Technical

• Linear scaling techniques are used to speed up calculations on large unit cells
• SCF convergence based on a Direct Inversion of Iterative Subspace (DIIS) method
• The implementation is built upon a highly optimized numerical integration scheme for the
  evaluation of matrix elements of the Hamiltonian, property integrals involving the charge density, etc.
  This is the same numerical integration scheme as used in ADF.
• The program has been parallelized and vectorized
• Basis functions are Slater-Type Orbitals (STOs) and/or Numerical Orbitals (NOs).
• Fit functions are Slater-type exponential functions centered on the atoms and are used to fit the
  deformation density, which is the difference between the final density and the startup density.
  The deformation density has zero charge and will in general be small. The fitted deformation density
  is used for the calculation of the Coulomb potential and the derivatives of the total density (needed
  for the gradient corrections in the exchange-correlation functionals). In both cases the main part,
  due to the startup density, is calculated accurately by a numerical procedure, and only
  the small part from the deformation density is obtained via the fit.
• A frozen core facility is provided to allow efficient treatment of the inner atomic shells.
• Space group symmetry is used to reduce the computational effort in the integrals over the Brillouin
  zone.