Abstract: In recent ycars,GPU has heavily shaked CPU's traditional application. It has been applied to many kinds of high performance computation, especially in the field of network security. In order to overcome the disadvantage of tra- ditional hardware acceleration, we first describe GPU's basic architecture and its parallel computing principle, and com- pare the difference between CUDA GPU programming and CPU programming. Besides, we analyze the GPU's imple- mentation of some algorithms referred to network security, and carry out experiments to test its performance according- ly. It shows that GPU parallel computing gains great advantage in computing,when optimization is reasonable.

Key words: GPU programming, Parallel computing, Network security