Note: It is recommended to use the Visual Studio Code dev-container as long as you only need to run calculations on your local computer. Ensuring the relevant dependencies are correctly installed is a difficult problem. Only follow these instructions if you need a custom install, e.g. because you cannot use Docker, need to run on a custom configured supercomputer, or need access to special hardware such as GPU.

System requirements

Required software and libraries	Further information
git	https://git-scm.com/
CMake (Minimum version 3.1)	https://cmake.org
BLAS	http://www.netlib.org/blas
LAPACK	http://www.netlib.org/lapack

Download TBTK

git clone https://github.com/dafer45/TBTK/

Select version

Always work agains a public release. This is done by checking out the latset release using

git checkout v2.6.5

A full list of public releases can be found at https://github.com/dafer45/TBTK/releases. By remembering the version number that an application is compiled with, it is guaranted that the application will be possible to build again anytime in the future.

TBTK uses semantic versioning. This means that any application that compiles with v2.x.y also will work with any higher version v2.x'.y'. An update to the first (major) version number may break older applications.

Install TBTK

Unix like operating systems such as Linux and Mac OS

Begin by creating a build folder in parallel with the TBTK source folder (i.e. not inside).

mkdir TBTKBuild

The folder structure should now be

SomeFolder/
    TBTK
    TBTKBuild

Then do the following.

cd TBTKBuild
cmake ../TBTK
make
sudo make install

Local installation

It is possible to perform a local install if administrator privileges are not available. To do so, add the flag -DCMAKE_INSTALL_PREFIX=install/path/ to the cmake call above, where install/path/ is the preferred installation path. Make sure to also set the environment variables CPLUS_INCLUDE_PATH, LIBRARY_PATH, and PATH if this is done. The paths should be install/path/include/, install/path/lib/, and install/path/bin/, respectively.

Customized installation

It is possible to customize the build by passing these flags to cmake.

Flag	Description
-DCMAKE_INSTALL_PREFIX:PATH=XXX	Specify a custom insallation path. For example a local path if administrator privileges are not available.
-DARPACK_LIBRARY_PATH:PATH=XXX	Specify a non standard search path for ARPACK library files.
-DSUPER_LU_INCLUDE_PATH:PATH=XXX	Specify a non standard search path for SuperLU include files.
-DSUPER_LU_LIBRARY_PATH:PATH=XXX	Specify a non standard search path for SuperLU library files.

where XXX is to be replaced by the relevant path.

Extensions

Additional features will also be available if one or more of the following libraries are installed.

Optional libraries	Further information
ARPACK	http://www.caam.rice.edu/software/ARPACK
CUDA	https://developer.nvidia.com/cuda-downloads
cURL (libcurl)	https://curl.haxx.se
FFTW3	http://www.fftw.org
GMP	https://gmplib.org/
Matplotlib (Python)	https://matplotlib.org/
NumPy (Python, C library)	https://numpy.org/
Python	https://www.python.org/
SuperLU (v5.2.1)	http://crd-legacy.lbl.gov/~xiaoye/SuperLU

The following table shows the optional libraries that are required to enable the respective TBTK components.

	ARPACK	CUDA	cURL	FFTW3	GMP	Python with matplotlib and numpy	SuperLU (v5.2.1)
ArbitraryPrecision					X
ArnoldiIterator	X						X
ChebyshevExpander (GPU execution)		X
FourierTransform				X
LinnearEquationSolver							X
LUSolver							X
Plotter						X
Resource			X

Update TBTK

If TBTK is already installed, it can be updated to v2.6.5 as follows. We here assume that TBTK has been downloaded and built in two parallel folders TBTK and TBTKBuild, as described above.

cd TBTK
git pull
git checkout v2.6.5
cd ../TBTKBuild
cmake ../TBTK
make
sudo make install

Next: Manual