Fine-grained NUMA-aware parallel scheme for anisotropic mesh adaptation
Rakotoarivelo, Hoby, Franck Ledoux, Franck Pommereau
Posters, 25th International Meshing Roundtable, Sandia National Laboratories, September 26-30 2016
25th International Meshing Roundtable
Washington DC, U.S.A.
September 26-30, 2016
Hoby Rakotoarivelo, CEA, FR, firstname.lastname@example.org
Franck Ledoux, CEA, FR, email@example.com
Franck Pommereau, IBISC, FR, firstname.lastname@example.org
In this work, we provide a fine-grained parallel scheme for anisotropic mesh adaptation on non-uniform memory access (NUMA) multicore architectures. Data dependencies are expressed by a task graph for each remeshing stage. Concurrency is extracted through speculative parallel graph coloring.
To ensure data consistency, tasks are structured into bulk-synchronous steps using the queuing shared-memory (QSM) bridging model. Thus interleaved mesh requests/updates patterns are avoided. To ensure performance portability, theoretical guarantees on asymptotical execution time and load imbalance are given. Furthermore, the impact of degree distribution and memory accesses penalties on scalability is highlighted.
Download Full Paper (PDF)
Contact author(s) or publisher for availability and copyright information on above referenced article