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

High performance computing with WINDOWS HPC Server 2008

Dare Podgornik (2010) High performance computing with WINDOWS HPC Server 2008. EngD thesis.

[img] PDF
Download (2428Kb)


    With the development of hardware, lower prices for hardware and the huge amounts of data in daily work, systems for High Performance Computing have become greatly expanded. Even Microsoft, with its cluster management programming environment called Windows HPC Server 2008, set foot in the field of High Performance Computing. In this thesis we describe the installation process of Windows HPC Server 2008 on a cluster of computers, its configuration, monitoring and diagnostic tools and user interface. The installation and configuration of the system is very simple, installation is supported with wizard and a list of necessary tasks for system configuration. The user interface is well designed and intuitive, and greatly contributes to the usability and transparency of the system. Here we introduced the possibility of different implementations for distributed computing applications on the case of applications for the calculation of π number. In the second part of the thesis we tested capabilities of Windows HPC Server 2008 on two cases of text documents classification, using a method called Support Vector Machines (SVM). We compared these results with the results of text documents classification within Condor environment. It turns out that Windows HPC Server 2008 is faster then Condor environment when processing small amounts of data, when processing large amounts of data Condor environment provides better results.

    Item Type: Thesis (EngD thesis)
    Keywords: Windows HPC Server 2008, cluster, text document classification, High Performance Computing, Condor
    Number of Pages: 82
    Language of Content: Slovenian
    Mentor / Comentors:
    Name and SurnameIDFunction
    izr. prof. dr. Uroš Lotrič270Mentor
    Link to COBISS: http://www.cobiss.si/scripts/cobiss?command=search&base=50070&select=(ID=8067156)
    Institution: University of Ljubljana
    Department: Faculty of Computer and Information Science
    Item ID: 1200
    Date Deposited: 15 Oct 2010 13:55
    Last Modified: 13 Aug 2011 00:38
    URI: http://eprints.fri.uni-lj.si/id/eprint/1200

    Actions (login required)

    View Item