Mixed formulations and iterative solver for the parallelisation of a forging simulation software by mesh partitioning - Mines Paris Accéder directement au contenu
Communication Dans Un Congrès Année : 1999

Mixed formulations and iterative solver for the parallelisation of a forging simulation software by mesh partitioning

Résumé

A parallel implementation of the Preconditioned Conjugate Residual (PCR) method is described. The method is used to solve the discretised generalised Stokes problem derived from the simulation of complex 2D axisymetrical hot metal forming processes. We numerically show that the use of a simple Block Diagonal (BDS) preconditioner results in a high level of parallelism. However in 2D, the iterative solver with BDS is much slower than the direct one. Hence, better sequential and parallel preconditioners, like incomplete factorisations ILU, are studied. It results in an almost twice faster parallel code, but with a loss in scalability due to the increase of communications between processors.

Domaines

Matériaux
Fichier non déposé

Dates et versions

hal-00538294 , version 1 (22-11-2010)

Identifiants

Citer

Etienne Perchat, Lionel Fourment, Thierry Coupez. Mixed formulations and iterative solver for the parallelisation of a forging simulation software by mesh partitioning. 3rd Euro-Conference on Parallel and Distributed Computing for Computational Mechanics, Mar 1999, Weimar, Germany. p.67-72, ⟨10.4203/ccp.57.3.1⟩. ⟨hal-00538294⟩
138 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More