matt-chan / numopt

A C++ library for performing numerical optimization.

numopt v1.6.0

A C++ library for performing numerical optimization.

Dependencies

Installation

To install this library:

1. clone the master branch, which contains the latest release:

    git clone https://github.com/GQCG/numopt.git --branch master --single-branch --recurse-submodules
    cd numopt
   

2. perform an out-of-source cmake build:

    mkdir build && cd build
    cmake -DINSTALLATION_PREFIX=prefix ..
    make && make test && sudo make install
   

   where

   • prefix is the installation prefix (defaulted to /usr/local) you want the library to be installed at:
     • the library libnumopt.a will be installed in prefix/numopt/lib
     • the header files (and cmake files, see Usage) will be installed in prefix/numopt/include
   • if you want to use as a backend for BLAS/LAPACK operations, you should add the USE_MKL option

   cmake -DINSTALLATION_PREFIX=prefix -DUSE_MKL=ON ..    
   

Usage

Basic usage of this library can be found in the tests directory. If you use CMake in other projects, you can add the following CMake command to the CMakeLists.txt-file:

find_package(numopt 1.6.0)

CMake then provides the commands numopt_INCLUDE_DIRS to be used in your target_include_directories and the library numopt to be used in your target_link_libraries.