Computing electrostatic potential

[JunHyeong1]

Hi. Thanks for developing this nice code.

I am wondering whether gpu4pyscf supports computing electrostatic potential generated by atomic orbitals using cartesian grids.

Specifically, I am interested in calculating the following term, which can be computed by pyscf.df.incore.aux_e2.

[wxj6000]

@JunHyeong1 We haven't implemented this function yet. But you can use other routines for this functionality. Here is the script.

import pyscf
import numpy as np
from pyscf import gto, df
from gpu4pyscf.df import int3c2e

atom = '''
O       0.0000000000    -0.0000000000     0.1174000000
H      -0.7570000000    -0.0000000000    -0.4696000000
H       0.7570000000     0.0000000000    -0.4696000000
'''

mol = pyscf.M(atom=atom)
coords = np.random.rand(10,3)

fakemol = gto.fakemol_for_charges(coords)
int3c_gpu = int3c2e.get_int3c2e(mol, fakemol)

int3c_cpu = df.incore.aux_e2(mol, fakemol)

print(np.linalg.norm(int3c_cpu - int3c_gpu.get()))

[JunHyeong1]

@wxj6000 Thank you very much!