Efficient parallel implementation of sequence analysis algorithms using a global address space model

  • Authors:

  • Stanley M. Dunn;John Fedyna;Joseph E. Peters

  • Affiliations:

  • Department of Biomedical Engineering and Department of Computer Science Rutgers University, Piscataway, New Jersey 08855-0909, U.S.A. and University of Medicine and Dentistry of New Jersey Newark, ...;Department of Biomedical Engineering Rutgers University, Piscataway, New Jersey 08855-0909, U.S.A.;Department of Computer Science Rutgers University, Piscataway, New Jersey 08855-0909, U.S.A.

  • Venue:
  • Mathematical and Computer Modelling: An International Journal
  • Year:
  • 1992

Quantified Score

Hi-index 0.98

Visualization

Abstract

In this paper, we consider the problem of efficiently writing and using computer programs on large scale parelled computers. Specifically we will propose a model of computation that allows the user to write the program for a single abstract machine, and use the program on small-scale processor systems and large parallel systems alike. A good examples of problem where this migration is necessary are ssequence analysis problems: the comparison of DNA and protein sequences and the visualization of their seconfary and tertiary structure. Although many programs for these problems have been written for personal computers, these same programs cannot be used on parallel machines. This paper has three basic parts. First, there is a basic introduction to sequence analysis and the associated computational problems. Second, there is a general description of computer systems used for these sequence analysis problems. Third, the programming model of a global linear address space is presented followed by examples of its use in sequence analysis programs. These programs run effeciently on personal computers and distributed memory parallel computers. This illustrates that effeciently write one program to solve different size problems on different target architectures.