Algorithmic Of SHA256 Optimization Based On Network-on-Chip

With the development of cryptography theory and distributed network,block chain technology has entered a period of rapid development.Digital money,a typical application of block chains,has sprung up like bamboo shoots after a spring rain.Block chain technology is a decentralized distributed network structure based on cryptography.In the whole block chain,the core is the calculation of hash value.After calculating the effective hash,new blocks can be created.In block chains,SHA256 algorithm is the most widely used algorithm to calculate hash value.The internal execution process of SHA256 algorithm is the whole cycle iteration.The result of this cycle iteration is the input data of the result of the next iteration.In this way,a serious dependence on the whole algorithm is formed.Moreover,the algorithm is mainly composed of logical operation instructions,and there is a lot of instruction redundancy.The main content of this paper is to optimize the instruction redundancy and serial execution of SHA256 algorithm.The main contents of this paper are as follows:1.Instruction optimization of SHA256 algorithmIn the chapter of instruction optimization,we focus on the separation of constants and variables in the whole cycle of SHA256 algorithm to reduce the waste of computing power caused by invalid transmission of constants.Aiming at the functions of two large cycles in the algorithm,we design and implement the special instructions of SHA256 algorithm,which greatly reduces the number of instructions in the whole process of algorithm execution and speeds up the algorithm.Finally,the judgment of effective hash needs to pass two judgment conditions,but the possibility of satisfying the two conditions is very different.So the two judgment conditions are separated,and the special judgment instruction of SHA256 algorithm is designed and implemented.2.Implementation of SHA256 Pipeline ComputingSHA256 algorithm is a multi-iteration process in the actual implementation process.The data used in each iteration is the result of the last iteration.There is a serious data dependence in the two iterations.Therefore,the whole execution process of SHA256 algorithm is serial execution,and parallel computation cannot be carried out.However,in the actual application process,SHA256 needs to be executed many times,so pipeline computing can be applied in the implementation process of SHA256 algorithm.In addition,the SHA256 algorithm is decomposed into data flow graph for calculation in the experiment.Based on this premise,the ACK mechanism is designed and implemented.Through the ACK mechanism,the parallel computation of some nodes in the data flow graph is realized,and the execution efficiency of the whole algorithm is improved.3.Application of multitask dislocation mapping methodAfter Instruction Optimization and pipeline computing are completed,the algorithm is mapped to the network on chip,and the actual component utilization rate is still very low.The essence of the problem is that SHA256 algorithm has dependence and PE idle rate is high.Aiming at the high spare rate of components in the network on chip,an optimization scheme of multi-task dislocation mapping is proposed,which maps multi-task in the network on chip and performs dislocation mapping.In this way,the idle computing power is utilized to improve the utilization rate of components in the network on chip.