A High Efficient and Fast kNN Algorithm Based on CUDA

The k Nearest Neighbor (kNN) algorithm is an effective classification approach in the statistical methods of pattern recognition. But it could be a rather time-consuming approach when applied on massive data, especially facing large survey projects in astronomy. NVIDIA CUDA is a general purpose parallel computing architecture that leverages the parallel compute engine in NVIDIA graphics processing units (GPUs) to solve many complex computational problems in a fraction of the time required on a CPU. In this paper, we implement a CUDA-based kNN algorithm, and compare its performance with CPU-only kNN algorithm using single-precision and double-precision datatype on classifying celestial objects. The results demonstrate that CUDA can speedup kNN algorithm effectively and could be useful in astronomical applications.