The Installation process needs three steps:
Before you are able to install AMfe you need a Python environment. We recommend to install the Anaconda distribution. It is a free python distribution which is both available for windows and for linux/mac and focuses on scientific computing which is best suitable for programs like AMfe. The distribution may be used for business or private purposes. You do not need root access to install Anaconda. You may install it in your home-directory.
- Download Anaconda from http://www.continuum.io/downloads.
- Install Anaconda with installation instructions from http://www.continuum.io/
Important
Make sure that you install an Anaconda version that provides a Python-version newer or equal Python 3.5
Note
We recommend to add anaconda to your PATH-variable after installation.
When you have installed anaconda successfully the most important packages for using AMfe, such as numpy and scipy, should also have been installed automaticly.
If you do not want to install anaconda and want to use another python distribution, check if you have a python version greater than 3.5 and that you have installed the newest versions of the packages listed in the following table:
Module Numpy Scipy Pandas Sphinx Numpydoc h5py Min. Version 1.11.1 0.18.1 0.18.1 1.4.6 0.6.0 2.6.0
Usually you can check the version number by first opening a pyhton3 console via:: bash
user@host:~$ python3
and then entering:
>>> import modulename
>>> print(modulename.__version__)
For a matlab like environment we recommend Spyder. Spyder includes a python editor with syntax-highlighting, IPython Condole, Documentation-Window, Variable explorer and more. In Anaconda it is usually already installed.
There are different ways to install AMfe. The way which is best for you depends on your application and plans for using AMfe. The following table lists different ways and potiential users
| Way of installation | Recommended for | Platforms |
| Git clone/fork | Developers, Users who want to change/edit source-code or want to contribute to AMfe-project | linux/win/mac |
| Non-Developer source-code | Users who like to build source-code by themselves, but who do not intend to change the source-code. People who do not intend to change the source but want best performance of AMfe by using compilers on their own target machine. | linuc/win/mac |
| Already built code | Users who are not familiar with compilers/computers or users who want a system out of the box. | linux/win/mac |
Click on one of the following links for details:
1. Installation git clone/fork
For speedup of the solver-routines, it is recommended to install pyMKL. This is package is a wrapper for speed-up the Pardiso-Solver by using the Intel Math Kernel Library. The package pyMKL can be downloaded from https://github.com/c-meyer/pyMKL
Download the repository, unzip it to an arbitrary folder change to it and run:
python setup.py install
This command installs the pyMKL-package to your python environment.