Numerical diagonalization of 3x3 matricesA common scientific problem is the numerical calculation of the eigensystem of symmetric or hermitian 3x3 matrices. If this calculation has to be performed many times, standard packages like LAPACK, the GNU Scientific Library, and the Numerical Recipes Library may not be the optimal choice because they are optimized mainly for large matrices. This website offers C and FORTRAN implementations of several algorithms which were specifically optimized for 3x3 problems. All algorithms are discussed in detail in the following paper: Joachim KoppEfficient numerical diagonalization of hermitian 3x3 matrices Int. J. Mod. Phys. C 19 (2008) 523-548 arXiv.org: physics/0610206 |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Download and InstallationAll algorithms are released as C and FORTRAN code. The tarballs contain the respective source code including all dependencies and header files. You have the following download options: The complete package (including Makefile to generate a shared library)
Algorithms for real symmetric matrices
Algorithms for complex hermitian matrices
This software cannot be used as a standalone program. Instead, it is intended to be used as a dynamically linked library, or copied directly into your own programs, as is fit for your purposes. Make sure that you also include all required header files and dependencies. The C code follows the C99 standard, which includes the data type double complex to support complex numbers. For the GNU compiler this requires usage of the option -std=c99. For optimum performance we recommend switching on all optimizations provided by your compiler. For the GNU compilers, this is done by the option -O3. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Terms of UseThis software is freely available under the terms of the GNU Lesser General Public License (LGPL). However, since it has mainly been developed for academic puposes, the author would appreciate to be given academic credit for it. If you use one of the algorithms to prepare a scientific publication or talk, please cite the reference given above. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Bug reports / ContactAlthough this software has been thoroughly tested, we cannot guarantee that it is free of bugs. Therefore, bug reports, comments, and suggestions are always welcome. If you want to contact the author, you can do so under the following email address: jkopp [at] fnal.gov |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Latest changes12 March 2012
October 2006
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Last modified: 12 Mar 2012 04:55:26 CDT |