"Comparison of data assimilation methods based on the classical, ensemble and local Kalman filter by the example of the advection equation and Lorenz system"
Rostilov D.A., Kaurkin M.N., and Ibraev R.A.

When parallelizing the solution of three-dimensional boundary value problems, especially in domains with complex geometry, the сomputational technologies and data structureы are important. The amount of stored information and the computational time depend on them. In this paper we propose the technologies for parallelizing the method of decomposition of the computational domain into subdomains conjugated without overlapping on a quasistructured grid. Parallel grid data structures oriented mainly to work with structured data arrays are developed. An illustrative example clarifying the fundamentals of the proposed approach is discussed.

Keywords: boundary value problems, quasistructured grids, parallelization technologies, data structures, structured arrays.

  • Rostilov D.A. – Lomonosov Moscow State University, Faculty of Computational Mathematics and Cybernetics; Leninskie Gory, Moscow, 119992, Russia; Student, e-mail: danilrostilov@gmail.com
  • Kaurkin M.N. – Shirshov Institute of Oceanology, Russian Academy of Sciences; Nakhimovskiy prospect 36, Moscow, 117218, Russia; Ph.D., Scientist, e-mail: kaurkinmn@gmail.com
  • Ibraev R.A. – Marchuk Institute of Numerical Mathematics, Russian Academy of Sciences; ulitsa Gubkina 8, Moscow, 119333, Russia; Dr. Sci., Professor, Corresponding Member of Russian Academy of Sciences, Principal Scientist, e-mail: ibrayev@mail.ru