DOI

Spreadsheets are used extensively for calculations in several domains, especially in finance and insurance. Spreadsheets offer a clear benefit to their users: they are an easy to learn application in which to express their business needs, however, there are downsides too. Like software, spreadsheets can have a long life span in which they are used by several people. This leads to maintainability issues, including errors, but also often to issues with performance. In this paper we present a case study in which a model for shortfall calculations, originally implemented in a spreadsheet, was adapted to run on an HPC cluster. We present the design, analysis and implementation of the solution which clearly improved the performance of the spreadsheet, with a factor of 50 in some cases. We subsequently reflect on challenges related to reverse engineering, testing and scalability. Finally, we identify opportunities that would provide automatic support to refactoring, dependency recognition and performance profiling in future spreadsheet optimization projects.
Original languageEnglish
Title of host publication2016 IEEE 23rd International Conference on Software Analysis, Evolution, and Reengineering, SANER 2016
EditorsA. Jiu
Place of PublicationPiscataway, NJ
PublisherIEEE
Pages673-677
Number of pages5
Volume1
ISBN (Electronic)978-1-5090-1855-0
DOIs
StatePublished - 23 May 2016
EventSANER 2016 - Osaka, Japan

Conference

ConferenceSANER 2016
CountryJapan
CityOsaka
Period14/03/1618/03/16

    Research areas

  • simulation models, spreadsheet optimization, large-scale, high-performance

ID: 13969270