The application used is based on DNS (Direct Numerical Simulation)-, LES- and RANS (Reynolds–Averaged Navier–Stoke) unified C/C++ code developed at the MSU Research Computing Center (RCC MSU) and the Marchuk Institute of Numerical Mathematics. The code is designed for studies of geophysical turbulent flows and was evaluated in a number of the atmospheric boundary layer experiments. The model utilizes hybrid MPI/OpenMP/CUDA approach for computations on parallel architectures. In this paper we consider MPI-only implementation. The code supports spatial decomposition of the computational grid among MPI-processes with common optimizations for improving scaling on HPC systems. Numerical methods utilized in the application are based on conservative in momentum and energy second-order finite-difference approximations explicit third-order Adams–Bashforth time integration scheme. The new version of the RCC MSU – INM RAS LES model contains a rich set of additional parameterizations of physical processes in the atmospheric boundary layer, such as: modules for calculating the radiation transfer in the short- and long-wave range, cloud microphysics, as well as an interactive model of the soil layer.

• Developing since 2009
• Filtered Navier-Stockes for viscous incompressible fluid
• Boussinesq approximation for stratified fluid
• DNS-, LES-, RANS- turbulence modelling support
• High performance computations: MPI/OpenMP/CUDA
• Fully suited for comprehensive atmospheric turbulence modelling
• Mixed dynamic subfilter/subgrid closure for LES
• Cloud microphysics
• Radiative transfer module
• Rich parameterization package for atmosphere / surface interaction
• Explicit building treatment
• Atmospheric chemistry / pollution transport in both Eularian and Lagrangian frameworks