Low-Power Design of Hybrid Instruction Cache Based on Branch Prediction and Drowsy Cache

Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 246)


In the design of modern high-performance embedded processors, Branch prediction and larger capacity instruction cache greatly improves the performance of processors. They also consume a larger proportion of total power consumption of processors at the same time. There are many methods to reduce their power consumption. However, there is few method of combining them to reduce the power consumption. Based on these two technologies having a similar structure, this paper presents a hybrid cache structure of combining them to a new structure named B-Cache. By the method, we reduce the number of Comparisons of tag or PC. And based on spatial locality of instruction cache, we propose the method to reduce the number of tag comparison of instruction cache. By the experiments, we found this method compared with the common method of two-port BTB used to wake up cache line on demand can significantly improve the performance and reduce the power consumption of processors.


Branch prediction Instruction Cache B-Cache Tag 


  1. 1.
    Zheng W-m, Tang z-z (1998) Computer architecture, 2nd edn. Tsinghua University Press, Beijing, In ChineseGoogle Scholar
  2. 2.
    Chang YJ (2006) Lazy BTB: reducing BTB energy consumption using dynamic profiling. Proceedings of conference on Asia South Pacific design automation, Yokohama, 2006, 917–922. doi:  10.1109/ASPDAC.2006.1594803
  3. 3.
    Deris KJ, Baniasadi A (2006) Branchless cycle prediction for embedded processors. Proceedings of ACM symposium on applied computing, Dijon, 2006, 928–932. doi:  10.1145/1141277.1141492
  4. 4.
    K Flautner, Kim NS et al. (2002) Drowsy caches: simple techniques for reducing leakage power. In: SIGARCH proceedings of the 29th annual international symposium on computer architecture. IEEE Computer Society, Washington, 148–157. doi:  10.1109/ISCA.2002
  5. 5.
    Zhang C, Zhou HW et al. (2006) Architectural leakage power reduction method for instruction cache in ultra deep submicron microprocessors. In: The 11th Asia-Pacific computer systems architecture conference. Springer, Heidelberg, 588–594. doi: 10.1007/11859802_62Google Scholar
  6. 6.
    Hu J et al. (2003) Exploiting program hotspots and code sequentiality for instruction cache leakage management. In: International symposium on low power electronics and design (ISLPED’03). Springer, Heidelberg, 25–27. doi: 10.1109/LPE.2003.1231936
  7. 7.
    Chung SW, Skadron K (2006) Using branch prediction information for near-optimal I-Cache leakage. In: The 11th Asia-Pacific computer systems architecture conference. Springer, Heidelberg, 24–37. doi:  10.1007/11859802_4
  8. 8.
    Kim NS, Flautner K et al. (2004) Single-VDD and single-VT super-drowsy techniques for low-leakage high performance instruction caches. In: International symposium on low power electronics and design (ISLPED’04). Springer, Heidelberg, 54–57. doi:  10.1145/1013235.1013254
  9. 9.
    Kim NS et al. (2002) Drowsy instruction caches. Leakage power reduction using dynamic voltage scaling and cache sub-bank prediction. In: 35th Annual IEEE/ACM international symposium on microarchitecture (MICRO’02), Istanbul, Turkey. Springer, 219–230. doi:10.1109/MICRO.2002.1176252Google Scholar
  10. 10.
    Kim NS, Flautner K et al (2004) Circuit and microarchitectural techniques for reducing cache leakage power. IEEE T VLSI Syst 12(2):167–184. doi: 10.1109/TVLSI.2003.821550 CrossRefGoogle Scholar
  11. 11.
    Hong-wei Z, Min-xuan Z (2008) The research on power controlling policies for instruction cache with architecture level methods. Acta Ectronica Sinica 36(11):2107–2112, In ChineseGoogle Scholar
  12. 12.
    Mingyan Y, Xiangjian Z, Bing Y (2010) Low power branch target buffer design based on hopping access. Journal of Computer-Aided Design & Computer Graphics 22(4):695–702, In ChineseCrossRefGoogle Scholar
  13. 13.
    Zmily A, Kozyrakis C (2005) Energy-efficient and high-performance instruction fetch using a block-aware ISA. In: International symposium on low power electronics and design. Springer, Heidelberg, 36–41. doi: 10.1145/1077603.1077614Google Scholar

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  1. 1.Xi’an Microelectronics Technology InstituteXianChina

Personalised recommendations