#HODLR PACKAGE : Fast dense solver for hierarchically off-diagonal low-rank matrices

This software package is a highly felxible HODLR solver that can be used in a variety of applications. The package can solve matrices that are preallocated in memmory or defined by a kernel function.

A variety of low-rank approximation methods (SVD, partial pivoting ACA, full pivoting ACA, BDLR, etc.) are available. Furthermore, the package also accepts any custom HODLR partitioning from the user which is very handy in many applications.

####Author :

Amirhossein Aminfar: amir_aa127@yahoo.co.uk

####Citation:

If you use the implementation or any part of the implementation in your work, kindly cite as follows:

####Article:

@article{aminfar2014fast,

author={{A}minfar, {A}mirhossein and {A}mbikasaran, {S}ivaram and {D}arve, {E}ric},

title={A Fast Block Low-Rank Dense Solver with Applications to Finite-Element Matrices},

journal={arXiv:1403.5337},

year={2014}

}

####Code

@MISC{aminfar2014HODLR,

author = {{A}minfar, {A}mirhossein},

title = {A fast direct solver for HODLR matrices},

howpublished = {https://github.com/amiraa127/Dense_HODLR},

year = {2014}

}

####Version 1.00

Date: September 30th, 2014

Copyleft 2014: Amirhossein Aminfar

Developed by Amirhossein Aminfar

####License

This program is free software; you can redistribute it and/or modify it under the terms of MPL2 license. The Source Code Form is subject to the terms of the Mozilla Public License, v. 2.0. If a copy of the MPL was not distributed with this file, You can obtain one at http://mozilla.org/MPL/2.0/.

####Build

The easiest way to build the library is to use CMake. Go to the project directory and run:

mkdir build
cd build
cmake ../
make

####Documrntation

The documentation for this package is currently located at http://amiraa127.github.io/Dense_HODLR/. It's currently incomplete but I will add material to it gradually.