Accelerating Keyword Search for Large RDF Data on Many-Core Systems

Choksuchat C, Chantrapornchai C, Haidl M, Gorlatch S

Abstract

Resource Description Framework (RDF) is the commonlyused format for Semantic Web data. Nowadays, huge amounts of dataon the Internet in the RDF format are used by search engines for pro-viding answers to the queries of users. Querying through big data needssuitable searching methods supported by a very high processing power,because the traditional, sequential keyword matching on a semantic webserver may take a prohibitively long time. In this paper, we aim at ac-celerating the search in big RDF data by exploiting modern many-corearchitecture based on Graphics Processing Units (GPUs). We developseveral implementations of RDF search for many-core architectures us-ing two programming approaches: OpenMP for systems with CPUs andCUDA for systems comprising CPUs and GPUs. Experiments show thatour approach provides a speedup of 20.5 times over the sequential search.