Technology of parallelization for explicit high accuracy algorithms of computational fluid dynamics and aeroacoustics on unstructured grids

The paper is devoted to the technology to adapt the codes initially developed for sequential computations to the distributed?memory architecture of multiprocessor?computer systems. Under using high accuracy algorithms and unstructured grids, there is a serious difficulty in providing the high efficiency of parallelization even in the case of explicit methods. The paper describes in detail a technology of parallelization for explicit high accuracy algorithms of computational fluid dynamics on example of code NOISETTE for solving aeroacoustics problems on unstructured triangular grids. A special attention is paid to some specific techniques to keep the high parallelization efficiency peculiar to explicit methods. The given results on test problems and the problem on acoustic wave suppression in resonance-type liners demonstrate the parallelization quality and the capability of parallel NOISETTE version as a whole.