Project summary

Supercomputer architecture is moving quickly to multi-core and many-core architectures. An additional trend is the increasing use of special purpose accelerators, e.g. in form of graphics cards, the Cell processor, or reconfigurable hardware. This has the potential to deliver unprecedented performance at lower cost and reduced power consumption. However, this trend opens many unanswered questions on how these devices can be used effctively in real life supercomputing applications, since these accelerators are only effective on certain algorithmic structures that exhibit a high degree of regularity and they are ineffcient or very diffcult to program for others. At the LSS we have developed (partly with KONWIHR support) application software whose basic architecture will permit the use of heterogeneous architectures. For this project, we will focus on the waLBerla framework for simulating complex ows in a wide range of applications. The software waLBerla is designed such that it distinguishes between modules that require complex algorithmic logic and regular structures that can exploit a very high degree of parallelism. While the latter parts are very well suited for a potential use of multi-core parallelism and accelerator technologies, the algorithmically more complex parts can e.g. be run on the conventional parts of a heterogeneous system. We therefore propose to extend waLBerla to exploit current and future heterogeneous multi-core architectures.

KONWIHR funding

• KONWIHR funding of waLBerla-MC: 9/2008 – 8/2013

Contact:

• Prof. Dr. Ulrich Rüde, Lehrstuhl für Systemsimulation, Uni-Erlangen

Project staff:

• Dipl.-Inf. Markus Stürmer, Lehrstuhl für Systemsimulation, Uni-Erlangen
• Jan Götz, M.Sc., Lehrstuhl für Systemsimulation, Uni-Erlangen
• Stefan Donath, M.Sc., Lehrstuhl für Systemsimulation, Uni-Erlangen

Publications and presentations

• J. Götz, K. Iglberger, C. Feichtinger, S. Donath, U. Rüde: Coupling Multibody Dynamics and ComputationalFluid Dynamics on 8 192 Processor Cores, Parallel Computing, Elsevier, 36 (2010) 142-151.
• Jan Götz, Klaus Iglberger, Markus Stürmer, Ulrich Rüde: Direct Numerical Simulation of Particulate Flows on 294 912 Processor Cores, Accepted at Supercomputing Conference 2010, New Orleans, LA, (2010).
• S. Donath, C. Feichtinger, T. Pohl, J. Götz, U. Rüde: Localized Parallel Algorithm for Bubble Coalescence in Free Surface Lattice Boltzmann Method, Lecture Notes in Scientific Computing, Euro-Par 2009 (Ed: H. Sips, D. Epema, H.-X. Lin), Springer (Berlin, Heidelberg), 5704 (2009) 735-746.
• S. Donath, C. Feichtinger, T. Pohl, J. Götz, U. Rüde: A Parallel Free Surface Lattice Boltzmann Method for Large-Scale Applications, Parallel CFD 2009, 21st International Conference on Parallel Computational Fluid Dynamics (Ed: R. Biswas), (2009), pp. 198-202.
• M. Stürmer, J. Götz, G. Richter, A. Dörfler, U. Rüde: Fluid flow simulation on the Cell Broadband Engine using the lattice Boltzmann method, Computers , 58:5 (2009) 1062 – 1070.
• J. Götz, C. Feichtinger, K. Iglberger, S. Donath, U. Rüde: Large scale simulation of fluid structure interaction using Lattice Boltzmann methods and the „physics engine“, Proceedings of the 14th Biennial Computational Techniques and Applications Conference, CTAC-2008 (Ed: G.N. Mercer, A.j. Roberts), ANZIAM Journal (Canberra, Australia), (2008), pp. C166-C188.