Authors: J.C. Cajasaa, G. Houzeauxa and B. Eguzkitzaa
aBarcelona Supercomputing Center – Centro Nacional de Supercomputación
aaNEXUS II Building, c/ Jordi Girona 29, 08034 Barcelona (Spain)
Abstract: Domain Composition Methods are techniques to couple local solutions of physical problems, solved on local meshes, to obtain a global solution on the union of these meshes. This work consisted in implementing such techniques at the algebraic level, making the coupling independent of the physics to be considered. This approach enables us to solve multi-domain and multi-physics problems, using both single and multi-code approaches. Both explicit and implicit coupling were implemented, for surface and volume couplings. The implementation was carried out for distributed memory supercomputers, using MPI. Several physical examples demonstrate the reliability of the proposed implementation.
Authors: Yusuf Yılmaz,Can Özturan, Oğuz Tosun, Ali Haydar Özer, Seren Soner
Dept. of Computer Engineering, BogaziciUniversity, Istanbul, Turkey
Abstract: The main goal of this project is to develop a parallel tetrahedral mesh generator based on existing sequential mesh generation software. As sequential mesh generation software, the Netgen mesh generator was used due to its availability as LGPL open source software and its wide user base. Parallel mesh generation routines were developed using the MPI libraries and the C++ language. The parallel mesh generation algorithms developed proceed by decomposing the whole geometry into a number of sub-geometries sequentially on a master node at the beginning and then mesh each sub-geometry in parallel on multiple processors. Three methods were implemented. The first decomposes the CAD geometry and produces conforming surface sub-meshes that are sent to other processors for volume mesh generation. The second and third methods are refinement based methods that also make use of the CAD geometry information. Advantages and disadvantages of each method are discussed. Parallel repartitioning also need to be done in the first method. To facilitate distributed element movements in parallel, a migration algorithm that utilizes “owner updates” rule is developed. Timing results obtained on the Curie supercomputer are presented. In particular, results show that by using a refinement based method, one can generate over a billion element meshes in under a minute.
Authors: PavlaKabelikova, Ales Ronovsky, Vit Vondraka
Dept. of Applied Mathematics,VSB-Technical University of Ostrava, Tr. 17. listopadu 15, 708 00Ostrava, Czech Republic
Abstract: In this whitepaper, a new mesh multiplication package developed for Code_Saturne is described. The package implements parallel global refinement of hexahedral meshes for Code_Saturne to allow creating meshes with more than 1 billion cells. This enables running Code_Saturne’s extremely large CFD simulations on PRACE Tier-0 systems. The effectiveness of the implemented multiplication algorithm is demonstrated on practical examples, which were carried out on CURIE system at CEA.
Authors: G.Houzeaux,R.delaCruz, M.V ́azquez
Barcelona Supercomputing Center,Edificio NEXUS I, Gran Capit ́an 2-4, 08034 Barcelona, Spain
Abstract: The objective of the present project is to implement a parallel uniform mesh multiplication in a HPC code developed atBarcelona Supercomputing Center named Alya. The mesh multiplication consists in subdividing recursively the mesh inparallel in order to obtain a reﬁned mesh from an initial ”coarse” mesh. This mesh multiplication should be implementedeﬃciently to be used the ﬂy, thus avoiding treating huge geometry ﬁles. In addition, the post-process can be carried outin a straightfoward way at any level of the mesh multiplication.
Rila Solutions EAD, Acad. G. Bonchevstr., bl. 27, Sofia 1113, Bulgaria
Abstract: The SPECFEM3D package simulates seismic wave propagation using a parallel implementation of a variation of the Galerkin procedure called Spectral Element Method (SEM). The advantage of this method is that it produces a diagonal mass matrix which allows for quick explicit solving of the seismic wave equations. The drawback is that using an explicit method the solver is not stable unless the Courant-Friedrichs-Lewy (CFL) condition for convergence holds, imposing a certain inequality relation between the time step and spatial size of the spectral elements. The internal mesher provided with the package uses a partitioning approach that directly links the number of parallel solver tasks with the spatial size of the elements. For any given mesh, increasing the number of parallel processes leads to decreasing the element size, and via the CFL condition to decreasing the time step. This directly compromises the scalability of the solver by requiring more time steps for the same amount of work with no gain in performance. The goal of this project is to improve the scalability of the package by reconsidering the partitioning approach of the internal mesher. We analyze the existing approach and propose a new one, based on using the SCOTCH partitioning library, already integrated in the workflow for solving the problem on externally generated mesh. Our results indicate that the modified partitioning scheme leads to substantial performance benefits for regional modeling with the internal mesher in petascale environments. We validate the results by running the examples included in the package and observe that the solver produces identical synthetic seismograms when run with meshes generated by the original and modified internal mesher.
These whitepapers have been prepared by the PRACE Implementation Phase Projects and in accordance with the Consortium Agreements and Grant Agreements n° RI-261557, n°RI-283493, or n°RI-312763.
They solely reflect the opinion of the parties to such agreements on a collective basis in the context of the PRACE Implementation Phase Projects and to the extent foreseen in such agreements. Please note that even though all participants to the PRACE IP Projects are members of PRACE AISBL, these whitepapers have not been approved by the Council of PRACE AISBL and therefore do not emanate from it nor should be considered to reflect PRACE AISBL’s individual opinion.
© 2014 PRACE Consortium Partners. All rights reserved. This document is a project document of a PRACE Implementation Phase project. All contents are reserved by default and may not be disclosed to third parties without the written consent of the PRACE partners, except as mandated by the European Commission contracts RI-261557, RI-283493, or RI-312763 for reviewing and dissemination purposes.
All trademarks and other rights on third party products mentioned in the document are acknowledged as own by the respective holders.