Research On Optimization Technology Of Office Password Recovery Based On Domestic Self-controllable Platform

At present,the storage,transmission and use of data generally adopt the form of encryption,the purpose is to improve the security of the data.At the same time,criminals will also use various encryption methods to hide criminal information,which brings new challenges to the work of the security department.Due to the continuous improvement of encryption algorithms,the number of digits in the key is increasing,and the form of the key combination is becoming more and more complex,resulting in a gradual increase in the calculation scale of password recovery.Existing CPU platforms cannot meet the computing requirements for password recovery,and GPU platforms have memory access problems,which limit the performance of password recovery.In view of this,there is an urgent need for a high-performance password recovery system to realize the dynamic aggregation of large-scale computing resources,and then to realize the efficient analysis and recovery of encrypted data.This thesis proposes a research on the optimization technology of Office password recovery,with the purpose of improving the performance of password recovery.The specific research content and contribution points are as follows:1.This thesis takes Office password recovery algorithm as the research object,analyzes and designs a password recovery system based on Sunway multi-core platform.At the same time,it carried out indepth parallelism analysis and correlation processing of its algorithm,tested and counted the hot functions of the password recovery algorithm,and found out the bottleneck that affects its performance.2.According to the characteristics of the domestic Sunway multi-core platform,six optimization strategies are proposed to improve and optimize the password recovery system.This optimization strategy makes good use of the secondary core computing performance and local storage advantages of Sunway multi-core platform.In order to solve the problem of long execution time of loop nested functions,64 slave cores of SW26010 are used to accelerate calculation.In order to solve the problem of large memory access delay in the function,it is proposed to use Local Direct Memory(LDM)to reduce the memory access delay.In order to solve the problem of low data transmission efficiency between functions,a batch transmission technology using DMA(Direct Memory Access)is proposed.In order to solve the problem of limited local storage space of 64 KB,the principle of circular block using tile statement is proposed.In order to solve the problem of reduced core computing performance or increased communication overhead,it is proposed to use a double buffering mechanism.In order to solve the 64 computational problems of intensive data from the core,vectorization using data-level parallelism is proposed.3.This thesis verifies the correctness of the optimization strategy,the effectiveness,scalability and controllability of the password recovery system,and the superiority of the Sunway multi-core platform.The experimental results show that the optimized system has the characteristics of strong parallelism,faster recovery speed,and larger throughput.Compared with the open-source software Hashcat,the password recovery performance has increased by 2.06 times,and compared with the commonly used password recovery tools AOPR,EDPR and Accent Software,it has increased by 97.21,19.96 and 15.31 times respectively.This optimization strategy makes good use of the advantages of the domestic Sunway multicore platform,improves the efficiency of password recovery,and alleviates actual application requirements.