/ParMETIS

ParMETIS - Parallel Graph Partitioning and Fill-reducing Matrix Ordering

Primary LanguageCOtherNOASSERTION

ParMETIS

ci

ParMETIS is an MPI-based library for partitioning graphs, partitioning finite element meshes, and producing fill reducing orderings for sparse matrices. The algorithms implemented in ParMETIS are based on the multilevel recursive-bisection, multilevel k-way, and multi-constraint partitioning schemes developed in our lab.

Downloading ParMETIS

You can download ParMETIS by simply cloning it using the command:

git clone https://github.com/KarypisLab/ParMETIS.git

To build ParMETIS you can follow the instructions below:

cmake --workflow --preset default

Definitions of supported data types

ParMETIS uses the same data types for integers and floating point numbers (32/64 bit integers and single/double precision floating point numbers) as used when configuring and building METIS.

Copyright & License Notice

Copyright 1998-2020, Regents of the University of Minnesota