I am totally new to math/numerical analysis programming and I was looking all over the internet to find appropriate libraries.
What I stumbled upon are CLAPACK, Boost::uBLAS, ATLAS. I need to solve SLEs (system of linear equations) so I guess BLAS itself would not开发者_如何学Go be sufficient without some implementation of LAPACK.
None of them seem to have easy and intuitive way to work on Windows platform.
Would there be any recommendations?
Intel MKL seems tempting but I cannot buy it because it is an academic project.
As far, as I understand the best way is to use Cygwin and compile ATLAS for my system and use LAPACK extension.
Thanks in advance for any thoughts.
EDIT
I decided upon using Eigen and trying Intel MKL under linux one day.Also, this seems to be a good brief overview of possible options http://gears.aset.psu.edu/hpc/software/numlib/
Mkl libraries are available free of charge under noncommercial license, search Google and you will find it. http://software.intel.com/en-us/articles/non-commercial-software-development/
I do not know what SEL are, system linear equations? boost ublas has LU decompose which you can use to solve a few problems. Ublas is fairly straightforward to install on Windows.
Does "SLEs" means systems of linear equations? If yes, try Eigen. It has stable and well tested with large matrices LU decomposition capabilities.
In my experience, GSL works fine on Windows.
The GNU Scientific Library (GSL) is a numerical library for C and C++ programmers. It is free software under the GNU General Public License. http://www.gnu.org/software/gsl/
In addition to mentioned ones (from which I prefer GSL and Eigen) there is also simple library called TNT
http://math.nist.gov/tnt/ which is very nice for simple tasks. It is very easy to use and does not require bunch of configuration.
精彩评论