VEDAS: an efficient GPU alternative for store and query of large RDF data sets

Abstract Resource Description Framework (RDF) is commonly used as a standard for data interchange on the web. The collection of RDF sets can form large graph which consumes time to query. It known that modern Graphic Processing Units (GPUs) be employed execute parallel programs in order speedup running time. In this paper, we propose novel representation along with query processing algorithm suitable GPU processing. Since main challenges architecture are limited memory sizes, transfer latency, and vast number cores. Our system designed strengthen use cores reduce effect transfer. We consists indices column-based ID requirement. indexing pre-upload filtering techniques then applied between host memory. add index swapping process facilitate sorting joining based given variable step size results’ storage, experimental results show our about 35% smaller than traditional NT format 40% less compared gStore. ranging from 1.95 397.03 when RDF3X gStore WatDiv test suite. achieves 578.57 62.97 LUBM benchmark RDF-3X analysis shows cases gain benefits approach.