TOWARDS A FAST PARALLEL SPARSE MATRIX-VECTOR MULTIPLICATION
Roman Geus, Stefan Röllin · 2000
The sparse matrix-vector product is an important computational kernel that runs ineffectively on many computers with super-scalar RISC processors. In this paper we analyse the performance of the sparse matrix-vector product with symmetric matrices originating from the FEM and describe techniques that lead to a fast implementation. It is shown how these optimisations can be incorporated into an efficient parallel implementation using messagepassing. We conduct numerical experiments on many different machines and show that our optimisations speed up the sparse matrix-vector multiplication substantially. Key words: sparse matrices, matrix-vector multiplication, source code optimisation, parallel linear algebra 1 Performance analysis of the sparse matrix-vector product In this paper we focus on large symmetric sparse matrices, that do not fit into the memory cache. While our matrices are stored in symmetric sparse skyline format (SSS), our ideas can be applied to general sparse matrice...