期刊文献+

一种同时多线程指令队列竞争缓解策略 被引量:1

A Kind of Instruction Queue Competition Easing Strategy for Simultaneous Multithreading Architecture
下载PDF
导出
摘要 同时多线程结构利用线程级并行和指令级并行的优势,提高了指令吞吐率,但线程对关键资源(如指令队列)的竞争会削弱这种优势,造成资源浪费,又会降低处理器性能.提出了指令队列利用参数,通过分析指令队列利用率与处理器性能的关系,用实验评估了在四线程情况下,典型静态指令队列竞争缓解策略(如Dwarn,2OP_Block,Static)及其组合对处理器性能的影响.给出了load依赖链模型,分析了基于load依赖链的基准程序线程特性,提出了一种结合线程特性的指令队列竞争缓解策略.实验结果表明,该策略能够加速执行指令吞吐率较高的线程,通过提升此类线程的性能使整体指令吞吐率进一步增加. The simultaneous multi-threading ( SMT ) technique boosts instructions per clock (IPC) by adopting thread level parallelism and instruction level parallelism. However, the competition of key resources between threads do weaken such advancement. Instruction queue (IQ) is proved as one key resource and its competition always results into performance degradation. Typical IQ competition easing strategies include Dwarn, 20P_Block and Static. This paper presets two IQ utilization parameters to estimate the relationships between IQ usage and system performance. Competition easing capability of typical IQ strategies and their combination are compared. A load dependency chain model isbuilt and analysis of thread characteristics based on the model is given. Then a new IQ competition easing strategy combining with thread characteristics is proposed. The experimental results show that such strategy can achieve total IPC improvement by accelerating high IPC threads.
出处 《同济大学学报(自然科学版)》 EI CAS CSCD 北大核心 2013年第12期1889-1897,共9页 Journal of Tongji University:Natural Science
基金 国家自然科学基金(60903033)
关键词 同时多线程 指令队列 load依赖链 竞争缓解策 线程特性 simultaneous multi- threading instruction queue load dependency chain competition easing strategy threadcharacteristics
  • 相关文献

参考文献15

  • 1Knijnenburg P M W,Ramirez A,Latorre F. Branch classification to control instruction fetch in simultaneous multithreaded architectures[A].Kauai:IEEE Computer Society,2002.67-76.
  • 2Raasch S E,Reinhardt S K. The impact of resource partitioning on SMT processors[A].New Orleans:IEEE Computer Society/ACM/IFIP,2003.15-25.
  • 3Joshua J Y,Ajay J,Resit S. Analyzing the processor bottlenecks in SPEC CPU 2000[A].Austin:SPEC,2006.
  • 4印杰.同时多线程处理器容错技术研究[D]上海:同济大学电子与信息工程学院,2011.
  • 5Cazorla F J,Ramirez A,Valero M. Dcache warn:an i-fetch policy to increase SMT efficiency[A].Santa Fe:IEEE Computer Society,2004.74-83.
  • 6Sharkey J J,Ponomarev D V. Efficient instruction schedulers for SMT processors[A].Austin:IEEE Computer Society,2006.288-298.
  • 7Wang H,Sangireddy R,Baldawa S. Optimizing instruction scheduling through combined in-order and O-O-O execution in SMT processors[J].{H}IEEE Transactions on Parallel and Distributed Systems,2009,(03):389.
  • 8Brekelbaum E,Rupley J,Wilkerson C. Hierarchical scheduling windows[A].Istanbul:IEEE Computer Society/ACM,2002.27-36.
  • 9Cristal A,Ortega D,Llosa J. Out-of-order commit processors[A].Madrid:IEEE Computer Society,2004.48-59.
  • 10印杰,江建慧.缓解同时多线程结构中线程对关键资源的竞争[J].计算机科学,2010,37(3):256-261. 被引量:1

二级参考文献48

  • 1何立强,刘志勇.一种具有QoS特性的同时多线程处理器取指策略[J].计算机研究与发展,2006,43(11):1980-1984. 被引量:4
  • 2张盛兵,王晶.同时多线程结构的线程预构[J].西北工业大学学报,2007,25(2):159-163. 被引量:2
  • 3Evers M, Yeh T-Y. Understanding Branches and Designing Branch Predictors for High-Performance Microprocessors[J]. Proceedings of the IEEE,2001,89(11) :1610-1620.
  • 4Kang D, Gaudiot J-L. Speculation-aware Thread Scheduling for Simultaneous Multithreading [ J]. IEE Electronics Letters, 2004,40(5) : 296-298.
  • 5Kang D, Gaudiot J-L. Speculation Control for Simultaneous Multithreading[C]// Proceedings of the 18th International Parallel and Distributed Processing Symposium. Santa Fe, New Mexico, April 2004 : 76-85.
  • 6Falcon A, Santana O J, Ramirez A, et al. Tolerating Branch Predictor Latency on SMT[C] //Proceedings of the 5^thInternational Symposium on High Performance Computing. Tokyo,Japan, October: 86-98.
  • 7Tullsen D M, Brown J A. Handling Long-latency Loads in a Simultaneous Multithreading Proeessor[C]// Proc. of the 34^th IEEE International Symposium on Mieroarchiteeture. Austin, USA,Dec 2001 : 318-327.
  • 8Ei-Moursy A, Albonesi D H. Front-End Policies for Improved Issue Efficiency in SMT Processors[C]//Proc. of the 9th International Symposium on High-Performance Computer Architecture. Anaheim,California, USA, February: 31-40.
  • 9Cazorla F J, Ramirez A, Valetor M, et al. Dcache Warn: an I- Fetch Policy to Increase SMT Efficiency[C]//Proc. of the 18^th International Symposium on Parallel and Distributed Processing. Santa Fe,New Mexico,USA, April 2004:74-83.
  • 10Suh G E, Devadas S, Rudolph L. A New Memory Monitoring Scheme for Memory-Aware Scheduling and Partitioning [C]// Proc. of the 8^th International Symposium on High-Performance Computer Architecture. Boston, Massachusetts, USA, February 2002:117-128.

共引文献2

同被引文献4

引证文献1

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部