Parallel Computation of Two-dimensional Linearized Euler Equations

Perform numerical analysis of sound generation and propagation with accuracy and with lower computational time are some of the main challenges for Computational Aeroacoustics (CAA), which had experienced some major improvements in the last decade. In the present work it was performed numerical study of two-dimensional propagation of waves by solving LEE equations applied to a Gaussian pulse in pressure and density in an undisturbed mean flow with different Mach numbers. A 4-order 4-steps Runge-Kutta scheme was adopted for time integration. The spatial derivatives were calculated by a 4-order DRP and a 6-order compact finite difference schemes. Additionally, the problem was parallelized adopting MPI using 1-D and 2-D domain decompositions. A speed-up and efficiency analysis showed that the computation time is reduced using this technique, being more effective for the compact finite difference scheme. However, there was a decrease in efficiency probably due time spent in communication and synchronization.