基于CUDA的AES并行算法优化

为提升高级加密标准(AES)的加密性能,利用显卡的通用计算能力,在统一计算设备架构(CUDA)平台上实现AES的128位、192位和256位3个版本的GPU并行算法,并提出优化的AES并行算法。在考虑块内线程数量、共享存储器容量和总块数的基础上,根据分块最优值的经验数据指导AES算法在GPU上的最优分块。实验结果表明,与未优化的AES并行算法相比,该算法的3个版本在Nvidia Geforce G210显卡上的加密速度分别提高5.28%,14.55%和12.53%,而在Nvidia Geforce GTX460显卡上的加密速度分别提高12.48%,15.40%和15.84%,且能更好地对SSL数据进行加密。

Optimization of AES Parallel Alogorithm Based on CUDA

In order to enhance the efficiency of Advanced Encryption Standard（AES）and make use of general computing ability of Graphics Processing Unit（GPU）,all the three versions of GPU parallel AES,namely128 bit version,192 bit version and256 bit version,are implemented on Compute Unified Device Architecture（CUDA）. Then,it proposes optimization alogorithms of parallel AES with3 versions. These alogorithms first consider threads amount in a block,shared memory size and total blocks,then use the experience data of optimal value of block size to guide AES alogorithm＇s optimal block on GPU. Experimental results show that compared with unoptimized parral AES,these alogorithms can obtain encryption mean speedup by5. 28%,14.55% and12.53% respectively on Nvidia Geforce G210 graphics card,while by12.48%,15.40% and15.84% on Nvidia Geforce GTX460 graphics card. In addition,these alogorithms are better at improving encrypting of Secure Socket Layer（SSL）.