Abstract:
In our meteorological application, the clustering algorithm was adopted for analysis and processing of radar reflectivity data.While facing problems of large scale of dataset and high dimension of feature vector, the clustering algorithm is too time-consuming to satisfy the real-time constraint in our applications.This paper proposes a parallelized clustering algorithm using GPU-CPU pipeline scheme to solve this problem.In our method, we utilized the feature of asynchronous execution between GPU and CPU, and organized the process of clustering into pipeline-style, with which we can largely exploit the parallelism in algorithm.The experimental results show that our GPU-CPU pipeline based parallelized clustering algorithm outperform normally parallelized clustering algorithm using CUDA without GPU-CPU pipeline by 38%.Compared to the serial code on CPU, out approach can achieve a 47x performance improvement, which makes it satisfy the requirements of real-time applications.