[gmx-developers] strange results

[Dongsheng Zhang]

Dear GMX developers,

I tried to use gromacs to check my code. At the end of day, I got some
strange results. I calculated the interaction energy between  one atom
in group CCT and twenty atoms (1-20) in group CNT. I made two cases. The
only difference is the charge group number for atom 1 in itp file. I got
different answers for these two cases.

In case 1, part of  cnt.itp file is as following:

[ moleculetype ]
; Name            nrexcl
cnt           3

[ atoms ]
;   nr       type  resnr residue  atom   cgnr     charge       mass
typeB    chargeB      massB
     1       CNTC      1    NTN     C1      1          0     12.011   ;
qtot 0
     2       CNTC      1    NTN     C2      1          0     12.011   ;
qtot 0


In case 2, part of  cnt.itp file is as following:


[ moleculetype ]
; Name            nrexcl
cnt           3

[ atoms ]
;   nr       type  resnr residue  atom   cgnr     charge       mass
typeB    chargeB      massB
     1       CNTC      1    NTN     C1    174          0     12.011   ;
qtot 0
     2       CNTC      1    NTN     C2      1          0     12.011   ;
qtot 0

The index file is as:

[ System ]
1 .....8992

[ CNT ]
   1    2    3    4    5    6    7    8    9   10   11   12   13    
  14   15  16   17   18   19   20 

[ CCT ]
4497 

The energy for case one is 0.000000, for case 2, it is -0.000684. In
fact, none is correct. The energy value in case 2 equals to the
interaction energy between atom 1 in CNT reacting and atom 4497 in CCT.
My question is why the charge group number changes the calculation
result. 

I have prepared all files. If you think it might be a bug, I can submit
all files. The result is reproducible.