Modeling communication overhead: MPI and MPL performance on the IBM SP2

被引:39
|
作者
Xu, ZW [1 ]
Hwang, K [1 ]
机构
[1] UNIV HONG KONG,DEPT ELECT & ELECTR ENGN,HONG KONG,HONG KONG
来源
关键词
D O I
10.1109/88.481662
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
The authors use timing experiments on the IBM SP2 to develop an overhead-quantifying method for evaluating communication performance on message-passing multicomputers. MPP designers and users can apply this method to reveal architectural bottlenecks, and to trade off between computations and communications for parallel applications for parallel applications optimization.
引用
下载
收藏
页码:9 / 23
页数:15
相关论文
共 50 条
  • [1] Modeling communication overhead: MPI and MPL performance on the IBM SP2 (vol 4, pg 13, 1996)
    Xu, ZW
    Hwang, K
    IEEE PARALLEL & DISTRIBUTED TECHNOLOGY, 1996, 4 (02): : 3 - 3
  • [2] Modeling the communication performance of the IBM SP2
    Abandah, GA
    Davidson, ES
    10TH INTERNATIONAL PARALLEL PROCESSING SYMPOSIUM - PROCEEDINGS OF IPPS '96, 1996, : 249 - 257
  • [3] Assessing the performance of the new IBM SP2 communication subsystem
    Miguel, J
    Arruabarrena, A
    Beivide, R
    Gregorio, JA
    IEEE PARALLEL & DISTRIBUTED TECHNOLOGY, 1996, 4 (04): : 12 - +
  • [4] A comparative characterization of communication patterns in applications using MPI and shared memory on an IBM SP2
    Karlsson, S
    Brorsson, M
    NETWORK-BASED PARALLEL COMPUTING: COMMUNICATION, ARCHITECTURE, AND APPLICATIONS, 1998, 1362 : 189 - 201
  • [5] THE COMMUNICATION SOFTWARE AND PARALLEL ENVIRONMENT OF THE IBM SP2
    SNIR, M
    HOCHSCHILD, P
    FRYE, DD
    GILDEA, KJ
    IBM SYSTEMS JOURNAL, 1995, 34 (02) : 205 - 221
  • [6] Evaluating and modeling communication overhead of MPI primitives on the Meiko CS-2
    Folino, G
    Spezzano, G
    Talia, D
    RECENT ADVANCES IN PARALLEL VIRTUAL MACHINE AND MESSAGE PASSING INTERFACE, 1998, 1497 : 27 - 35
  • [7] Distributed shared memory on IBM SP2
    Ramesh, S
    Lakshmi, R
    Govindarajan, R
    1997 INTERNATIONAL CONFERENCE ON PARALLEL AND DISTRIBUTED SYSTEMS, PROCEEDINGS, 1997, : 338 - 345
  • [8] Distributed Shared Memory on IBM SP2
    Motorola India Electronics Ltd, Bangalore, India
    Proc Int Conf Parallel Distrib Syst ICPADS, (338-345):
  • [9] Evaluating MPI collective communication on the SP2, T3D, and paragon multicomputers
    Hwang, K
    Wang, CM
    Wang, CL
    THIRD INTERNATIONAL SYMPOSIUM ON HIGH-PERFORMANCE COMPUTER ARCHITECTURE - PROCEEDINGS, 1997, : 106 - 115
  • [10] Utilization and predictability in scheduling the IBM SP2 with backfilling
    Feitelson, DG
    Weil, AM
    FIRST MERGED INTERNATIONAL PARALLEL PROCESSING SYMPOSIUM & SYMPOSIUM ON PARALLEL AND DISTRIBUTED PROCESSING, 1998, : 542 - 546