|
|
ILUPACK V2.4
|
 |
ILUPACK provides several multilevel ILU preconditioners for general
real and complex matrices as well as real and complex symmetric
(Hermitian)
positive definite systems.
The basic idea of ILUPACK is based on inverse-based ILUs, that are
incomplete LU decompositions that control the growth of the inverse
triangular
factors.
ILUPACK V2.4 update 2016
New features
- simple user-friendly interface that automatically selects the appropriate diver (similar to MATLAB toolbox)
- some MATLAB drivers added
- some bugs fixed
ILUPACK V2.4 released June 2011
New features
- some simple example C++ programs added
- MATLAB interfaces partially updated to better interact with recent releases
- some bugs fixed
ILUPACK V2.3 released December 2010
New features
- some simple example programs that demonstrate more easily the use of ILUPACK
- more MATLAB interfaces for maximum weight matchings and reorderings, transposed solver
- block-structured solvers (still under development)
- mixed precision (single/double)
ILUPACK V2.2 released November 2008
New features
- New user-friendly interface similar to MATLAB
- Some experimental multigrid solvers (with smoothing, V-, W-cycle, etc)
- elbow space only optional (recommendation) parameter
- MATLAB drivers for metis reordering and symmetric maximum weight matchings
ILUPACK V2.1 released in January 2006
New features
- ILUPACK toolbox for MATLAB
- drivers that combine different preconditioners and iterative solvers
- interface ILUPACK to be used from FORTRAN77
ILUPACK V2.0 released in May 2005
New features
- preconditioners for real symmetric, complex Hermitian and complex symmetric systems
-
Symmetric maximum weight matching (used for symmetric indefinite preconditioners)
- new dropping strategies using aggressive dropping
ILUPACK V1.1 released on December 3, 2004
New features
- Many more interfaces with respect to existing software packages
- interface for AMD from UMFPACK V4.3
- interface for MPS (maximum weight matching) from PARDISO
- interface for MeTiS multilevel nested dissection
- various combination of maximum weight matching (MC64 and MPS) followed by
symmetric reorderings
- backward error as stopping criterion
- several variants of preconditioning (besides right preconditioning, which is default)
- some more features ...
ILUPACK V1.0 released on May 14, 2004
m.bollhoefer@tu-bs.de
Last modified: August 31, 2016