Research And Implementation Of GPU Hash Algorithm Computing Library And WLAN Module

After computing of hash function,a plaintext can be transformed to be a ciphertext.Theoretically,in case of kowning the ciphertext,we can find the corresponding plaintext by using the all of the plaintext to be computed with hash function.If the hash function is complex and the plaintext is long and varity,only to depend on traditional computing power of CPU can hardly achieve it. The emergence of GPU greatly improve the scale and the speed of parallel computing,and it provide a new research direction of plaintext searching task.This paper mainly discusses the implementation of the hash function computing library on GPU,on this basis, the large scale parallel computing of WLAN hash function based on GPU is realized.The library contains two aspects:the plaintext fundation library and the based hash function library.Under the integral plaintext space,usually represented by regular expression,finding out all the combination of plaintext is the base of plaintext searching.The plaintext library also include two aspects.the first part is funcion library of the constrcut of plaintext’s attribute at the host-side,the second part is the public function library of reading plaintext at the GPU device-side.To achieve the combination of plaintext,this paper provides a new method,it made operating of plaintext reading more simply and more effectivly.The based hash function library mainly packaged the common and based hash function to be a single module,so other complex hash function can call them easily.The library can reduce the development cycle,also it can develop the efficiency of program.Under the studing of WLAN hash function,this paper refines 3 based hash function,and puts them into the based hash function library.On the basis of these works,the author has accomplished the large scale parallel computing of WEP and WPA hash function on GPU.According to the features of hash function and GPU architecture,the paper proposes a series of optimization,and get a good result.At last,the paper illustrates the rationality and efficiency of these designs by the experimental data.