VASP 5

VASP.5 is currently not distributed, expected release date not before 2007.

VASP.5 is a significant update from vasp.4.X. Internally it has been rewritten to separate the data representation from algorithms. With the new version it will be possible to implement different basis sets (for instance finite elements) without a major code rewrite.

VASP.5 supports or will support a large number of additional features (for internal tests vasp.5.1.21 is presently available):

• Optical properties, in particular the real and imaginary part of the frequency dependent dielectric function are supported (beta stage). (Sec. 6.64.1).
• Linear response with respect to an external field and with respect to the ionic positions is supported (beta stage) (Sec. 6.64.4).
• Most ``second order'' response functions, such as internal strain tensor, piezoelectric tensor, Born effective charges, interatomic force constants (alpha stage).
• Exact exchange and hybrid functionals (PBE0) are supported both in the $\Gamma$-point only version and for the full k-point version. The $k-$point sampling can be performed in the IRZ, because routines for the symmetrisation of the wavefunctions are now implemented. The estimated computing requirements will however increase dramatically; expect something like two orders of magnitude (beta stage).
• Screened exchange (beta stage) and model GW in the COHSEX (alpha stage) will be supported.
• Exact exchange in the framework of the optimized effective potential method will be supported (alpha stage, but unlikely to be ever released publicly: first very excited, later not so exciting at all).
• Full frequency dependent GW at the speed of the plasmon pole model: fully parallel, very fast (beta stage, Si 128 bands, $6\times 6 \times 6$ k-points takes 500-1000 seconds on dual Opteron). Very affordable, indeed. (Sec. 6.65).
• Bethe-Salpether, TD-DFT and TD-HF for excitons (alpha stage).
• Relaxed core PAW is supported (beta stage).
• A new direct optimization scheme for the KS functional is supported, which performs (almost) as fast and robust as the charge density mixing schemes (beta stage).
• A new matrix diagonalisations routine will be implemented, which will scale essentially like $N^2$ (concept exists, but work not yet in progress).
• Embedding in electrostatic point charges might be supported.
• Finite elements might be supported (work not yet in progress).