Stable Version armadillo-3.820.1.tar.gz   Old Versions armadillo-3.810.3.tar.gz armadillo-3.800.3.tar.gz   These versions are no longer maintained; they are kept in case there are any regressions in the new stable version

---

Installation and Configuration Notes
All Platforms · Linux · Mac OS X · Windows
All Platforms

• Please see the README.txt file in the .tar.gz package
   
• If you encounter any problems or regressions, please report them
   
• If you use Armadillo in your research and/or software, please cite the overview tech report. Citations are useful for the continued development and maintenance of the library.

Linux   (Fedora, Ubuntu, Red Hat, SUSE, Debian, etc)

• Armadillo can work without external libraries. However, to get the most functionality it's recommended to install (in advance) the LAPACK, BLAS and ATLAS libraries, as well as Boost libraries (version 1.34 or later), along with the corresponding development/header files.
   
• For best performance, we recommend using the multi-threaded OpenBLAS library instead of standard BLAS.
   
• Many Linux-based operating systems provide pre-built Armadillo packages, eg. Fedora, Debian, Ubuntu, openSUSE. These packages may not be the latest version. If you're encountering problems, please use the official packages provided here.
   
• Recommended packages for Fedora & Red Hat (installed before Armadillo): cmake, blas-devel, lapack-devel, atlas-devel, boost-devel. You may wish to install ATLAS packages that are optimised for your CPU, eg. atlas-sse2.
   
• Recommended packages for Ubuntu & Debian (installed before Armadillo): cmake, libopenblas-dev, liblapack-dev, and libboost-dev.
   
• If you're updating from a previous version, it's a good idea to remove all the old files before updating. The files are typically in /usr/include/armadillo* as well as the library files in /usr/lib/ or /usr/lib64/

Mac OS X

• The "Accelerate" framework is used for accessing BLAS and LAPACK functions. See the README.txt file in the package for more information.
   
• If you're updating from a previous version, it's a good idea to remove all the old files before updating. The files are typically in /usr/include/armadillo* or /usr/local/include/armadillo* as well as the library files in /usr/lib/ or /usr/local/lib/

Windows

• As a courtesy, the Armadillo package contains pre-compiled 32 bit versions of LAPACK and BLAS for Windows, as well as MSVC project files to compile the example programs.
   
• If you're getting messages such as "use of LAPACK needs to be enabled", manually modify "include/armadillo_bits/config.hpp" to enable the use of LAPACK. See the README.txt file in the package for more information.
   
• The MSCV project files were tested on 32 bit Windows XP with Visual C++ 2008. You may need to make adaptations for 64 bit systems, later versions of Windows and/or the compiler.
   
• If the pre-compiled versions of LAPACK and BLAS don't work for you, or you want more speed, try these:
  
  
  • http://software.intel.com/en-us/intel-mkl/
  • http://www.amd.com/.../amd-core-math-library-acml/
  • http://xianyi.github.com/OpenBLAS/
  • http://www.stanford.edu/~vkl/code/libs.html
  • http://icl.cs.utk.edu/lapack-for-windows/lapack/
  • http://www.fi.muni.cz/~xsvobod2/misc/lapack/


• For more info on the use of high-speed BLAS and LAPACK replacement libraries (such as Intel MKL and AMD ACML), see the FAQ.
   
• If you're using a 64 bit system (such as Windows 7), dedicated 64 bit versions of LAPACK and BLAS are considerably faster. You may wish to see this example of using such 64 bit versions.
   
• For faster performance, use the high-quality MinGW GCC C++ compiler. For best results we also recommend using an operating system that's more suitable for heavy duty work, such as the Linux-based Ubuntu or Fedora operating systems.