期刊文献+

同时多线程结构的线程预构 被引量:2

Simultaneous Multi-Threading Architecture Using Novel Thread Pre-Construction Concept
下载PDF
导出
摘要 同时多线程结构中,多个线程共享一个取指单元。要选择最合适的取指线程,需要在调度策略里综合考虑流水线反馈条件或者预测流水线行为,这是非常困难的。提出了线程预构的概念,把线程取指调度分解为线程预构调度和就绪线程取指调度,分别解决错误路径取指问题和指令队列阻塞问题。提出了线程缓冲的链表结构,实现了线程预构的概念,与国际公认的基准SMT结构相比,IPC平均提高9.39%。 Simultaneous Multi-Threading (SMT), which allows multiple threads to share all the processor's resources every cycle, can exploit both the instruction level parallelism and thread levelparallelism. Instruction fetch policy is important for acquiring higher performance. To choose the most appropriate thread from all active threads, the thread fetch schedule policy must take into account the feedbacks and also the predicted behavior of different pipeline stages. As the fetch schedule mechanism is very complex, its implementation presents great difficulties. So we propose a novel thread pre-construction concept for SMT processor to decompose the thread fetch schedule into two related schedules. The threadpre-costruction schedule focuses on reducing wrong path instruction ratio;the ready thread fetch schedule focuses on reducing Instruction Queue(IQ)clog.Also,athread buffer structure is proposed to implenent the thread pre-construction mechanism.Simulation results obtained by SMTSIM tool set ,shown in Figs,3,4,and 5 in the full paper,show preliminarily that using the novel thread pre-construction concept improves the processor's performance by 9.39%over that of the standard SMTprocessor with ICOUNT2.8 policy.
作者 张盛兵 王晶
出处 《西北工业大学学报》 EI CAS CSCD 北大核心 2007年第2期159-163,共5页 Journal of Northwestern Polytechnical University
基金 国家自然科学基金(60573107) 新世纪优秀人才支持计划资助
关键词 同时多线程处理器 取指策略 线程预构 线程缓冲区 线程调度 Simultaneous Multi-Threading (SMT), instruction fetch, thread pre-construction, threadbuffer, thread fetch schedule
  • 相关文献

参考文献9

  • 1Sohi G,Roth A.Speculative Multithreaded Processors.IEEE Computer,2001,34(4):66~71
  • 2Tullsen D,Eggers S,Emer J,Levy H,Lo J,Stamm R.Exploiting Choice:Instruction Fetch and Issue on an Implementable Simultaneous Multithreading Processor.Proc of the 23rd Annual International Symposium on Computer Architecture,1996,191~202
  • 3Luo Kun,Franklin Manoj,Mukherjee Shubhendu S,Sezne Andre.Boosting SMT Performance by Speculation Control.Proc of the 15th International Parallel and Distributed Processing Symposium (IPDPS′01),2001,9~16
  • 4Limousin C,Sebot J,Vartanian A,Drach-Temam N.Improving 3D Geometry Transformations on a Simultaneous Multithreaded SIMD Processor.Proc of the ICS,2001,236~245
  • 5Tullsen D M,Brown J A.Handling Long-Latency Loads in a Simultaneous Multithreading Processor.Proc of MICRO′34,2001,172~180
  • 6Knijnenburg P M W,Ramirez A,Latorre F,Larriba J,Valero M.Branch Classification to Control Instruction Fetch in Simultaneous Multithreaded Architectures.Proc of the International Workshop on Innovative Architecture for Future Generation High-Performance Processors and Systems (IWIA′02),2002,67~76
  • 7El-Moursy Ali,Albonesi David H.Front End Policies for Improved Issue Efficiency in SMT Processors.Proc of the Ninth International Symposium on High-Performance Computer Architecture (HPCA-9′03),2003,25~33
  • 8Shin Chulho,Lee Seong-Won,Gaudiot Jean-Luc.Dynamic Scheduling Issues in SMT Architectures.Proc of the International Parallel and Distributed Processing Symposium (IPDPS′03),2003,187~192
  • 9He Liqiang,Liu Zhiyong.An Effective Instruction Fetch Policy for Simultaneous Multithreaded Processors.Proc of the 17th International Conference on High Performance Computing and Grid in Asia Pacific Region(HPCAsia′04),2004,243~248

同被引文献47

  • 1韩俊芳.缓存一致性问题研究[J].科技情报开发与经济,2005,15(14):239-240. 被引量:1
  • 2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
  • 3刘近光,梁满贵.多核多线程处理器的发展及其软件系统架构[J].微处理机,2007,28(1):1-3. 被引量:22
  • 4屈文新,樊晓桠,张盛兵.多核多线程处理器存储技术研究进展[J].计算机科学,2007,34(4):13-16. 被引量:8
  • 5Evers M, Yeh T-Y. Understanding Branches and Designing Branch Predictors for High-Performance Microprocessors[J]. Proceedings of the IEEE,2001,89(11) :1610-1620.
  • 6Kang D, Gaudiot J-L. Speculation-aware Thread Scheduling for Simultaneous Multithreading [ J]. IEE Electronics Letters, 2004,40(5) : 296-298.
  • 7Kang 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.
  • 8Falcon 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.
  • 9Tullsen 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.
  • 10Ei-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.

引证文献2

二级引证文献4

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