SUPERCELL CLUSTERS ARE CODE

I have ran the code on a lot of vanadium tellurites structures (32 in total): except for IXASUW and VOLMAL, others' running time range from 15 seconds to 4 mins. I don't understand whether there will be such a big difference on running time if there is none in the fractional coordinates. However, this change didn't seem to result in a shorter running time (I have tested the influence of the change on running time on VOLMAL). I have tried to put the to_be_bonded line and add_bond line under the 'for mol in expansion' loop, hoping that the running time will decrease if we find to_be_bonded on the smaller scale. To_be_bonded of 2*2*2 supercell was: 144 seconds for VOLMAL and 149.72 seconds for IXASUW. I timed each sections of my code and it was the to_be_bonded of 3*3*3 supercell that took the most time: 1796 seconds for VOLMAL and 1766 seconds for IXASUW. I set the upper limit for the atoms_not_of_interest list to be a parameter called 'threshold' and I called dimension_test in the main function so that I could run it twice to calculate max_atoms in 2*2*2 supercell and max_atoms in 3*3*3 supercell. What I did was that I wrote a little function called dimension_test which covers from expansion to finding max atoms using ponents (please see the attached code for more details). I am sorry that I did not make it clear at first. I need the max3/max2 ratio to find out the dimensionality of the structure.

The reason why mine took such a long time was that I not only did the calculation for the 2*2*2 supercell but also did it for the 3*3*3 supercell. I put my extra line on the exact same spot as you.

Supercell clusters are