Prepare your System¶
LFA Lab depends on the following packages:
Furthermore, the following packages provide extra functionality, but are optional:
You can find the commands to install the dependencies for different operating systems below.
Currently, LFA Lab should work with Python 2 and pre C++11 compilers. Using those, however, is unsupported. It is recommended to use Python 3 and a recent C++ compiler.
To install the dependencies on Debian, run the following command:
sudo apt-get install -y \ g++ cmake python3-dev python3-numpy python-six swig \ libeigen3-dev liblapack-dev python-matplotlib
To install the dependencies on Fedora, run the following command:
sudo dnf install -y \ gcc-c++ cmake python3-devel python3-numpy python-six swig \ eigen3-devel lapack-devel python3-matplotlib
Mac OS X¶
First, you need to install a current version of Python. Visit the Python Website and download a recent version of Python 3, which you then install. Make sure that this Python version is your path by running the Update Shell Profile.command. After running this command, you need to close all terminal windows.
Next, we need to install the C++ dependencies, needed to compile the native Python extension of LFA Lab. An easy way to install these dependencies is the Homebrew package manager for Mac. Make sure you have Homebrew installed. Then, change into the source directory of LFA Lab and execute the following command:
Then, you can use PIP to install the remaining (Python) dependencies:
pip3 install --user --upgrade -r requirements.txt
Warning: You might have multiple Python versions on your machine. You have to make sure that all dependencies and LFA Lab are installed with the same version.
After you have prepared your system, you can build and install LFA Lab using pip. Change into the source directory of LFA Lab and execute:
pip3 install --user .
To check your installation, open a new terminal window and run:
You should see a message, which tells you that LFA Lab has been installed successfully.
If you want to customize the installation, e.g., in case the build fails, see the manual installation below.
Build LFA Lab¶
Execute the following commands in the source directory of LFA Lab:
cmake [OPTIONS] . make
(Do not forget the dot at the end.)
In case the build fails or you want to tweak your installation you can use the following options.
Install into a per-user directory:
Turns the compiler optimization on:
Choose to use LAPACK for certain operations. If a good LAPACK/BLAS implementation is available, this will speed up the program essentially:
Choose to use ARPACK. This will speed up the program if large spectra need to be analyzed. Arpack, however, might not be able to compute the spectra for certain tricky problems:
Set other prefices wich will be searched. For example if you installed some of the libraries in $HOME/.local run:
cmake -DCMAKE_PREFIX_PATH=$HOME/.local [OTHER OPTIONS] .
To install LFA Lab just run:
sudo make install
If you just want to use the software without installation, you can run:
instead. This command will setup the current shell session such that you can use LFA Lab.
You can now use LFA Lab. Take a look at the Tutorial page to find out how to use it.