Parallel Computing For The Data Of The Noise

It is helpful to develop the earthquake monitoring,the earthquake prediction, earthquake engineering, and the seismology through the research on the data of the noise. In order to improve the computing capability for the data of the noise, parallel computing is introduced in this paper. First of all, this paper describes the actuality in our country and the computing process of the data of the noise, and then it put forward and judge the reasonable task partitioning and mathematical model for the tasks allocation. Obviously it is a NP complete scheduling. On the premise of the load balance, it gave two different arithmetic for the task allocation. One is the genetic algorithm-the currency optimization algorithm. The paper proposes a partheno-genetic algorithm and gives the special mutation operator to speed up the convergence.The other is the greedy algorithm.Through the research on the graph theory, it gives a greedy tactic which can gets the stable result.But they have diffient characteristics. The result of the greedy algorithm is stable and it can ensure data balance for every node.Though the partheno-genetic algorithm can make the total data less transformation than the greedy algorithm, it can not ensure the data balance.So the nodes must be chosen to adopt different practical problems.After taking task allocation in the node for the different threads, it propose a appropriate algorithm to determine the number of the thread which can influence the efficiency of the parallel computing in our experiment in the paper. According to analysing the capability of the parallel computing for the data of the noise, it is easy to find the extendibility of the parallel computing.