Nowadays, highly portable and low-energy computing environments require programming\napplications able to satisfy computing time and energy constraints. Furthermore, collaborative\nfiltering based recommender systems are intelligent systems that use large databases and perform\nextensive matrix arithmetic calculations. In this research, we present an optimized algorithm\nand a parallel hardware implementation as good approach for running embedded collaborative\nfiltering applications. To this end, we have considered high-level synthesis programming for\nreconfigurable hardware technology. The design was tested under environments where usual\nparameters and real-world datasets were applied, and compared to usual microprocessors running\nsimilar implementations. The performance results obtained by the different implementations were\nanalyzed in computing time and energy consumption terms. The main conclusion is that the\noptimized algorithm is competitive in embedded applications when considering large datasets and\nparallel implementations based on reconfigurable hardware.
Loading....