期刊文献+

支持循环自动流水线的粗粒度可重构阵列体系结构 被引量:11

原文传递
导出
摘要 可重构计算兼顾定制计算的高效性与通用计算的灵活性,是人们正在探索的一类新型计算机体系结构.文中介绍一种粗粒度可重构多核协处理器体系结构,用于加速计算密集或数据密集类算法.在该结构中采用具有猜测执行机制的循环自动流水线技术,将流水线并行执行和数据驱动执行紧密结合.采用以循环程序为单位的固定指令多数据流执行模式,在循环执行期间,每个单元固定执行一种指令,配置信息量减少,支持快速动态重构.设计了多种特殊数据传递指令,如非平衡数据复制指令、数据选择指令等,充分利用循环迭代内部、循环迭代间的数据相关,开发数据重用性,减少对存储器带宽的需求.以大容量FPGA开发板为基础,构建了8×8规模的可重构阵列实验环境.多个计算密集型程序的实际运行结果表明,该体系结构在流水线吞吐率方面可以达到每拍1个循环迭代,访问存储器的请求数仅为典型RISC处理器的3%,并且具有单拍切换配置现场的能力.
出处 《中国科学(E辑)》 CSCD 北大核心 2008年第4期579-591,共13页 Science in China(Series E)
基金 国家自然科学基金(批准号:60633050)资助项目
  • 相关文献

参考文献13

  • 1Cardoso J M P. Dynamic loop pipelining in data-driven architectures. In: Bagherzadeh N, Valero M, Ramirez A,eds. Proc 2nd Int Conf Comput Front. New York: ACM Press, 2005. 106-115.
  • 2Baumgarte V, Ehlers G, May F, et al. PACT XPP-A self-reconfigurable data processing architecture. J Super- comput, 2003, 26(2): 167--184.
  • 3Mei B, Vernalde S, Verkest D, et al. Exploiting loop-level parallelism on coarse-grained reconfigurable architectures using modulo scheduling. In: When N, Verkest Di, eds. Proc Conf Des, Aurora Test Eur. New York: IEEE Press, 2003. 10296--10301.
  • 4Barat F, Jayapala M, Beeck P O. Software pipelining for coarse-grained reconfigurable instruction set processors. In: Proc IEEE Int Conf VLSI Des. New York: IEEE Press, 2002. 338--344.
  • 5Hauser J R, Wawrzynek J. Garp: A MIPS processor with a reconfigurable coprocessor. In: Proc IEEE Int Syrup Field-Programmable Custom Computing Machines. New York: IEEE Press, 1997.16--21.
  • 6Rau B R. Iterative modulo scheduling: An algorithm for software pipelining loops. In: Proc ACM 27th Int Symp Microarchitecture. New York: IEEE Press, 1994. 63--74.
  • 7Lee M H, Singh H, Lu G, et al. Design and implementation of the MorphoSys reconfigurable computing processor. J VLSI Signal Processing-Systems for Signal, Image and Video Technology, 2000, 24:147--164.
  • 8Dennis J B, Gao G R. An efficient pipelined dataflow processor architecture. In: Proc ACM Int Conf on Supercomputing, New York: IEEE Press, 1988. 363--373.
  • 9Arvind, Nikhil R S. Executing a program on the MIT tagged-token dataflow architecture. IEEE Trans Comput, 1990, 39(3): 300--318.
  • 10Iannucci R A. Toward a dataflow/von Neumann hybrid architecture. In: Siegel H, ed. Proc Int Symp Computer Architecture. New York: IEEE Press, 1988. 131--140.

同被引文献99

  • 1陈桂茸,窦勇,徐炜遐.基于SUIF2的程序特征分析技术研究[J].计算机研究与发展,2007,44(z1):254-258. 被引量:3
  • 2Katherine Compton, Scott Hauck. Reconfigurable computing: A survey of systems and software. ACM Computing Surveys, 2002, 34(2): 171-210
  • 3Hartenstein R. A decade of reconfigurable computing: A visionary retrospective//Proceedings of the DATE2001. Munich, Germany, 2001:642-649
  • 4Shrivastava Aviral, Earlie Eugene, Dutt Nikil, Nicolau Alex. PBExplore: A framework for compiler-in-the-loop exploration of partial bypassing in embedded processors//Proceedings of the DATE 2005. Washington, 2005:1264-1269
  • 5Fan Kevin, Kudlur Manjunath, Park Hyunchul, Mahlke Scott A. Increasing hardware efficiency with multifunction loop accelerators//Proceedings of the CODES+ISSS. Seoul, Korea, 20061 276-281
  • 6Yoon Jonghee W, Shrivastava Aviral, Park Sanghyun, Ahn Minwook, Jeyapaul Reiley, Pack Yunheung. SPKM: A novel graph drawing based algorithm for application mapping onto coarse-grained reconfigurable architectures//Proceedings of the ASP-DAC. Seoul, Korea, 2008: 776-782
  • 7Lee J, Choi K et al. Mapping loops on coarse-grain reconfigurable architectures using memory operation sharing. Center for Embedded Computer Systems, University of California, Irvine: TR 02-34, 2002
  • 8Kejariwal Arun, Veidenbaum Alexander V, Nieolau Alexandru, Girkar Milind, Tian Xinmin, Saito Hideki. Challenges in exploitation of loop parallelism in embedded applications// Proceedings of the CODES+ISSS. Seoul, Korea, 2006: 173-180
  • 9Baumgarten Volker, Ehlers G, May F, Nuckel Armin, Vorbach Martin, Weinhardt Markus. PACT XPP-A self-reconfigurable data processing architecture. The Journal of Supercomputing, 2003, 26(2): 167-184
  • 10Singh H, Lee M, Lu G, Kurdahi F J, Bagherzadeh N, Filho E, Maestre R. Morhposys: Case study of a reconfigurable computing system targeting multimedia applications//Proceedings of the DAC' 00. Los Angeles, California, 2000: 573- 578

引证文献11

二级引证文献26

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