1. REQUIREMENTS 
    Addibit can be compiled with g++ 4.4 or above. It requires GMP (The GNU Multiple Precision Arithmetic Library: http://gmplib.org/) and
    GSL (The GNU Scientific Library: http://www.gnu.org/software/gsl/). Alternatively, you can use the precompiled binary.

2. COMPILE
    In order to compile execute
     1. ./configure --prefix=$PWD USE_TEMPLATES=1 INFO=1 DEBUG=0 COMPRESSION=1
     2. make
     3. make install
    This will create a new executable file 'addibit' in the folder 'bin'. If you compile on a machine supporting the sse4.2 instruction
    set then the program will use the POP_CNT instruction. You may turn it off by configuring with 'USE_SSE4_2=0'.

3. INPUT 
    In order to run the program on a single problem select one of the configuration files 'settings_bp-trees.conf',
    'settings_bp-qbn-trees.conf' or 'settings_vp-trees.conf' and run './addibit PATH_TO_INPUT_FILE PATH_TO_CONFIGURATION_FILE'.
    The input file should be in the conventional 'ine' format:
        ..
        begin
        rows columns integer
        1 3 4 3 ...
        ...
        end
        ..

    The program assumes that the input matrix has full column rank. Providing another matrix will
    result in an error message and program termination. Please note that (for the moment) we support
    only integers in the input files. Using floating point numbers or rationals leads to unspecified
    behavior.

4. OUTPUT
    On success the extreme rays will be written into a new file which has the name of the input one plus
    a suffix '.out'. You can specify the output form for the extreme rays in the configuration file. Valid
    options are
        - ACTIVE_SET for binary vectors identifying the row vectors of the input matrix which the extreme ray satisfies with equality,
        - INTEGER_VEC for calculation of arbitrary precision integer vectors using GMP and
        - DOUBLE_VEC for calculation of floating point vectors using GSL.

5. BENCHMARK
    In order to generate the benchmark results run './benchmark.sh benchmark PATH_TO_ADDIBIT_BINARY'. This will create a new folder
    'benchmark/paper/doc' containing three log files for each problem: one for the calculation with bp-trees, one for bp-qbn-trees
    and one for vp-trees. A summary of all results can be found under 'benchmark/doc/benchmark.html'. If you want to run the program
    on a different set of problems just create a new folder 'my_new_benchmark' and run the script on it:
    './benchmark.sh my_new_benchmark PATH_TO_ADDIBIT_BINARY'.