Evaluation of the characteristics of a series-parallel statistical process modeling

Abstract

The article is devoted to evaluation of speed of transformation to be performed while modeling the statistical characteristics of system based on GPGPU complex of  sequential -parallel data processing. A generalized task of coordinating a sequential-parallel process of statistical modeling for sequential steps of transformation of the initial information. The processes of coordination analyzed using Gantt charts. Terms of performance of modeling operations based on the operator conversion method of distribution laws of probabilities, that uses integral operators. In tbe article were synthesized and tested for adequacy generalized model of coordination of parallel processes and considered its application to image processing. The developed approach allows you to create criteria for the selection of the optimum size of the data blocks in parallel processing and to reduce the simulation time.