SC Conference - Activity Details

A Scalable Domain Decomposition Method for Ultra-parallel Arterial Flow Simulations

Leopold Grinberg  (Brown University)
John Cazes  (Texas Advanced Computing Center)
George Karniadakis  (Brown University)
Posters Session
Tuesday,  05:15PM - 07:00PM
Room Rotunda Lobby
Ultra-parallel flow simulations on hundreds of thousands of processors require new multi-level domain decomposition methods. We present a new two-level method that couples discontinuous and continuous Galerkin formulations. At the coarse level the domain is subdivided into several big patches and within each patch a spectral element discretization is employed. New interface conditions for the Navier-Stokes equations are developed to connect the patches, relaxing the C^0 continuity and minimizing data transfer at the patch interface. A Multilevel Communicating Interface (MCI) has been developed to enhance the parallel efficiency. The potentially great advantage of the new approach is in simulating very large scale problems. It addresses limited available memory per processor and tight communication between thousands of processors. We present the implementation of the MCI in our research code Nektar and results of 3D flow simulations of benchmark problems and of arterial flows to evaluate the performance of the new method.
   IEEE Computer Society  /  ACM     2 0   Y E A R S   -   U N L E A S H I N G   T H E   P O W E R   O F   H P C