The impact of symmetry on software distributed shared memory

被引:2
|
作者
Keleher, PJ [1 ]
机构
[1] Univ Maryland, Dept Comp Sci, College Pk, MD 20742 USA
关键词
software DSM; consistency protocols; shared memory; home-based;
D O I
10.1006/jpdc.2000.1637
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
A homeless protocol is one in which all nodes are treated identically when they access common resources. By contrast, home-based protocols assign a home or manager to each resource. Use of the resource by the home incurs less overhead than use by other processors. The key to good performance in such systems is to ensure that the asymmetry of the underlying protocol is skewed in the same way as that of the application. This paper presents a comparative evaluation of invalidation-based homeless and home-based software DSM protocols, We pay particular attention to those performance differences caused by symmetric and asymmetric features of the protocols. We then show how the picture changes when update protocols are targeted. We show that a modified home-based protocol can significantly outperform more general protocols in this application domain because of reduced protocol complexity. We further optimize our protocol by completely eliminating such memory manipulation calls from the steady-stale execution. Our resulting protocol improves average application performance by a further 34%, on top of the 19% improvement gained by our initial modification of the home-based protocol. (C) 2000 Academic Press.
引用
收藏
页码:1388 / 1419
页数:32
相关论文
共 50 条
  • [41] Working-set based adaptive protocol for software distributed shared memory
    Lee, SW
    Yoo, KY
    HIGH PERFORMANCE COMPUTING - HIPC 2000, PROCEEDINGS, 2001, 1970 : 73 - 82
  • [42] Software distributed shared memory over virtual interface architecture: Implementation and performance
    Rangarajan, M
    Iftode, L
    USENIX ASSOCIATION PROCEEDINGS OF THE 4TH ANNUAL LINUX SHOWCASE AND CONFERENCE, ATLANTA, 2000, : 341 - 352
  • [43] Accurate data redistribution cost estimation in software distributed shared memory systems
    Morris, DG
    Lowenthal, DK
    ACM SIGPLAN NOTICES, 2001, 36 (07) : 62 - 71
  • [44] Distinguishing sharing types to minimize communication in software distributed shared memory systems
    Liang, TY
    Ueng, JC
    Shieh, CK
    Chuang, DY
    Lee, JQ
    JOURNAL OF SYSTEMS AND SOFTWARE, 2000, 55 (01) : 73 - 85
  • [45] Using the resizable working-set notion for software distributed shared memory
    Lee, SW
    Yoo, KY
    CLUSTER 2000: IEEE INTERNATIONAL CONFERENCE ON CLUSTER COMPUTING, PROCEEDINGS, 2000, : 373 - 374
  • [46] Achieving performance under OpenMP on ccNUMA and software distributed shared memory systems
    Chapman, B
    Bregier, F
    Patil, A
    Prabhakar, A
    CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE, 2002, 14 (8-9): : 713 - 739
  • [47] Design of a software distributed shared memory system using an MPI communication layer
    Ojima, Y
    Sato, M
    Boku, T
    Takahashi, D
    8th International Symposium on Parallel Architectures, Algorithms and Networks, Proceedings, 2005, : 220 - 227
  • [48] Flexible hardware/software support for message passing on a distributed shared memory architecture
    Francesco, P
    Antonio, P
    Marchal, P
    DESIGN, AUTOMATION AND TEST IN EUROPE CONFERENCE AND EXHIBITION, VOLS 1 AND 2, PROCEEDINGS, 2005, : 736 - 741
  • [49] ORION: An adaptive home-based software distributed shared memory system
    Ng, MC
    Wong, WF
    SEVENTH INTERNATIONAL CONFERENCE ON PARALLEL AND DISTRIBUTED SYSTEMS, PROCEEDINGS, 2000, : 187 - 194
  • [50] Checkpointing Distributed Shared Memory
    Luis M. Silva
    João Gabriel Silva
    The Journal of Supercomputing, 1997, 11 : 137 - 158