Abstract
This chapter discusses the direct numerical simulation method using PC clusters. PC clusters with conventional 100 Mbits/s networks provide a high computing power to cost ratio compared with conventional parallel computers. However, their low bandwidth and high latency are serious obstacles that prevent their efficient use for many conventional parallel CFD (computational fluid dynamics) algorithms. A code for the direct numerical simulation (DNS) of turbulent flows, that provides fairly good scalability up to about 36 processors has been developed. It is based on a finite control-volume formulation and uses a spectro-consistent numerical scheme. The main bottleneck and the Poisson equation are solved with a Direct Schur-Fourier Decomposition (DSFD). This algorithm only needs one all-to-all communication episode to solve the Poisson equation almost to machine-accuracy and therefore is especially useful on high latency computers. The statistics of a turbulent natural convection flow in a tall cavity of aspect ratio 4, Pr=0.71 and Raz = 6.4 x 108 are presented and compared with the previous results available from the literature, that assumes a two-dimensional behavior.
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 callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
