SMOOTH: Sparse Matrix Object-Oriented Ordering Methods. The ordering of sparse matrices is an important part of solving sparse linear systems of equations. The SMOOTH software package provides this functionality with a collection of software objects in two ways. First, some important concepts from the underlying graph theory, (e.g., graphs, bipartite graphs, permutations, trees, elimination trees, heaps, and others), are encapsulated as abstract data types. Second, some important algorithms are also formulated as objects, (e.g., minimum degree, Kernighan-Lin, and some new graph partitioning methods). This collection of objects forms a toolkit for both the researcher and end user, to be used to order sparse matrices as well as perform related tasks. The ordering algorithms in this package have shown themselves to be competitive with state-of-the-art software. However, we hope that the objects in the SMOOTH package, both data and algorithms, will serve as the building blocks for better ordering algorithms and extended functionality. This release of the package is totally within the public domain; there are absolutely no licensing restrictions as with other software packages. The development of this software was funded by ARPA (Contract DABT63-95-C-0122) with the express purpose that others (academic, government, industrial and commercial) could easily incorporate the data structures and algorithms into application codes. All we ask is an acknowledgement in derivative codes and any publications from research that uses this software. And, we hope that any improvements will be communicated to others.
References in zbMATH (referenced in 1 article )
Showing result 1 of 1.
- Çatalyürek, Ümit V.; Aykanat, Cevdet; Kayaaslan, Enver: Hypergraph partitioning-based fill-reducing ordering for symmetric matrices (2011)