If the orbitals are initialized using a random number generator (the default in VASP), the initial orbitals are usually unreasonable and the iterative matrix diagonalization will require 5-10 steps to obtain reasonable orbitals. The charge density corresponding to the initial orbitals is also, at best, erratic. It is hence advisable to perform a few electronic steps while keeping the initial Hamiltonian fixed. This initial Hamiltonian is usually determined from a superposition of atomic charge densities (see ICHARG).

Such a 'delay' is absolutely necessary. if the SCF-convergence is slow and problematic (e.g. for surfaces or metallic clusters, low dimensional system). Without a delay, VASP will most likely not converge, or at least the convergence speed is slowed significantly.

NELMDL might be set to a positive or negative value. A negative value means that the delay is only performed in the first ionic step (usually the recommended option). A positive number means that a delay is employed after each ionic movement. This can improve the convergence speed in VASP.6 (see below) but is not recommended in VASP.5.

按照这个说明，我觉得我可以用vasp6重新试一试。现在用的版本是vasp5