[gmx-developers] Need help to get the last parts of gromacs/gamess-uk interface to work

[Gerrit Groenhof]

>From the output you need the energy, and gradients on MM and gradients
on QM atoms.

After reading, you need to store these in rvecs QMgrad[],rvec MMgrad[].

Back in calculate_QMMM, these gradient are then used to update the forces.

Hope this helps.
Gerrit


Ake Sandgren wrote:
> Hi!
>
> I'm trying to get the interface between gromacs/gamess-uk written by
> Huub vanDam working and the only remaining part is getting the result
> back properly.
>
> At the moment the interface doesn't agree completely with gamess-uk on
> what to expect.
>
> I simply need help identifying what to read back to which variables in
> gromacs from the various blocks of output gamess-uk generates.
>
> The output from gamess contains these blocks (for the test case i'm
> working with at the moment)
> block = fragment.sequence records = 0
> block = coordinates index =   1 records =  1209 unit = au \
> block = gradients records =    34 unit = au \
> block = bq_gradients records =  1175 unit = au \
> block = update_coordinates index =   2 records =  1209 unit = au \
> block = gradients records =    34 unit = au \
> block = bq_gradients records =  1175 unit = au \
> block = update_coordinates index =   3 records =  1209 unit = au \
> block = gradients records =    34 unit = au \
> block = bq_gradients records =  1175 unit = au \
> block = fragment records = 0
> block = coordinates records =    34 unit = au \
> block = point_charges records =  1175 unit = au
> block = total_energy records = 1
>
> Can someone with more gromacs knowledge help me figuring out exactly
> what to read back to where?
>
> The input from gromacs to gamess-uk contained 34 QM atoms and 1175 MM
> atoms (marked BQ in gamess input)
>
> I have a hunch but I'm not 100% sure...
>
>