Parallel implementation of 2-dimensional Toeplitz solver on MasPar with applications to image restoration

Kin-wai Mak, Raymond H. Chan

Research output: Chapters, Conference Papers, Creative and Literary WorksRGC 32 - Refereed conference paper (with host publication)peer-review

Abstract

Image restoration problems can be transformed into problems of solving a linear system Tx = b where T is a block-Toeplitz or near-block-Toeplitz matrix. However, for many of these problems, the size of the matrix T is usually very large. For instance, if we are going to manipulate a 1024-by-1024 pixel image, then T will be of the order 10242-by-10242. In this paper, we implement a parallel version of our existing 2-Dimensional Toeplitz solver in a data-parallel fashion on MasPar (DECmpp) by fully utilizing its massively parallel processing power. The package we developed is portable and easy-to-use. We will demonstrate how to use our package to recover a satellite image which is blurred by atmospheric turbulence. The implementation details and performance results will also be presented.
Original languageEnglish
Title of host publicationProceedings: High Performance Computing on the Information Superhighway: HPC Asia '97
PublisherIEEE Computer Society
Pages389-394
ISBN (Print)0818679018, 0818679026
DOIs
Publication statusPublished - Apr 1997
Externally publishedYes
EventProceedings of the 1997 2nd High Performance Computing on the Information Superhighway, HPC Asia'97 - Seoul, Korea, Republic of
Duration: 28 Apr 19972 May 1997

Publication series

NameProceedings of the Conference on High Performance Computing on the Information Superhighway

Conference

ConferenceProceedings of the 1997 2nd High Performance Computing on the Information Superhighway, HPC Asia'97
PlaceKorea, Republic of
CitySeoul
Period28/04/972/05/97

Research Keywords

  • High Performance Fortran
  • Data Parallel
  • Toeplitz matrix
  • Circulant matrix

Fingerprint

Dive into the research topics of 'Parallel implementation of 2-dimensional Toeplitz solver on MasPar with applications to image restoration'. Together they form a unique fingerprint.

Cite this