Getting the DFT++ source code

The preferred way is through the darcs repository. We strongly encourage you to use this approach since you will not only get the latest bugfixes but it will allow you to easily contribute back any fixes to bugs you may find. Here is how to get started:

Obtain darcs, either from your OS distribution, by putting a statically linked binary in your path, or by compiling darcs from source.
Download the latest version of dft:
darcs get --partial http://dft.physics.cornell.edu/dft new_dft_dir
sh autogen.sh in the newly created directory in order to generate the configure- and make- files. You need a recent version of the autoconf/automake tools. Older versions of these tools sometimes trigger errors, but in almost all cases running by hand the 4 commands in autogen.sh does the trick.
./congfigure (use ./configure --help to see the various options for mpi, blas, etc.)
make will create the executable dft in the src/ subdirectory.
darcs pull to update your repository with the latest changes (after say few days).
To contribute your improvements
- darcs record --no-test, will interactively prompt you to name a set of changes.
- darcs send will email your changes to the dft-users mailing list in a form that we can easily apply to the code. If your system does not have configured email use darcs send -o FILENAME to create a patch file and email it to us as an attachment.

Alternatively you can download the tarball from the latest release. In this case you do not need the automake/autoconf tools installed, but you still need FFTW.
- Download the latest tarball from here .
To see dft in action
- make check will run in the subdirectory tests a set of simple examples that we use to test the code integrity.
- make in the subdirectory bench will run a set of more realistic examples.