Title :
Challenges of Scaling Algebraic Multigrid Across Modern Multicore Architectures
Author :
Baker, Allison H. ; Gamblin, Todd ; Schulz, Martin ; Yang, Ulrike Meier
Author_Institution :
Center for Appl. Sci. Comput., Lawrence Livermore Nat. Lab., Livermore, CA, USA
Abstract :
Algebraic multigrid (AMG) is a popular solver for large-scale scientific computing and an essential component of many simulation codes. AMG has shown to be extremely efficient on distributed-memory architectures. However, when executed on modern multicore architectures, we face new challenges that can significantly deteriorate AMG´s performance. We examine its performance and scalability on three disparate multicore architectures: a cluster with four AMD Opteron Quad-core processors per node (Hera), a Cray XT5 with two AMD Opteron Hex-core processors per node (Jaguar), and an IBM Blue Gene/P system with a single Quad-core processor (Intrepid). We discuss our experiences on these platforms and present results using both an MPI-only and a hybrid MPI/OpenMP model. We also discuss a set of techniques that helped to overcome the associated problems, including thread and process pinning and correct memory associations.
Keywords :
distributed memory systems; message passing; parallel architectures; AMD Opteron Hex-core processors; AMD Opteron Quad-core processors; Cray XT5; IBM Blue Gene/P system; algebraic multigrid; distributed-memory architecture; hybrid MPI/OpenMP model; large-scale scientific computing; modern multicore architectures; Deformable models; Interpolation; Laboratories; Multicore processing; Program processors; Three dimensional displays;
Conference_Titel :
Parallel & Distributed Processing Symposium (IPDPS), 2011 IEEE International
Conference_Location :
Anchorage, AK
Print_ISBN :
978-1-61284-372-8
Electronic_ISBN :
1530-2075
DOI :
10.1109/IPDPS.2011.35