Computation and I/O benchmarks on networks of workstations

被引:0
|
作者
Olesen, S [1 ]
Sunderam, V [1 ]
机构
[1] Emory Univ, Dept Math & Comp Sci, Atlanta, GA 30322 USA
关键词
benchmark input output cluster workstations;
D O I
暂无
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
This paper examines the performance characteristics of several parallel and I/O benchmark suites implemented with the PVM and PIO US systems for network of workstation computing environments. The results are compared to previous performance reports for the same benchmarks. The benchmarks include the NAS parallel benchmarks, the NHT-1 I/O benchmarks, and the PAT benchmark. Our results indicate performance improvements that are consistent with increase in processor speeds and system memory size. In some cases we obtained performance comparable to Cray Y/MP-1 on a cluster of 12 high-end workstations with ATM interconnection, in other cases, performance within an order of magnitude of Cray Y/MP-1 was obtained.
引用
收藏
页码:705 / 708
页数:4
相关论文
共 50 条
  • [41] Unstructured mesh computations on networks of workstations
    Jones, MT
    Plassmann, PE
    COMPUTER-AIDED CIVIL AND INFRASTRUCTURE ENGINEERING, 2000, 15 (03) : 196 - 208
  • [42] I/O and computation overlap on SIMD systolic arrays
    Lavenier, D.
    Raimbault, F.
    Frison, P.
    Journal of VLSI Signal Processing Systems for Signal, Image, and Video Technology, 1995, 9 (03): : 153 - 165
  • [43] DSP-CC: I/O Efficient Parallel Computation of Connected Components in Billion-scale Networks
    Kim, Min-Soo
    Lee, Sangyeon
    Han, Wook-Shin
    Park, Himchan
    Lee, Jeong-Hoon
    2016 32ND IEEE INTERNATIONAL CONFERENCE ON DATA ENGINEERING (ICDE), 2016, : 1504 - 1505
  • [44] Performance and power analysis for high performance computation benchmarks
    Issa, Joseph
    OPEN COMPUTER SCIENCE, 2013, 3 (01): : 1 - 16
  • [45] Linear algebra computation benchmarks on a model grid platform
    Storchi, L
    Manuali, C
    Gervasi, O
    Vitillaro, G
    Laganà, A
    Tarantelli, F
    COMPUTATIONAL SCIENCE - ICCS 2003, PT II, PROCEEDINGS, 2003, 2658 : 297 - 306
  • [46] DSP-CC-: I/O Efficient Parallel Computation of Connected Components in Billion-Scale Networks
    Kim, Min-Soo
    Lee, Sangyeon
    Han, Wook-Shin
    Park, Himchan
    Lee, Jeong-Hoon
    IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, 2015, 27 (10) : 2658 - 2671
  • [47] Parallel Computation of Shallow-water Model on Workstations Cluster
    $$$$Song Junqiang
    Sun An-clang
    Li Xiaomei(epartment Of CO
    Wuhan University Journal of Natural Sciences, 1996, (Z1) : 522 - 525
  • [48] Evaluation of scheduling methods in the network of workstations with heterogeneous computation power
    Shim, YC
    Kim, JS
    INTERNATIONAL CONFERENCE ON PARALLEL AND DISTRIBUTED PROCESSING TECHNIQUES AND APPLICATIONS, VOLS I-V, PROCEEDINGS, 1999, : 2629 - 2635
  • [49] Computation time and idle time of tiling transformation on a network of workstations
    Sathe, SR
    Nawghare, PM
    INTERNATIONAL JOURNAL OF HIGH SPEED COMPUTING, 2000, 11 (03): : 129 - 149
  • [50] The GSN library and FORTRAN level I/O benchmarks on the NS-IIIHPC system
    Fujita, N
    Ookawa, H
    HIGH PERFORMANCE COMPUTING, 2003, 2858 : 458 - 467