Concurrent Number Cruncher:
    An Efficient Sparse Linear Solver on the GPU

    This is a package developed by Luc Buatois: you can find the original reference at http://alice.loria.fr (includes paper and software).

    The original package was developed in a Windows environment, but I needed to make it available on my local machine (GNU/Linux, Ubuntu 9.04, CUDA 2.3, with an NVidia GTX 295 graphical card).

    I have repackaged the source code using the GNU autotools (autoconf and automake) to build it with the GNU toolchain.

    If you have to use this package in a Linux box, feel free to

    DOWNLOAD cnc-gpu-cuda-0.9-2.tar.gz

    Here you can find an overview on Sparse Matrix formats
    joomla visitor