If a solute perturbation is defined in MORPH.pert, but the initial and final
non-bonded parameters are identical, and there are no perturbations in bond,
angle or dihedrals, then the gradients should be zero.
This is indeed observed, but only if the atom types do not involve dummy atoms
("du").
julien@node004:methanol-discharge-solvated$ head MORPH.pert
version 1
molecule LIG
atom
name C
initial_type c3
final_type c3
initial_charge 0.11670
final_charge 0.11670
initial_LJ 3.39967 0.10940
final_LJ 3.39967 0.10940
julien@node004:methanol-discharge-solvated$python debug.py 0.500 1
julien@node004:methanol-discharge-solvated$ cat gradients-0.5000.dat
1 -0.0000000000
2 -0.0000000000
3 -0.0000000000
4 -0.0000000000
5 -0.0000000000
6 -0.0000000000
7 -0.0000000000
8 -0.0000000000
9 -0.0000000000
10 -0.0000000000
* Then edit of MORPH.pert for the first atom
julien@node004:methanol-discharge-solvated$ head MORPH.pert
version 1
molecule LIG
atom
name C
initial_type c3
final_type du
initial_charge 0.11670
final_charge 0.11670
initial_LJ 3.39967 0.10940
final_LJ 3.39967 0.10940
julien@node004:methanol-discharge-solvated$python debug.py 0.500 1
julien@node004:methanol-discharge-solvated$ cat gradients-0.5000.dat
1 0.7942053587
2 1.1477001496
3 1.0496873523
4 0.9552297854
5 1.1364659597
6 0.9471439288
7 1.2354505913
8 0.8188560249
9 1.0472695641
10 0.9962898923
Original issue reported on code.google.com by julienm...@googlemail.com on 12 Jul 2013 at 10:35
Original issue reported on code.google.com by
julienm...@googlemail.com
on 12 Jul 2013 at 10:35