Semilocal DFT at the GGA level underestimates
long-range dispersion interactions.
In the case of graphite, PBE predicts the
interlayer binding energy of ~1 meV/atom
which is too small compared to the RPA 
reference of 0.048 eV/atom
(Lebgue et al., PRL 105, 195401 (2010)).
 
In this example, the interlayer binding energy 
of graphite in its experimental structure is 
determined using the TS method 
of Tchatchenko and Scheffler (PRL 102, 073005 (2009)),
which performs well in description of the 
structure of graphite (see e.g. example 
graphiteDistance_ts).

The calculation is performed in two steps (sigle-point
calculations) in which the energy for bulk graphite 
and for graphene are obtained. The binding energy 
is computed automatically and it is written in the 
file results.dat.

Even though the TS method predicts a reasonable geometry
it overestimates the energetics strongly:
the computed binding energy of -0.083 eV/atom is too 
large compared to the RPA reference of 0.048 eV/atom
This overestimation is - at least in part - due to
neglecting the many-body interactions (see example 
graphiteBinding_mbd).

Details of implementation of TS in VASP + a number
of tests:
Bucko et al., PRB 87, 064110 (2013).

