|if WAVECAR exists||1|
This flag determines whether to read the file WAVECAR or not.
The set of plane waves will be redefined and re-padded according to the new cell size/shape (POSCAR) and the new plane wave cut-off (INCAR). These values might differ from the old values, which are stored in the file WAVECAR. If the file WAVECAR is missing or if file WAVECAR contains an inappropriate number of bands and / or k-points the flag ISTART will be set to 0 (see above). In this case VASP starts from the scratch and initialize the wave functions according to the flag INIWAV.
The usage of ISTART=1 is recommended if the size/shape of the supercell (see section 9.6) or the cut-off energy changed with respect to the last run and if one wishes to redefine the set of plane waves according to a new setting.
ISTART=1 is the usual setting for convergence tests with respect to the cut-off energy and for all jobs where the volume/cell-shape varies (e.g. to calculate binding energy curves looping over a set of volumes).
Mind: main.f can be recompiled with new settings for NGX,NGY,NGZ,NPLWV ... between different runs, the program will correctly repad and reorganize the 'storage layout' for the wavefunction arrays etc. In addition it is also possible to change the k-point mesh if the number of k-points remains constant. This might be of importance if a loop over a set of k-points (band-structure calculations) is performed.
The set of plane waves will not be changed even if the cut-off energy or the cell size/shape given on files INCAR and POSCAR are different from the values stored on the file WAVECAR. If the file WAVECAR is missing or if the file WAVECAR contains an inappropriate number of bands and/or k-points the flag ISTART will be set to 0 (see above). In this case VASP starts from the scratch and initialize the wave functions according to the flag INIWAV. If the cell shape has not changed then ISTART=1 and ISTART=2 lead to the same result.
ISTART=2 is usually used if one wishes to restart with the same basis set used in the previous run.
Mind: Due to Pullay stresses (section 9.6) there is a difference between evaluating the equilibrium volume with a constant basis set and a constant energy cut-off -- unless absolute convergence with respect to the basis set is achieved! If you are looking for the equilibrium volume, calculations with a constant energy cut-off are preferable to calculations with a constant basis set, therefore always restart with ISTART=1 except you really know what you are looking for (see section 9.6).
There is only one exception to this general rule: All volume/cell shape relaxation algorithms implemented in VASP work with a constant basis set, so continuing such jobs requires to set ISTART=2 to get a 'consistent restart' with respect to the previous runs (see section 9.6)!
Same as ISTART=2 but in addition a valid file TMPCAR must exist containing the positions and wave functions at time steps t(N-1) and t(N-2), which are needed for the wavefunction and charge prediction scheme (used for MD-runs).
ISTART=3 is generally not recommended unless an operating system imposes serious restriction on the CPU time per job: If you continue with ISTART=1 or 2, a relatively large number of electronic-iterations might be necessary to convergence the wave functions in the second and third MD-steps. ISTART=3 therefore saves time and is important if a MD-run is split into very small pieces (NSW<10). Nevertheless, we have found that it is saver to restart the wavefunction-prediction after 100 to 200 steps. If NSW>30 ISTART=1 or 2 is strongly recommended.
Mind: If ISTART=3, a non-existing WAVECAR or TMPCAR file or any inconsistency of input data will immediately stop execution.