A high performance linear equation solver on the VPP500 parallel supercomputer

Author

Nakanishi, Makoto ; Ina, Hiroshi ; Miura, Kenichi

Author_Institution

Middleware Div. III, Fujitsu Ltd., Shizuoka, Japan

fYear

1994

fDate

14-18 Nov 1994

Firstpage

803

Lastpage

810

Abstract

This paper describes the implementation of two high performance linear equation solvers developed for the Fujitsu VPP500, a distributed memory parallel supercomputer system. The solvers take advantage of the key architectural features of VPP500: scalability for an arbitrary number of processors up to 222 processors; flexible data transfer among processors provided by a crossbar interconnection network; vector processing capability on each processor; and overlapped computation and data transfer. The general linear equation solver based on the blocked LU decomposition method achieves 120.0 GFLOPS performance with 100 processors in the LINPACK Highly Parallel Computing benchmark

Keywords

distributed memory systems; linear algebra; mathematics computing; parallel algorithms; performance evaluation; reconfigurable architectures; vector processor systems; 120 GFLOPS; Fujitsu VPP500; Highly Parallel Computing benchmark; LINPACK; VPP500 parallel supercomputer; blocked LU decomposition method; computation; crossbar interconnection network; data transfer; distributed memory parallel supercomputer; flexible data transfer; high performance linear equation solver; performance; scalability; vector processing; Concurrent computing; Distributed computing; Drives; Equations; Middleware; Multiprocessor interconnection networks; Parallel processing; Scalability; Supercomputers; Vector processors;

fLanguage

English

Publisher

ieee

Conference_Titel

Supercomputing '94., Proceedings

Conference_Location

Washington, DC

Print_ISBN

0-8186-6605-6

Type

conf

DOI

10.1109/SUPERC.1994.344346

Filename

344346