Parallel optical architectures for some comparison-based problems

(1992) Parallel optical architectures for some comparison-based problems. Masters thesis, King Fahd University of Petroleum and Minerals.

[img]
Preview
PDF
9992.pdf

Download (2MB) | Preview

Arabic Abstract

-

English Abstract

Parallel computing requires architectures that are radically different from the inherently sequential computer architectures of today. In addition to processor-level parallelism of parallel computation models, where many processors collectively and concurrently work on a given task, parallelism at the basic instruction-level within each individual processor, is equally sought. While the former objective is achievable through different parallel computation models i.e., SIMD, MIMD models, the latter is difficult to achieve in the electronic domain. Optics provides parallelism at the basic instruction-level due to its massive inherent parallelism, large space-bandwidth product, and cross-talk-free free-space connectivity. In this study, attempts have been made to combine the field of Parallel Computing and Optical Computing to solve some comparison-based problems. Some parallel optical architectures along with an Optical Vector Processor and associated algorithms are developed to solve problems that are based on comparison. Optical implementation of Clos Nonblocking-Multicast-Switching Network and an O(1) network control algorithm have been developed. Optical Reconfigurable Bus System has been developed to solve unary-sorting problem in O(1) time. The class of Binary-Tree-Computations has been simulated on optical architectures along with algorithms for Parallel Select and Parallel Quicksort problems.

Item Type: Thesis (Masters)
Subjects: Computer
Department: College of Computing and Mathematics > Information and Computer Science
Committee Advisor: Ghanta, Subbarao
Committee Members: Al-Darwish, Hussain Nasir and Guizani, Mohsin and Akyildiz, E.
Depositing User: Mr. Admin Admin
Date Deposited: 22 Jun 2008 13:53
Last Modified: 01 Nov 2019 13:53
URI: http://eprints.kfupm.edu.sa/id/eprint/9992