Communication studies of single-threaded and multithreaded distributed-memory machines

Author

Sohn, Andrew ; Paek, Yunheung ; Ku, Jui-Yuan ; Kodama, Yuetsu ; Yamaguchi, Yoshinori

Author_Institution

Dept. of Comput. & Inf. Sci., New Jersey Inst. of Technol., Newark, NJ, USA

fYear

1999

fDate

9-13 Jan 1999

Firstpage

310

Lastpage

314

Abstract

This report explicates the communication overlapping capabilities of three distributed-memory machines, SGI/Cray T3E, IBM SP-2 with wide nodes, and the ETL EM-X. Bitonic sorting and Fast Fourier Transform are selected for experiments. Various message sizes are used to determine when, where, how much and why the overlapping takes place. Experimental results with up to 64 processors indicated that the communication performance of EM-X is insensitive to various message sizes while SP-2 is the most sensitive. T3E stayed in between. The EM-X gave the highest communication overlapping capability while T3E did the lowest. The experimental results are compared with the analytical results based on LogP and LogGP communication models

Keywords

distributed memory systems; fast Fourier transforms; ETL EM-X; IBM SP-2; LogGP communication models; LogP; SGI/Cray T3E; bitonic sorting; communication overlapping capabilities; communication studies; distributed-memory machines; fast Fourier transform; Computer architecture; Delay; Fast Fourier transforms; Image segmentation; Lapping; Multithreading; Performance analysis; Sorting; Switches; Yarn;

fLanguage

English

Publisher

ieee

Conference_Titel

High-Performance Computer Architecture, 1999. Proceedings. Fifth International Symposium On

Conference_Location

Orlando, FL

Print_ISBN

0-7695-0004-8

Type

conf

DOI

10.1109/HPCA.1999.744384

Filename

744384