基于异构平台的代数密码分析并行加速技术研究
发布时间：2018-07-31 16:46
【摘要】：作为网络信息安全的重要支撑技术,密码分析在信息时代的作用日益凸显,它的发展受到广泛关注。代数密码分析是当前发展最快、应用前景最好的主流密码分析技术之一,已经取得了许多突破性的成果。然而随着研究的深入,代数密码分析技术的发展面临密码代数系统求解的高计算复杂性的挑战。针对这一问题,本文结合并行计算技术对代数密码分析的加速技术进行研究,主要取得了以下研究成果:1.为了有效实现密码代数系统求解在异构平台上的并行加速,本文在深入研究一般并行计算模型和密码代数系统求解的特点的基础上,提出了密码代数系统求解加速的异构并行计算模型,为密码代数系统求解相关算法在异构平台上的并行化实现建立了基本框架。2.为了有效利用协处理器实现对密码代数系统求解运算的加速,本文针对现有求解算法,设计实现了基于二元域矩阵运算的协处理器加速技术。3.为了提高基协处理器对密码代数系统求解的加速效果,实现对协处理器的计算资源利用率最大化,本文基于Roofline性能评价模型提出了针对协处理器的性能评估和优化方法。4.本文基于CPU/GPU异构平台,完成了对吴特征列算法完成了并行化实现,并将它应用到具体的代数密码分析中进行了相关测试。通过测试分析,验证了本文研究的技术对密码代数系统求解加速的有效性。
[Abstract]:As an important supporting technology of network information security, cryptographic analysis plays an increasingly important role in the information age. Algebraic cryptography analysis is one of the most popular cryptographic analysis techniques with the fastest development and the best application prospect. Many breakthrough results have been achieved. However, with the development of algebraic cryptographic analysis technology, the development of algebraic cryptosystem is facing the challenge of high computational complexity. Aiming at this problem, this paper combines parallel computing technology to study the accelerated technology of algebraic cryptography analysis, and obtains the following research results: 1. In order to realize the parallel acceleration of cryptographic algebraic system solving on heterogeneous platforms effectively, this paper studies the general parallel computing model and the characteristics of cryptographic algebraic system solving. An accelerated heterogeneous parallel computing model for cryptographic algebraic systems is proposed. The basic framework of parallelization of cryptographic algebraic system solving algorithms on heterogeneous platforms is established. In order to effectively use coprocessor to accelerate the solution of cryptographic algebraic system, this paper designs and implements the coprocessor acceleration technology based on binary matrix operation. In order to improve the acceleration effect of base coprocessor to solve cryptographic algebra system and realize the maximum utilization of computing resources for coprocessor, a performance evaluation and optimization method for coprocessor based on Roofline performance evaluation model is proposed in this paper. Based on the heterogeneous platform of CPU/GPU, the parallel implementation of Wu's characteristic sequence algorithm is completed, and it is applied to the analysis of algebraic cryptography. The effectiveness of the proposed technique for accelerating the solution of cryptographic algebraic systems is verified by test and analysis.
【学位授予单位】：国防科学技术大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN918.1
本文编号：2156194
[Abstract]:As an important supporting technology of network information security, cryptographic analysis plays an increasingly important role in the information age. Algebraic cryptography analysis is one of the most popular cryptographic analysis techniques with the fastest development and the best application prospect. Many breakthrough results have been achieved. However, with the development of algebraic cryptographic analysis technology, the development of algebraic cryptosystem is facing the challenge of high computational complexity. Aiming at this problem, this paper combines parallel computing technology to study the accelerated technology of algebraic cryptography analysis, and obtains the following research results: 1. In order to realize the parallel acceleration of cryptographic algebraic system solving on heterogeneous platforms effectively, this paper studies the general parallel computing model and the characteristics of cryptographic algebraic system solving. An accelerated heterogeneous parallel computing model for cryptographic algebraic systems is proposed. The basic framework of parallelization of cryptographic algebraic system solving algorithms on heterogeneous platforms is established. In order to effectively use coprocessor to accelerate the solution of cryptographic algebraic system, this paper designs and implements the coprocessor acceleration technology based on binary matrix operation. In order to improve the acceleration effect of base coprocessor to solve cryptographic algebra system and realize the maximum utilization of computing resources for coprocessor, a performance evaluation and optimization method for coprocessor based on Roofline performance evaluation model is proposed in this paper. Based on the heterogeneous platform of CPU/GPU, the parallel implementation of Wu's characteristic sequence algorithm is completed, and it is applied to the analysis of algebraic cryptography. The effectiveness of the proposed technique for accelerating the solution of cryptographic algebraic systems is verified by test and analysis.
【学位授予单位】：国防科学技术大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN918.1
【共引文献】
相关期刊论文 前4条
1 施伟;郑兆鹏;班琪;;基于改进Hough算法的地铁屏蔽门异物检测技术[J];城市轨道交通研究;2016年11期
2 郑兆鹏;张祥福;;基于改进Hough算法的地铁屏蔽门与列车间异物自动检测技术[J];内蒙古科技与经济;2015年14期
3 张国英;宋科科;赵鹏;程金龙;;一种采用容错宽度Hough变换的路网优化方法[J];测绘科学技术学报;2014年03期
4 陈洪波,王强,徐晓蓉;用于线段特征提取的改进Hough变换[J];计算机工程与应用;2004年21期
，本文编号：2156194
本文链接：https://www.wllwen.com/kejilunwen/wltx/2156194.html
