Closed
Description
I was playing around with the library and noticed that for a 74 atom molecule with C and H atoms, the generation of grids is very slow.
I am using the following code:
for center_index in range(num_centers):
min_num_angular_points = 86
max_num_angular_points = 302
start=timer()
context = numgrid.new_atom_grid_bse(radial_precision,
min_num_angular_points,
max_num_angular_points,
proton_charges[center_index],
basis_set_name)
duration = timer() - start
print('Duration for context: ',duration)
#num_points = numgrid.get_num_grid_points(context)
start = timer()
# generate an atomic grid in the molecular environment
x, y, z, w = numgrid.get_grid(context,
num_centers,
center_index,
x_coordinates_bohr,
y_coordinates_bohr,
z_coordinates_bohr,
proton_charges)
duration = timer() - start
print('Duration for xyz,w values: ',duration)
The context creation takes a fraction of a second, however, the numgrid.get_grid method takes a lot of time around 7 seconds for radial_precision = 1.0e-10
Is this expected or I could do something better?
For example, I understand that for the same type of atoms, I could just create the context once. But that is hardly the dominant step, so I don't think it makes much of a difference.
If there is nothing wrong with how I am using it, then should I look for parallelization?
Metadata
Metadata
Assignees
Labels
No labels