随着计算机和网络技术应用的日益普及，网络安全问题

日显突出。来自计算机系统外部和内部的入侵者可能会对系

统进行破坏，或者试图窃取客户的个人信息。为了保证计算

机系统及信息的安全，网络入侵检测技术得到了发展和应用。-In order to optimize test efficiency of Intrusion Detection System(IDS) based on Support Vector Machine(SVM), a new intrusion

detection method based on Graphics Processing Unit(GPU) and feature selection is proposed. During the process of intrusion detection, GPU-based

parallel computing model is adopted and features of samples are reasonable selected. Experimental results demonstrate that the proposed method can

reduce time consumption in the training procedure of IDS and the performance for intrusion detection is kept as usual.