–31 CrossRefGoogle Scholar
  8. 8.
    Gabbay F, Mendelson A (1997) Speculative execution based on value prediction. Technical Report 1080, Technion – Israel Institute of Technology Google Scholar
  9. 9.
    Goeman B, Vandierendonck H, de Bosschere K (2001) Differential FCM: increasing value prediction accuracy by improving table usage efficiency. In: Proc of the 7th int symp on high-performance computer architecture Google Scholar
  10. 10.
    Hu Z et al (2002) Applying decay strategies to branch predictors for leakage energy savings. In: Int conf on computer design, Sep Google Scholar
  11. 11.
    Juang P et al. (2004) Implementing branch-predictor decay using quasi-static memory cells. ACM Trans Archit Code Optim, 1 Google Scholar
  12. 12.
    Kaxiras S, Hu Z, Martonosi M (2001) Cache decay: exploiting generational behavior to reduce cache leakage power. In: Proc of the 28th int symp on computer architecture Google Scholar
  13. 13.
    Kesharvarzi A (1997) Intrinsic iddq: Origins, reduction, and applications in deep submicron low-power CMOS ICs. In: Proc of the IEEE international test conference Google Scholar
  14. 14.
    Kim NS et al (2003) Leakage current: Moore’s law meets static power. In: IEEE Computer Google Scholar
  15. 15.
    Kırman N, Kırman M, Chaudhuri M, Martínez JF (2005) Checkpointed early load retirement. In: Proc of the intl symp on high-performance comp architec Google Scholar
  16. 16.
    Li Y et al (2004) State-preserving vs non-state-preserving leakage control in caches. In: Proc of the DATE conference Google Scholar
  17. 17.
    Lipasti M, Wilkerson C, Shen J (1996) Value locality and load value prediction. In: Proc of the 7th int conf on architectural support for programming languages and operating systems Google Scholar
  18. 18.
    Powell MD et al (2000) Gated-Vdd: a circuit technique to reduce leakage in deep-submicron cache memories. In: Proc of the int symp on low power electronics and design Google Scholar
  19. 19.
    Sazeides Y, Smith JE (1997) The predictibility of data values. In: Proc of the 30th annual international symposium of microarchitecture Google Scholar
  20. 20.
    Yang S et al (2001) An integrated circuit/architecture approach to reducing leakage in deep-submicron high-performance I-Caches. In: Proc of the 7th int symp on high-performance computer architecture Google Scholar
  21. 21.
    Zhang Y, Paritkh D, Sankaranarayanan K, Skadron K, Stan M (2003) HotLeakage: a temperature-aware model of subthreshold and gate leakage for architects Google Scholar
  22. 22.
    Zhou H, Toburen MC, Rotenberg E, Conte TM (2001) Adaptive mode-control: A static-power-efficient cache design. In Proc of the int conf on parallel architectures and compilation techniques Google Scholar
  23. 23.
    Bhargava R, Kurian L Latency and energy aware value prediction for high-frequency processors. In: Proc of the int conference on supercomputing Google Scholar

Copyright information

© Springer Science+Business Media, LLC 2010

Authors and Affiliations

  • Juan M. Cebrián
    • 1
  • Juan L. Aragón
    • 1
  • José M. García
    • 1
  • Stefanos Kaxiras
    • 2
  1. 1.Dept. of Computer EngineeringUniversity of MurciaMurciaSpain
  2. 2.Dept. of Electrical and Computer EngineeringUniversity of PatrasPatrasGreece

Personalised recommendations