On the Performance of an Algebraic MultigridSolver on Multicore Clusters [electronic resource].
- Washington, D.C. : United States. Dept. of Energy, 2010.
Oak Ridge, Tenn. : Distributed by the Office of Scientific and Technical Information, U.S. Dept. of Energy.
- Physical Description:
- PDF-file: 16 pages; size: 1 Mbytes
- Additional Creators:
- Lawrence Berkeley National Laboratory, United States. Department of Energy, and United States. Department of Energy. Office of Scientific and Technical Information
- Restrictions on Access:
- Free-to-read Unrestricted online access
- Algebraic multigrid (AMG) solvers have proven to be extremely efficient on distributed-memory architectures. However, when executed on modern multicore cluster architectures, we face new challenges that can significantly harm AMG's performance. We discuss our experiences on such an architecture and present a set of techniques that help users to overcome the associated problems, including thread and process pinning and correct memory associations. We have implemented most of the techniques in a MultiCore SUPport library (MCSup), which helps to map OpenMP applications to multicore machines. We present results using both an MPI-only and a hybrid MPI/OpenMP model.
- Report Numbers:
- E 1.99:llnl-conf-429864
- Published through SciTech Connect.
Presented at: VECPAR 2010: High Performance Computing for COmputational Science, Berkeley, CA, United States, Jun 22 - Jun 25, 2010.
Schulz, M; Yang, U M; Baker, A H.
- Funding Information:
View MARC record | catkey: 14653831