San Diego, CA — We have put together a benchmark suite for iterative methods on sparse data called SparseBench. SparseBench is a benchmark suite of iterative methods on sparse data. Sparse matrices, such as derived from PDEs, form an important problem area in numerical analysis. Unlike in the case of dense matrices, handling them does not entail much reuse of data. Thus, algorithms for sparse matrices will be more bound by memory-speed than by processor-speed.
This benchmark uses common iterative methods, preconditioners, and storage schemes to evaluate machine performance on typical sparse operations. The benchmark components are: Conjugate Gradient and GMRES iterative methods, Jacobi and ILU preconditioners, diagonal storage and compressed row storage matrices.
The benchmark and instructions for running the suite can be found at http://www.netlib.org/benchmark/sparsebench/
Jack Dongarra, Victor Eijkhout Computer Science Department University of Tennessee Knoxville, TN 37996-1301, USA and Henk van der Vorst Universiteit Utrecht Utrecht, the Netherlands
For comments and questions, mail to [email protected]