Mohammad Lotfi-Jam
Alan Kent
Department of Computer Science,
RMIT, GPO Box 2476V,
Melbourne 3001, Australia.
mtlj@kbs.citri.edu.au
Department of Computer Science,
RMIT, GPO Box 2476V,
Melbourne 3001, Australia.
ajk@kbs.citri.edu.au
Abstract
Information retrieval techniques are being applied to increasingly
large volumes of data.
In order to obtain answers from these larger collections in reasonable
periods of time, many researchers have been looking at how to improve
the performance of existing techniques for these collections.
This paper investigates the suitability of using the CM2, a SIMD (Single
Instruction Multiple Data) parallel
computer with 8,192 CPUs, for implementing ranking.
An existing cosine-based ranking scheme that uses compressed inverted files
was selected due to its good sequential performance. A parallel version
of this algorithm is presented and analysed.
Conference Home Page