The parallelization of a multigrid solver for the quasi–3D and 3D RANS equations was implemented on a distributed shared memory parallel–architectured computer by the insertion of compiler directives. Tests performed by varying the number of blocks are presented in terms of speed–up and percentage of achieved parallelism. Typical time–consuming industrial applications which can benefit of this approach are discussed.
Turbomachinery Computations on Parallel Computers Using a Multigrid Method / A. Arnone; P. Boncinelli; M. Marconcini. - STAMPA. - Lecture notes in computational science and engineering, vol. 14:(2000), pp. 41-51. (Intervento presentato al convegno Multigrid Methods VI Proceedings of the Sixth European Multigrid Conference tenutosi a Gent, Belgium nel September 27-30) [10.1007/978-3-642-58312-4_5].
Turbomachinery Computations on Parallel Computers Using a Multigrid Method
ARNONE, ANDREA;MARCONCINI, MICHELE
2000
Abstract
The parallelization of a multigrid solver for the quasi–3D and 3D RANS equations was implemented on a distributed shared memory parallel–architectured computer by the insertion of compiler directives. Tests performed by varying the number of blocks are presented in terms of speed–up and percentage of achieved parallelism. Typical time–consuming industrial applications which can benefit of this approach are discussed.
I documenti in FLORE sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
