Could the old calculated results be reused in new DeviceLCAOCalculator ??

[anyuezhiji]

For instance：
1. An atom is slightly moved
2. An new atom is added
For an existing atom, read the value of each basis set from the old calculation used to generate the electron density
For an new added atom,  regard as neutral atom to generate the electron density

[Anders Blom]

This is possible for 1, but not for 2, today. We will consider 2 as a feature idea, although I think a more generic approach might be to take the real-space denstiy from the previous calculation and use that to initiate the scf loop instead.

[anyuezhiji]

This is possible for 1, but not for 2, today. We will consider 2 as a feature idea, although I think a more generic approach might be to take the real-space denstiy from the previous calculation and use that to initiate the scf loop instead.

Thanks very much for your reply.

If some atoms are monolithic moved, and the scattering region is resized,  could any information be read from the previous calculation and use that to initiate the scf loop ? If could ,can you provide an example script ?



For further,could some information be read from old calculations by specifying corresponding index number of atoms while the corresponding atoms is in similar environment? And regarded as neutral atoms while not specified.

[Anders Blom]

The rule is that the new configuration must have the same atoms, in the same order, and with the same basis set on each atom, as in the old calculation. If this is fulfilled, you initialize the new calculation from the old one using the keyword "initial_state". For an example, see http://quantumwise.com/publications/tutorials/mini-tutorials/143