engleski

Solution of applied statistical problems based on multiprocessor systems

The paper deals with multiprocessor modeling technologies for Monte Carlo problems. Calculation schemes are presented that provide an increase in the productivity and speed of calculations. The paper proposes a modified parallel computing algorithm based on the Monte Carlo method. Here, each calculator has its own random number generator. In this case, intermediate calculations are performed independently on different, separately taken blades of the cluster, “calculators”. Obviously, such a solution allows you to speed up the computation process. It is shown that parallel algorithms of the Monte Carlo method are stable to any input data and have the maximum parallel form and, therefore, the minimum possible implementation time on parallel computin.

multiprocessor modeling technologies ; Monte Carlo problems ; applied statistical problems

nije evidentirano