| In the past two decades to improve the performance of single-core processor architecturerelies mainly on improving and increasing the frequency of cache to achieve. But because ofthe physical materials, manufacturing processes and heat and other factors, the performance ofthe processor’s difficult to rely on this method has been further improved. At the same time,GPU as a new type of multi-core processors have gradually developed and lead the forefront ofhigh-performance floating-point arithmetic.Development of GPU computing is also to promote the progress of cryptography.Especially in recent years the rapid development of the Internet, a lot of important datatransmitted over the network and the need to save, which contains many important areas relatedto economic, political and other people’s daily lives and privacy of information. On the otherhand, the popularity of the network but also to show rapid growth in cyber crime trends,seriously affected the user’s Internet security. For the security of the encryption algorithm, thekey itself is whether brute force for a limited time is a key measure of the effectiveness of theencryption algorithm. GPU parallel computing capabilities powerful safety study provides aplatform encryption algorithm. This paper realized the password recovery system has sufficientpractical significance in network security and cyber crime forensics.This paper will design and implement a CUDA-based RAR file password recovery systems.The system not only can greatly improve the speed of recovery RAR file password, and can becustomized and extended with different specifications depending on the applicationrequirements. These characteristics make it very promising in the field of computer forensics.This article will discuss in detail the system of GPU resources management, task allocationstrategy, RAR compression format as well as specific file decryption process, and optimizationtechniques based on optimization theory and GPU-based parallel programs. |
PDF Full Text Request