Using existing JK object in psi4numpy

1

Hi all,

I’ve got a python code that first runs a self-consistent DFT energy calculation with a density fitted range-separated hybrid, and then reoptimizes the excited state orbitals. The reoptimization uses a JK object based on the basis set from the original calculation.

I would like to use the JK object from the initial calculation for the later calculations to avoid doubling a large memory object. Despite looking everwhere, and despite being fairly confident it must be available, I could not seem to find out how to do this.

Please help!

tl;dr - I have a DFT wfn and would like to access its existing JK object via python.

All the best,

Tim

2

Hi,

You should be able to access the JK object from the wavefunction if the save_jk keyword is set to true:

psi4.set_options({'save_jk': True})
e, wfn = psi4.energy('b3lyp/6-31g', return_wfn=True)
jk = wfn.jk()
3

Thank you! But, unfortunately when I first go to use the jk object I get a “forrtl: severe (174): SIGSEGV, segmentation fault occurred” error. Any idea what is going wrong?

molecule X {
0 1
Be
}

psi4.set_options({'save_jk': True})
e, wfn = psi4.energy('hse06/6-31g', return_wfn=True)
jk = wfn.jk()

C = wfn.Ca_subset("MO", "OCC")
jk.C_left_add(C)
jk.C_right_add(C)
jk.compute() # This seems to be the line that fails
print(jk.J()[0])