Computer Simulation of Carbon Nanostructure Synthesis Applying Multiprogramming Technologies

Abstract

Carbonnanostructures attract an increasing attention among researchers in various spheres due to their unique physical and chemical properties. The article considers the application of numerical methods to find solutions of different nanostructure formation models. Modelling problem of carbon nanostructure synthesis is decomposed into six subproblems and time taken to their solution has been analyzed. Anassumption was made that the most effective optimization is that of electromagnetic field parameter analysis algorithm as this subproblem requires the least time to solve. It was shown that modern mathematical models applied for carbon nanostructure synthesis description are characterized by high dimensions causing difficulties in calculation of model parameters. The results with low computational efficiency are more evident in case of parallel approach as computation flows have to be stopped from time to time to determine the stopping condition meeting. Theimplementationofdistributedparallelstructuresis assumed to provide parallelizing of computational processes within distributed computational environment. To solve the present problem method of multiprogramming has been applied. Thealgorithmofparallelcomputationand its implementation in the form of set of scenarios written on Python finding plasma cube potential numerical value of the prescribeddimensions via certain iterations was suggested. The suggested approach of computation optimizing can be implemented in case of large-particle method that combinesserial application of Euler and Lagrange approaches providing algorithm parallelizing. Theassumedparallelizingtechniquecanalsobeapplied bothto move from computational mesh nodes to the interaction between large computational particles and to solve Harlowparticle-in-cell method.