Skip to content

Installation from Source: Historical notes on older versions

Oliver Beckstein edited this page Nov 17, 2015 · 2 revisions

For installing the latest versions of MDAnalysis see the Install notes.

Older versions of MDAnalysis had different requirements and for historical reasons they are listed here:

up to 0.7.2

A LAPACK library is required (standard LAPACK/BLAS, ATLAS, or the Intel Math Kernel Library (MKL); on Mac OS X we simply use the native fast VecLib framework so you don't have to install anything else).

Fast math libraries

On Mac OS X we use the systems vecLib for fast linear algebra and nothing needs to be configured. On Linux you will probably want to use something such as ATLAS or the Intel Math Kernel Libraries for better performance for the rms fitting.

The fast math library paths are set in setup.cfg. By default we ship a template file setup.cfg.template that you can

cp setup.cfg.template setup.cfg

and edit. See the file itself for examples and the ones below.

Example MKL

Linux section of setup.cfg (your paths are probably different from the ones in this example)::

  [linux]
  fast_numeric_include = /opt/intel/cmkl/10.0.5.025/include
  fast_numeric_linkpath = /opt/intel/cmkl/10.0.5.025/lib/em64t
  fast_numeric_libs = mkl_lapack mkl guide

Example ATLAS LAPACK

If you want to use the ATLAS LAPACK libraries use something such as ::

[linux]
fast_numeric_include = /usr/include
fast_numeric_linkpath = /usr/lib/atlas
fast_numeric_libs = lapack

Your paths could be different. Use a command such as

locate lapack

or your package manager to get an idea of where the files are to be found.

On some versions of Ubuntu we had to use

[linux]
# Ubuntu 9.04 i686
fast_numeric_include = /usr/include
fast_numeric_linkpath = /usr/lib/sse2/atlas
fast_numeric_libs = lapack

Adding lapack packages

 sudo apt-get install liblapack-dev   # see below for notes on fast_numeric_*
Clone this wiki locally