LRC-DFT gradients time plan?

feature-request
1

Hi,

The manual states they are not implemented yet, and I was wondering where on the to-do list the gradients
for long-range corrected DFT stand?
Are they already worked upon, or merely planned ‘soon’?

cheers

2

The likeliest answer is that they will be implemented when one of the developers finds he needs them for an application project. I’m afraid they’re not scheduled, to my knowledge.

3

Alright thanks, that is what I wanted to know.

4

I just found out that there actually are LRC-DFT gradient available, just not for DF-DFT. It is mention in the DFT manual part. It is just that the python driver does not recognize them if I ask for e.g. opt(‘wb97’) or gradient(‘wb97’).
But something like below seems to work, and compares ok with the numerical gradient.

molecule water {
O       -1.76914200      -0.07618100       0.00000000
H       -2.06574500       0.83749200       0.00000000
H       -0.80903400       0.00131700       0.00000000
O        1.28389900       0.08888200       0.00000000
H        1.58000600      -0.42576000       0.75636100
H        1.58000600      -0.42576000      -0.75636100
 }

set global {
reference rks
basis cc-pVDZ
df_scf_guess false
scf_type direct
dft_functional wb97
}
gradient('scf')

Could someone confirm if they are working, or if terms are missing? (and therefore missing from the input parser)

As a side note, it might be nice to print out the obtained gradient by default when doing it by finite differences.