基于GPU的高性能彩虹表生成∗

摘要

Rainbow table, taking advantage of the compromise between time and space, is a powerful tool in terms of non-salted en-cryption algorithm. However, in order to generate rainbow table with certain success rate, the chains count and length should be large enough, thus resulting in long-time consumption of rainbow table pre-computing. Aiming at this, GPU is applied to accelerate the pre-computing of rainbow table by transferring the calculation of Hash and Reduct functions to GPU. Experiment shows that the pre-com-puting time is reduced to 1.89% ~ 2.70% of the original time.%利用时间空间折中思想的彩虹表是针对无盐口令加密算法的强大工具，然而为了生成有一定成功率的彩虹表，链长与链数必须足够大，因而在生成时需要消耗相当长的时间。针对这一问题，提出利用GPU加速彩虹表的生成，即将彩虹链的生成，主要是Hash以及Reduct函数的计算移植到GPU上进行。实验表明该方案能将彩虹表生成的时间缩短至原先的36.9~52.8分之一。