VLSI ARRAY DESIGN UNDER CONSTRAINT OF LIMITED I/O BANDWIDTH.

Philip S. Liu, Tzay Y. Young

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

Under the constraint of limited I/O bandwidth of the host system or the computing array, three configurations for the interfacing and controlling of a multiplication array to achieve optimal performance under different adverse situations are examined: multiplexing loading, processor row loading, and processor column group loading. A properly chosen configuration can significantly reduce computing time. For the inversion of covariance matrix which involves several distinct functional computations, a reconfigurable mesh connected square array is used to minimize the adverse effect of limited I/O bandwidth. Compared with a pipeline of functional computing arrays which also compute the inverse, the reconfigurable array takes much less time to accomplish the same task. A new array structure for L-U decomposition is proposed and compared with previous approaches.

Original languageEnglish
Pages (from-to)1160-1170
Number of pages11
JournalIEEE Transactions on Computers
VolumeC-32
Issue number12
StatePublished - Dec 1 1983

Fingerprint

Bandwidth
Covariance matrix
Multiplexing
Computing
Reconfigurable Mesh
Pipelines
Decomposition
Configuration
Design
Inversion
Multiplication
Distinct
Minimise
Decompose

ASJC Scopus subject areas

  • Hardware and Architecture
  • Electrical and Electronic Engineering

Cite this

VLSI ARRAY DESIGN UNDER CONSTRAINT OF LIMITED I/O BANDWIDTH. / Liu, Philip S.; Young, Tzay Y.

In: IEEE Transactions on Computers, Vol. C-32, No. 12, 01.12.1983, p. 1160-1170.

Research output: Contribution to journalArticle

Liu, PS & Young, TY 1983, 'VLSI ARRAY DESIGN UNDER CONSTRAINT OF LIMITED I/O BANDWIDTH.', IEEE Transactions on Computers, vol. C-32, no. 12, pp. 1160-1170.
Liu, Philip S. ; Young, Tzay Y. / VLSI ARRAY DESIGN UNDER CONSTRAINT OF LIMITED I/O BANDWIDTH. In: IEEE Transactions on Computers. 1983 ; Vol. C-32, No. 12. pp. 1160-1170.
@article{cd55023e90524b8a808eedb3d2a80d21,
title = "VLSI ARRAY DESIGN UNDER CONSTRAINT OF LIMITED I/O BANDWIDTH.",
abstract = "Under the constraint of limited I/O bandwidth of the host system or the computing array, three configurations for the interfacing and controlling of a multiplication array to achieve optimal performance under different adverse situations are examined: multiplexing loading, processor row loading, and processor column group loading. A properly chosen configuration can significantly reduce computing time. For the inversion of covariance matrix which involves several distinct functional computations, a reconfigurable mesh connected square array is used to minimize the adverse effect of limited I/O bandwidth. Compared with a pipeline of functional computing arrays which also compute the inverse, the reconfigurable array takes much less time to accomplish the same task. A new array structure for L-U decomposition is proposed and compared with previous approaches.",
author = "Liu, {Philip S.} and Young, {Tzay Y.}",
year = "1983",
month = "12",
day = "1",
language = "English",
volume = "C-32",
pages = "1160--1170",
journal = "IEEE Transactions on Computers",
issn = "0018-9340",
publisher = "IEEE Computer Society",
number = "12",

}

TY - JOUR

T1 - VLSI ARRAY DESIGN UNDER CONSTRAINT OF LIMITED I/O BANDWIDTH.

AU - Liu, Philip S.

AU - Young, Tzay Y.

PY - 1983/12/1

Y1 - 1983/12/1

N2 - Under the constraint of limited I/O bandwidth of the host system or the computing array, three configurations for the interfacing and controlling of a multiplication array to achieve optimal performance under different adverse situations are examined: multiplexing loading, processor row loading, and processor column group loading. A properly chosen configuration can significantly reduce computing time. For the inversion of covariance matrix which involves several distinct functional computations, a reconfigurable mesh connected square array is used to minimize the adverse effect of limited I/O bandwidth. Compared with a pipeline of functional computing arrays which also compute the inverse, the reconfigurable array takes much less time to accomplish the same task. A new array structure for L-U decomposition is proposed and compared with previous approaches.

AB - Under the constraint of limited I/O bandwidth of the host system or the computing array, three configurations for the interfacing and controlling of a multiplication array to achieve optimal performance under different adverse situations are examined: multiplexing loading, processor row loading, and processor column group loading. A properly chosen configuration can significantly reduce computing time. For the inversion of covariance matrix which involves several distinct functional computations, a reconfigurable mesh connected square array is used to minimize the adverse effect of limited I/O bandwidth. Compared with a pipeline of functional computing arrays which also compute the inverse, the reconfigurable array takes much less time to accomplish the same task. A new array structure for L-U decomposition is proposed and compared with previous approaches.

UR - http://www.scopus.com/inward/record.url?scp=0020896105&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0020896105&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0020896105

VL - C-32

SP - 1160

EP - 1170

JO - IEEE Transactions on Computers

JF - IEEE Transactions on Computers

SN - 0018-9340

IS - 12

ER -