ePrints.FRI - University of Ljubljana, Faculty of Computer and Information Science

Parallel string matching algorithms

Tomaž Hočevar (2012) Parallel string matching algorithms. EngD thesis.

Download (514Kb)


    This thesis presents different string searching algorithms. The string searching or string matching problem is one of the most basic problems on strings. It resurfaced with the development of bioinformatics and the need for DNA sequence analysis. It also presents a foundation for solving other, more complex string problems. First, we describe classical algoritms with linear time complexity such as Knuth-Morris-Pratt and Rabin-Karp. Then we turned our attention to the possibilites of parallelization. We present a simple parallelization scheme which finds the pattern in a string of size N in O(sqrt(N)) time on O(sqrt(N)) processors and a more complex Vishkin algoritm which solves it in O(log(N)) time. The algorithms were compared on real as well as on degenerate test cases. They were implemented in C++ and with the use of OpenMP library for parallelization. We determined that the basic algorithms were sufficient for most practical purposes. The degenerate cases can also be solved efficiently with sequential linear time algorithms. However, the experiments on parallelization showed that multicore computers these days differ too much from computation models to reach the expected theoretic speedup.

    Item Type: Thesis (EngD thesis)
    Keywords: string, substring, search, algorithm, parallel, hash
    Number of Pages: 48
    Language of Content: Slovenian
    Mentor / Comentors:
    Name and SurnameIDFunction
    prof. dr. Borut Robič28Mentor
    Link to COBISS: http://www.cobiss.si/scripts/cobiss?command=search&base=50070&select=(ID=00009390420)
    Institution: University of Ljubljana
    Department: Faculty of Computer and Information Science
    Item ID: 1792
    Date Deposited: 11 Sep 2012 15:06
    Last Modified: 24 Sep 2012 12:37
    URI: http://eprints.fri.uni-lj.si/id/eprint/1792

    Actions (login required)

    View Item