全部 标题 作者
关键词 摘要

OALib Journal期刊
ISSN: 2333-9721
费用:99美元

查看量下载量

相关文章

更多...
电子学报  2012 

HIP:一种基于提示值索引的间接转移预测技术

DOI: 10.3969/j.issn.0372-2112.2012.08.005, PP. 1523-1531

Keywords: 微处理器,间接转移预测,高能效

Full-Text   Cite this paper   Add to My Lib

Abstract:

随着模块化程序和面向对象语言的发展,间接转移预测已成为影响处理器性能的瓶颈.本文提出了一种基于提示值索引的间接转移预测技术(Hint-indexedIndirect-branchPrediction,HIP).该技术将间接转移指令目标地址保存在BTB中,每个目标地址使用一个提示值与之对应.HIP技术在进行间接转移预测时,首先使用已有的分支方向预测器区分不同的间接转移场景,并获得其中存储的提示值.该提示值与指令地址进行计算,生成一个虚拟地址,用于索引存储间接转移目标地址的BTB项.实验表明,该技术可以显著提高性能,而且不需额外的大容量存储结构.与常用的基于BTB的结构相比,HIP技术可将基础处理器性能提升20.38%.与已有基于硬件的方法相比,HIP比VPC方法性能提高8.66%,并且可以获得与48KBTTC预测器相同的性能提升效果.HIP还可以将处理器能耗平均降低14.34%.

References

[1]  P-Y Chang,E Hao,Y N Patt.Target prediction for indirect jumps.Proc of 24th Ann Int’l Symp Computer Architecture.Denver,Colorado,USA:ACM Express,1997.274-283.
[2]  S Gochman,et al.The Intel Pentium M processor:Microarchitecture and performance [J].Intel Technology Journal,2003,7(2):21-59.
[3]  O Azizi,A Mahesri,BC Lee,SJ Patel,M.Horowitz.Energy-performance tradeoffs in processor architecture and circuit design:a marginal cost analysis.Proc of 37th Ann Int’l Symp Computer Architecture.Saint-Malo,France:ACM Express,2010.26-36.
[4]  SPEC.Standard performance evaluation corporation.http://www.spec.org,2011.
[5]  K Driesen,U H?lzle.The cascaded predictor:Economical and adaptive branch target prediction.Proc of the 31st Annual Int’l Symp.Microarchitecture.Dallas,Texas,USA:ACM Express,1998.249-258.
[6]  A Seznec,P Michaud.A case for (partially) TAgged GEometric history length branch prediction [J].Journal of Instruction-Level Parallelism (JILP),2006,8(1):1-23.
[7]  M.Wolczko.Benchmarking Java with the Richards benchmark.http://research.sun.com/people/mario/java_benchmarking/richards/richards.html,2011.
[8]  E Perelman,G Hamerly,B Calder.Picking statistically valid and early simulation points.Proc of the 12th Int’l Conference on Parallel Architectures and Compilation Techniques.New Orleans,Louisiana,USA:ACM Express,2003.244-255.
[9]  S Thoziyoor,N Muralimanohar,J H Ahn,N P Jouppi.CACTI 5.1.Technical Report HPL-2008-20,2008.
[10]  沈立,王志英,鲁建壮,戴葵.基于控制流的混合指令预取[J].电子学报,2003,31(8):1141-1144. Shen Li,Wang Zhi-ying,Lu Jian-zhuang,Dai Kui.Hybrid instruction prefetch based on control low [J].Acta Electronica Sinica,2003,31(8):1141-1144.(in Chinese)
[11]  H Kim,J A Joao,O Mutlu,C J Lee,Y N Patt,R Cohn.VPC prediction:Reducing the cost of indirect branches via hardware-based dynamic devirtualization.Proc of 34th Ann.Int’l Symp.Computer Architecture.San Diego,CA,USA:ACM Express,2007.424-435.
[12]  B Calder,D Grunwald,B Zorn.Quantifying behavioral differences between C and C+ + programs [J].Journal of Programming Languages,1995,2(4):323-351.
[13]  T-Y Yeh,Y N Patt.Two-level adaptive training branch prediction.Proc of the 24st Annual Int’l Symp.Microarchitecture.Albuquerque,New Mexico,USA:ACM Express,1991.51-61.
[14]  T Austin,E Larson,D Ernst.SimpleScalar:An infrastructure for computer system modeling [J].IEEE Computer,2002,35(2):59-67.
[15]  D Brooks,V Tiwari,M Martonosi.Wattch:A framework for architectural-level power analysis and optimizations.Proc.of 27th Ann Int’l Symp Computer Architecture.Vancouver,BC,Canada:ACM Express,2000.83-94.
[16]  M U Farooq,L Chen,L K John.Value based BTB indexing for indirect jump prediction.Proc of the 16th High Performance Computer Architecture.Bangalore,India:IEEE Society,2010.1-11.

Full-Text

Contact Us

service@oalib.com

QQ:3279437679

WhatsApp +8615387084133