Efficient hardware acceleration of recommendation engines: A use case on collaborative filtering

Abstract

Recommendation engines are widely used in order to predict the rating that a user would give to an item based on the user’s past behavior. Modern recommendation engines are based on computational intensive algorithms like collaborative filtering that needs to process huge sparse matrices in order to provide efficient results. This paper presents a novel scheme for the acceleration of Alternating Least Squares-based (ALS) collaborative filtering for recommendation engines that can be used to speedup significantly the processing time and also reduce the energy consumption of computing platforms. The proposed scheme is implemented in reconfigurable logic and is mapped to the Pynq platform that is based on an all-programmable MPSoC Zynq system. The hardware acceleration is integrated with the Spark framework and evaluated on real benchmarks from movielens. The performance evaluation shows that the proposed scheme can achieve up to 120x kernel speedup and up to 12x energy-efficiency compared to the embedded ARM processor of Zynq.