期刊文献+

两级分配多可用重命名寄存器 被引量:2

Multi-Usable Rename Register with 2-Level Renaming and Allocating
下载PDF
导出
摘要 提出了一种新的寄存器重命名机制———两级分配多可用重命名寄存器,简称2L-MuRR,其特点如下:(1)对重命名寄存器文件(RRF)的占用要经过两级重命名和分配过程,减少了无效占用时间.(2)RRF被划分为多个不同宽度的字段,根据目标值的大小进行按字段分配,这样使每个重命名寄存器都是多可用的.(3)同一重命名寄存器内的多个目标值的写(或读)操作可以合并成一次完成,减少了RRF的访问次数和读/写端口的压力.模拟实验表明,2L-MuRR显著提高了RRF的存储能力(70%以上),能以较少的重命名寄存器满足SMT高并行度的要求,有效缓解了寄存器文件的设计压力. Great challenges must be overcome for designing the register file for SMT architecture. This paper proposes a novel register renaming scheme—Multi-usable Rename Register with 2-Level Renaming and Allocating (2L-MuRR), which features as follows. First, each renamed value undergoes 2-level of renaming and allocating before write-back to the Rename Register File (RRF), reducing useless occupancy. Second, the RRF is partitioned into several fields with different widths; either a single field or a field combination can be allocated to hold a renamed value, making each rename register multi-usable. Third, each access to a rename register may read or write multiple values in it, alleviating the pressure on the read/write ports. The simulations show that 2L-MuRR improves the RRF capacity greatly, over 70%. By using 2L-MuRR, SMT can achieve the same high performance with fewer rename registers, effectively reducing the design pressure on the register file.
出处 《计算机学报》 EI CSCD 北大核心 2006年第10期1729-1739,共11页 Chinese Journal of Computers
基金 国家"十五"预研基金(41316.1.2) 国家自然科学基金(60503015) 教育部博士点基金(20020213017)资助.
关键词 同时多线程 寄存器重命名 高性能 多可用 SMT register renaming high-performance multi-usable
  • 相关文献

参考文献23

  • 1Sima D..The design space of register renaming techniques.IEEE Micro,2000,20(5):70~83
  • 2Lipasti M.H.,Mestan B.R.,Gunadi E..Physical register inlining.In:Proceedings of the 31st International Symposium on Computer Architecture,Munich,2004,325~335
  • 3Balasubramonian R.,Dwarkadas S.,Albonesi D.H..Reducing the complexity of the register file in dynamic superscalar processors.In:Proceedings of the 34th IEEE/ACM International Symposium on Microarchitecture,Austin,2001,237~248
  • 4Postiff M.,Greene D.,Raasch S.et al.Integrating superscalar processor components to implement register caching.In:Proceedings of the International Conference on Supercomputing,Serento,2001,348~357
  • 5González A.,González J.,Valero M..Virtual-physical registers.In:Proceedings of the 4th International Symposium on High-Performance Computer Architecture,Los Alamitos,1998,175~184
  • 6Yang H.,Cui G.,Yang X.Z..Eliminating inter-thread interference in register file for SMT processors.In:Proceedings of the 6th International Conference on Parallel and Distributed Computing,Applications and Technologies.Dalian,2005,40~45
  • 7Cruz J.L.,González A.,Valero M.,Topham N.P..Multiple-banked register file architectures.In:Proceedings of the 27th International Symposium on Computer Architecture,Vancouver,2000,316~325
  • 8容红波,汤志忠.一种支持多重循环软件流水的寄存器结构[J].软件学报,2000,11(3):401-409. 被引量:1
  • 9Martin M.M.,Roth A.,Fischer C.N..Exploiting dead value information.In:Proceedings of the 30th IEEE/ACM International Symposium on Microarchitecture,Triangle Park,1997,125~135
  • 10Lo J.L.,Parekh S.S.,Eggers S.J.et al.Software-directed Register deallocation for simultaneous multithreaded processors.IEEE Transactions on Parallel and Distributed Systems,1999,10(9):922~933

二级参考文献8

  • 1汤志忠,王雷,钱江.多重循环的软件流水技术[J].软件学报,1996,7(7):422-427. 被引量:1
  • 2Wolfe A,Proc 4th Int Conf Architecture Support for Programming Languages and Operating Systems MAUSA,1991年,2页
  • 3Su B G,Microprocessing Microprogramming,1993年,39卷,1期,25页
  • 4Hwang K,Advanced Computer Architecture.Parallelism Scalability Programmability,1993年,457页
  • 5New Burn C J,Architectures and Compilation Techniques for Fine and Medium Grain Parallelism #?993年,39页
  • 6温钰洪,硕士学位论文,1991年
  • 7Yu T,Proceedings of the Conference on Advances in Parallel and Distributed Computing,1997年,345页
  • 8罗军,汤志忠,张赤红,于涛.软件流水中的一种数据分配算法[J].软件学报,1998,9(1):74-79. 被引量:1

共引文献3

同被引文献29

  • 1张承义,邓宇,王雷,等.现代处理器设计-超标量处理器基础[M].北京:电子工业出版社,2004.
  • 2Lo J L, PAREKH S S, EGGERS S J, LEVY H, TULLSEN D M. Software-directed register deallocation for simultaneous muhithreading processors [ J ]. IEEE Transactions on Parallel and Distributed System, 1999, 10 (9) : 922 - 933.
  • 3LIPASTI M H, MESTAN B R, GUNADI E. Physical register inlining [ C ]// Proceedings of the 31st International Symposium on Computer Architecture, Munich, 2004:325 - 335.
  • 4YANG Hua,Cul Gang, YANG Xiaozong. Eliminating interthread interference in register file for SMT processors [ C ]//Proceedings of the 6th International Conference on Parallel and Distributed Computing, Applications and Technologies, Dalian, 2005 : 40 - 45.
  • 5TOMASULO R M. An efficient algorithm for exploiting multiple arithmetic units [ J ]. IBM Journal, 1967, 44 (5) : 25 -33.
  • 6MCDOWEII L K, EGGERS S J, GRIBBLE S D. Improving server software support for simultaneous muhithreaded processors[ C]//Proceedings of the Ninth ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, San Diego, California, 2003:37-48.
  • 7FRANK M I. SUDS: automatic parallelization for raw processors [ D ]. Massachusetts Institute of Technology, 2003 : 1 - 67.
  • 8SWANSON S, MCDOWEII L K, SWIFT M M, EGGERS S J, LEVY H M. An evaluation of speculative instruction execution on simultaneous muhithreaded processors [ J ]. ACM Transactions on Computer Systems, 2003,21 ( 3 ) : 314 -340.
  • 9REDSTONE J, EGGERS S J, LEVY H M. Mini-threads: increasing TLP on small-scale SMT processors [ C ]// Proceedings of the Ninth International Symposium on High-Performance Computer Architecture,2003 : 19 - 30.
  • 10TULLSEN D M, EGGEGS S J, LEVY H M. Simultaneous multithreading: maximizing on-chip parallelism [ C ]// Proceedings of 22nd Annual International Symposium on Computer Architecture, 1995:392 - 403.

引证文献2

二级引证文献3

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