Solving 3D incompressible Navier-Stokes equations on hybrid CPU/GPU systems

Yushan Wang (1) , Marc Baboulin (2, 1) , Karl Rupp (3) , Olivier Le Maitre (4) , Yann Fraigneau (4)

1 LRI - Laboratoire de Recherche en Informatique
2 POSTALE - Performance Optimization by Software Transformation and Algorithms & Librairies Enhancement
3 TU Wien - Vienna University of Technology = Technische Universität Wien
4 LIMSI - Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur

This paper describes a hybrid multicore/GPU solver for the incompressible Navier-Stokes equations with constant coefficients, discretized by the finite difference method. By applying the prediction-projection method, the Navier-Stokes equations are transformed into a combination of Helmholtz- like and Poisson equations for which we describe efficient solvers. As an extension of a previous paper, this paper proposes a new implementation that takes advantage of GPU accelerators. We present numerical experiments on a current hybrid machine.

Mots clés: Graphics Processing Unit (GPU); Navier-Stokes equations; prediction-projection method; parallel computing

Domaines: Analyse numérique [cs.NA]

Yushan Wang, Marc Baboulin, Karl Rupp, Olivier Le Maitre, Yann Fraigneau. Solving 3D incompressible Navier-Stokes equations on hybrid CPU/GPU systems. High Performance Computing Symposium (HPC'14), Spring Simulation Multiconference, Apr 2014, Tampa, Florida, USA, United States. ⟨hal-01205305⟩