Abstract

Numerical solution of aerothermodynamics problems requires a lot of processor time, and for this reason, the numerical codes for such simulations must be efficiently parallelized. This paper presents a software platform based on a technological approach that greatly simplifies the parallelization of problems with unstructured grids. The paper formulates the principle of a unified mathematical address space of the problem for all used cluster nodes. The formalization of the presented technological approach and the implementation of its software basis in the form of data structures, exchange routines and work rules form a software platform on the basis of which parallel computational programs can be built. This approach is implemented and tested for the problem of modeling the characteristics of aircrafts, but it can be applied to other problems using unstructured grids with one-dimensional cell numbering. In this case, the physical and mathematical specifics of the problem are not important. Parallelization of the code with the new approach is carried out with minimal effort, without changing the main parts of the program. As a result, a single computational code can be created for all regimes — sequential, multi-threaded, and cluster. The performance results obtained with the new code confirm the good scalability of the parallelization method.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call