[ase-users] PreconLBFGS and broken symmetry

[Jens Jørgen Mortensen]

Den 25-05-2017 kl. 15:56 skrev James Kermode:
>> On 24 May 2017, at 13:34, Noam Bernstein <noam.bernstein at nrl.navy.mil> wrote:
>>
>> Another possibility it to use my symmetrizing calculator (which uses spglib) to symmetrize the trial step (won’t work if it's orthogonal to the symmetry allowed displacement subspace).
> This sounds good - we need a couple of trial configurations to do a finite difference estimate for the energy scale, so just moving along the forces is not really an option.

How about letting PrecondLBFGS ask the calculator to symmetrize the 
displacements for the trial configs (disp = 
calc.symmetrize_displacement(disp))?  It should be allowed for a 
calculator to not have that method.

Jens Jørgen

>
> James.