Abstract | ||
---|---|---|
Owing to the rising demands for resources integration, a lot of research efforts has been devoted to accelerating the data classification techniques. K-Nearest Neighbor(KNN)algorithm, as one of the most important data classification algorithms, is widely used in text categorization, predictive analysis etc. Two most vital issues of improving KNN are accelerating the speed of convergence and efficiently optimizing the parallel implementation. In this paper, we propose an efficient implementation on FPGA based heterogeneous computing system using OpenCL. An odd-even sort based KNN is designed to make full use of the parallel pipeline structure of FPGA. The results has shown that the performance of the proposed algorithm is much better than traditional GPU based KNN. |
Year | Venue | Keywords |
---|---|---|
2016 | 2016 INTERNATIONAL SOC DESIGN CONFERENCE (ISOCC) | OpenCL, KNN, Odd-even Sort, FPGA, Heterogeneous Computing |
Field | DocType | ISSN |
k-nearest neighbors algorithm,Convergence (routing),Computer science,sort,Parallel computing,Symmetric multiprocessor system,Field-programmable gate array,Odd–even sort,Data classification,Text categorization | Conference | 2163-9612 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Hai Peng | 1 | 0 | 0.68 |
Letian Huang | 2 | 38 | 7.92 |
john chen | 3 | 197 | 26.31 |