Programs and libraries for graph, mesh and hypergraph partitioning

Overview Repositories Revisions Requests Users Attributes Meta

Programs and libraries for graph, mesh and hypergraph partitioning

Programs and libraries for graph, mesh and hypergraph partitioning Its purpose is to apply graph theory, with a divide and conquer approach, to scientific computing problems such as graph and mesh partitioning, static mapping, and sparse matrix ordering, in application domains ranging from structural mechanics to operating systems or bio-chemistry.

The SCOTCH distribution is a set of programs and libraries which implement the static mapping and sparse matrix reordering algorithms developed within the SCOTCH project.

Sources inherited from project SUSE:SLE-15-SP3:GA
Download package
Checkout Package
osc -A https://api.opensuse.org checkout SUSE:SLE-15-SP7:GA/scotch && cd $_
Create Badge

Build Results
RPM Lint

Refresh

Source Files

Filename	Size	Changed
_multibuild	0000000457 457 Bytes	almost 4 years ago
scotch-Makefile.inc.in	0000000455 455 Bytes	almost 6 years ago
scotch-v6.1.0.tar.gz	0006292006 6 MB	almost 4 years ago
scotch.changes	0000008429 8.23 KB	almost 4 years ago
scotch.spec	0000018924 18.5 KB	almost 4 years ago

Latest Revision

Gustavo Yokoyama Ribeiro (gyribeiro) committed almost 4 years ago (revision 2)

Bring all relevant HPC Packages to the latest Version available
(jsc#SLE-14839).

- Sync SLE and openSUSE builds (jsc#SLE-12162).

- Update to version 6.1.0. The change in release number reflects a change  
  in behavior of the software, regarding two features:
  * Algorithms for sparse matrix reordering now fully take into account vertex
    weights. While the previous versions only accounted for them in the nested
    dissection method to compute and  optimize separator size/weight, they are
    now also accounted for in the minimum  degree and minimum fill algorithms.
    Consequently,  vertex  amalgamation  can  be  tuned so  that  the  size of
    (column-)blocks is tailored according to the real amount of computation to
    be performed on supervariables.
    The esmumps  interface  with MUMPS  has been  updated  so as to  take into
    account this new feature. 
  * The routines of the vertex  graph partitioning (with overlap)  module have
   been redesigned, leading to much smaller  run time and, most often, higher
   quality
- Scotch development is moving to gitlab
- Add support for gcc10 and build support for openmpi4.

Places