MORSy is an open source library for computing fill-reducing ordering of sparse matrix. It is based on the multilevel nested dissection algorithm with modifications for using vertex separator at all steps of multilevel scheme. MORSy has been developed at the Department of Computational Mathematics and Cybernetics at the State University of Nizhny Novgorod. It is used in the High Performance Computing Center of the State University of Nizhny Novgorod for solving sparse systems of linear equations in the process of finite element simulation of heart activity.
PMORSy is a parallel version of MORSy library for shared-memory systems. Parallel processing is done in a task-based fashion. It uses OpenMP 3.0 task parallelism relying on the dynamic load balancing implemented in the OpenMP runtime.
MORSy and PMORSy is distributed under the LGPL licence. They are cross-platform and can be used under Linux and Windows operating systems.
MORSy_1.0.0.zip (432 Kb)
PMORSy_1.2.0.zip (402 Kb)
DMORSy_1.0 (97 Kb)
PMORSySample.zip (205 Kb)
|MORSy v. 1.0.0
|The basic version of MORSy
|PMORSy v. 1.0.0
|The basic version of PMORSy (shared-memory parallelism)
|PMORSy v. 1.2.0
|New algorithms for graph coarsening and graph initial partitioning were added