论文信息 - A performance comparison of DSM, PVM, and MPI

A performance comparison of DSM, PVM, and MPI

We compare the performance of the Treadmarks DSM system with two popular message passing systems (PVM and MPI). The comparison is done on 1, 2, 4, 8, 16, 24, and 32 nodes. Applications are chosen to represent three classes of problems: loosely synchronous, embarrassingly parallel, and synchronous. The results show DSM has similar performance to message passing for the embarrassingly parallel class. However the performance of DSM is lower than PVM and MPI for the synchronous and loosely synchronous classes of problems. An analysis of the reasons is presented.

[1] Simon Haykin,et al. Neural Networks: A Comprehensive Foundation , 1998 .

[2] Stephen Cranefield,et al. A View-based Consistency Model based on Transparent Data Selection in Distributed Shared Memory , 2001 .

[3] David Jones. High performance , 1989, Nature.

[4] Anthony Skjellum,et al. A High-Performance, Portable Implementation of the MPI Message Passing Interface Standard , 1996, Parallel Comput..

[5] Geoffrey C. Fox,et al. Parallel Computing Works , 1994 .

[6] Jack Dongarra,et al. Pvm 3 user's guide and reference manual , 1993 .

[7] Martin K. Purvis,et al. View-based consistency and false sharing effect in distributed shared memory , 2001, OPSR.

[8] Alan L. Cox,et al. TreadMarks: shared memory computing on networks of workstations , 1996 .

[9] A. Tanenbaum. Computer recreations , 1973 .

[10] Alan L. Cox,et al. Quantifying the Performance Differences between PVM and TreadMarks , 1997, J. Parallel Distributed Comput..

[11] D. Skillicorn. Strategies for Parallelizing Supervised and Unsupervised Learning in Arti cial Neural Networks Using the BSP Cost Model , 1997 .

[12] B. Wilkinson,et al. A state-wide senior parallel programming course , 1999 .

[13] Catherine Blake,et al. UCI Repository of machine learning databases , 1998 .