Parallel hypergraph partitioning for scientific computing.

Conference · Fri Jul 01 04:00:00 EDT 2005

OSTI ID:968387

Heaphy, Robert; Devine, Karen Dragon; Catalyurek, Umit ^[1]; Bisseling, Robert ^[2]; Hendrickson, Bruce Alan; Boman, Erik Gunnar

---

1. Ohio State University, Columbus
2. Utrecht University, The Netherlands

Graph partitioning is often used for load balancing in parallel computing, but it is known that hypergraph partitioning has several advantages. First, hypergraphs more accurately model communication volume, and second, they are more expressive and can better represent nonsymmetric problems. Hypergraph partitioning is particularly suited to parallel sparse matrix-vector multiplication, a common kernel in scientific computing. We present a parallel software package for hypergraph (and sparse matrix) partitioning developed at Sandia National Labs. The algorithm is a variation on multilevel partitioning. Our parallel implementation is novel in that it uses a two-dimensional data distribution among processors. We present empirical results that show our parallel implementation achieves good speedup on several large problems (up to 33 million nonzeros) with up to 64 processors on a Linux cluster.

• Research Organization: Sandia National Laboratories
• Sponsoring Organization: USDOE
• DOE Contract Number: AC04-94AL85000
• OSTI ID: 968387
• Report Number(s): SAND2005-4327C
• Country of Publication: United States
• Language: English

Related Subjects

97 MATHEMATICS AND COMPUTING
ALGORITHMS
COMPUTER CALCULATIONS
DIAGRAMS
PARALLEL PROCESSING