high performance computing on graphics processing units: hgpu.org

Accurate and efficient data classification techniques are of vital importance to many problems, and are rapidly developing in recent decades. K-Nearest Neighbor algorithm (KNN), as one of the most important algorithms, is widely used in text categorization, predictive analysis, data mining and image recognition, etc. To accelerate the algorithm and to optimize the parallel implementation solution are two key issues of KNN. In this paper, we propose a new solution to speed up KNN algorithm on FPGA based heterogeneous computing system using OpenCL. Based on FPGA’s parallel pipeline structure, a specific bubble sort algorithm is designed and used to optimize KNN algorithm. The results have been shown that the efficiency of the solution in our paper is much higher than conventional GPU based KNN algorithm implementation.