FSPAI: Difference between revisions

From Sccswiki
Jump to navigation Jump to search
No edit summary
No edit summary
Line 23: Line 23:


== Implementation and Features ==
== Implementation and Features ==
<!--
The entire implementation is done in C++ with parallelization in MPI. Except the Block <br> SPAI approach, we cover the full functionality of SPAI 3.2.


=== Methodical improvements ===
* Single source providing
* Extension to target form: allow probing.
** highly scalable parallel implementation
* Support for explicit and inverse approximations, both in factorized <br> and unfactorized form and probing of Schur complements, see [http://www5.in.tum.de/software/mspai/mspai_variants.pdf pdf].
** sequential (MPI free) implementation
* Improve given factorized preconditioners such as ILU, AINV, FSAI, <br> FSPAI, etc subject to probing subspaces.
* Support for real and complex valued problems
* Compute sparse spectrally equivalent approximations to dense or <br> even full matrices.
* PCG solver available using HYPRE package
* Written in C/C++


=== Technical improvements ===
We are currently working on a next release which will cover
* Support for complex valued problems.
a highly scalable block version of FSPAI.
* Support for sparse QR methods using [http://www.cise.ufl.edu/research/sparse/CSparse/ CSparse] by Tim Davis.
* Caching approach to avoid redundant QR decompositions.
* Implementation of QR updates to accelerate pattern update steps.
* Support for maximum sparsity patterns.
* Arbitrary start patterns, i.e. possibility to compute a ''static'' SPAI <br> without pattern update steps.
 
=== Todo ===
* Mex interface for MATLAB.
* PetSc interface.
* Support for other LAPACK implementations than ATLAS.
* Wider coverage of file formats for sparse matrices, now support for <br> Matrix Market format only.
* Full support for complex problems in all features.
-->


== Download ==
== Download ==
Line 95: Line 81:


== Some References ==
== Some References ==
<!--
 
=== Papers ===  
=== Papers ===  


<pubsccs>nocaption=1&pubid=1463&lang=en</pubsccs><pubsccs>nocaption=1&pubid=1423&lang=en</pubsccs><pubsccs>nocaption=1&persid=53&utypid=1020&datum=2007&lang=en</pubsccs><pubsccs>nocaption=1&pubid=620&lang=en</pubsccs><pubsccs>nocaption=1&pubid=645&lang=en</pubsccs><pubsccs>nocaption=1&pubid=654&lang=en</pubsccs><pubsccs>nocaption=1&pubid=677&lang=en</pubsccs><pubsccs>nocaption=1&pubid=499&lang=en</pubsccs>
<pubsccs>nocaption=1&pubid=620&lang=en</pubsccs><pubsccs>nocaption=1&pubid=645&lang=en</pubsccs><pubsccs>nocaption=1&pubid=677&lang=en</pubsccs><pubsccs>nocaption=1&persid=53&utypid=2030&datum=2008&lang=en</pubsccs><pubsccs>nocaption=1&pubid=1140&lang=en</pubsccs>
 
=== Theses ===
 
<pubsccs>nocaption=1&persid=53&utypid=2030&datum=2008&lang=en</pubsccs><pubsccs>nocaption=1&pubid=1140&lang=en</pubsccs><pubsccs>nocaption=1&lang=en&persid=58&datum=2008&utypid=2040</pubsccs>


=== Further ===
=== Further ===


* Short summary on SPAI with core reference list: [http://www5.in.tum.de/software/mspai/summary.pdf summary.pdf]  
* Short summary on sparse approximate inverses with core reference list: [http://www5.in.tum.de/software/mspai/summary.pdf summary.pdf]  
* Extended reference list: [http://www5.in.tum.de/software/mspai/extended.pdf extended.pdf]
* Extended reference list: [http://www5.in.tum.de/software/mspai/extended.pdf extended.pdf]
-->
 
<!--
<!--
== Successful Applications ==
== Successful Applications ==

Revision as of 09:10, 15 July 2011

Factorized Sparse Approximate Inverses

This is a dummy textLatest release is 1.0


We developed a sequential and highly scalable parallel C/C++ implementation of the known FSPAI (Factorized SParse Approximate Inverses) algorithm.


Theory

Implementation and Features

  • Single source providing
    • highly scalable parallel implementation
    • sequential (MPI free) implementation
  • Support for real and complex valued problems
  • PCG solver available using HYPRE package
  • Written in C/C++

We are currently working on a next release which will cover a highly scalable block version of FSPAI.

Download

Tested environments

Some References

Papers

<pubsccs>nocaption=1&pubid=620&lang=en</pubsccs><pubsccs>nocaption=1&pubid=645&lang=en</pubsccs><pubsccs>nocaption=1&pubid=677&lang=en</pubsccs><pubsccs>nocaption=1&persid=53&utypid=2030&datum=2008&lang=en</pubsccs><pubsccs>nocaption=1&pubid=1140&lang=en</pubsccs>

Further

  • Short summary on sparse approximate inverses with core reference list: summary.pdf
  • Extended reference list: extended.pdf


Authors


License

FSPAI: Factorized Sparse Approximate Inverses
Copyright © 2010-2011, Matous Sedlacek
Research Unit Scientific Computing in Computer Science - Informatics V
Technische Universität München

This program is free software: you can redistribute it and/or modify it under the
terms of the GNU Lesser General Public License as published by the Free Software
Foundation, either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with
this program. If not, see http://www.gnu.org/licenses/.

If you obtain any results with FSPAI we would appreciate that you refer to FSPAI.


Further work on Sparse Approximate Inverses

Retrieved from "https://www5.in.tum.de/wiki/index.php?title=FSPAI&oldid=9302"