[gmx-users] 3.3 release and Forcefield nb.itp format

[Yang Ye]

Hi,

I have found that 3.3's grompp generates drastically different LJ_SR and 
LJ14 parameters for one of my self-defined forcefield (converted from 
AMBER).

 From gmxdump,
(3.3's tpr) tpx version 40
         ntypes=72
            functype[0]=LJ_SR, c6= 2.24468820e-20, c12= 3.18825866e-35
            functype[1]=LJ_SR, c6= 1.61870333e-32, c12= 0.00000000e+00
            functype[2]=LJ_SR, c6= 1.43123863e-20, c12= 1.22953165e-35
            functype[3]=LJ_SR, c6= 1.17580789e-26, c12= 0.00000000e+00
            ...

They should be
(3.3b's tpr) tpx version 32
         ntypes=72
            functype[0]=LJ_SR, c6= 3.35273775e-03, c12= 3.95092275e-06
            functype[1]=LJ_SR, c6= 8.76983977e-05, c12= 8.89523299e-09
            functype[2]=LJ_SR, c6= 3.08322068e-03, c12= 4.16509010e-06
            functype[3]=LJ_SR, c6= 4.73846623e-04, c12= 2.59686658e-07
            ...

My ffambernb.itp is
[ atomtypes ]
;name        mass      charge   ptype            c6           c12
     H     1.008     0.000     A  3.9229141156e-07  5.8568793913e-13
    HO     1.008     0.000     A  0.0000000000e+00  0.0000000000e+00
...
[ nonbond_params ]
; i    j func          c6           c12
     H     H   1  3.9229141156e-07  5.8568793913e-13
     H    HO   1  0.0000000000e+00  0.0000000000e+00
     H    HS   1  3.9229141156e-07  5.8568793913e-13
...
[ nonbond_params ]
; i    j func          c6           c12
     H     H   1  3.9229141156e-07  5.8568793913e-13
     H    HO   1  0.0000000000e+00  0.0000000000e+00
     H    HS   1  3.9229141156e-07  5.8568793913e-13
...

I found that FF provided by gromacs has different column width. Is this 
the reason?

Thanks.

Yang Ye