[ase-users] Different energy and forces from calculators.turbomole and io.read

[Jens Jørgen Mortensen]

On 07/29/2016 01:27 PM, Kondov, Ivan (SCC) via ase-users wrote:
> Dear all,
>
> i get different energies and forces from the turbomole calculator and from
> the 'gradient' reader method. It can be seen by extending the sample script
>
> https://gitlab.com/ase/ase/blob/master/ase/test/turbomole/turbomole_H2.py
>
> with the lines:
>
> # Run turbomole
> positions = atoms.get_positions()
> energy = atoms.get_potential_energy()
> forces = atoms.get_forces()
> print 'positions\n', positions
> print 'energy\n', energy
> print 'forces\n', forces
>
> # Read gradient file
> from ase.io import read
> cycles = read('gradient',index=':')
> calc = cycles[-1].get_calculator()
> print '\npositions\n', cycles[-1].get_positions()
> print 'energy\n', calc.results['energy']
> print 'forces\n', calc.results['forces']
>
> The output is:
>
> positions
> [[ 0.   0.   0. ]
>   [ 0.   0.   1.1]]
> energy
> -28.2056700793
> forces
> [[-0.         -0.          8.36908323]
>   [-0.         -0.         -8.36908323]]
>
> positions
> [[ 0.   0.   0. ]
>   [ 0.   0.   1.1]]
> energy
> -1.0365388985
> forces
> [[ 0.          0.         -0.08612504]
>   [ 0.          0.          0.08612504]]
>
> How do you think, is this an usage mistake or an ASE bug? I have the feeling
> that the units are not converted correctly (and the forces signs are
> opposite).
>
> Thank you in advance!

Thanks for finding this bug.  Is there a specification somewhere of the 
'gradient' file-format?

If anyone would like to try to fix this then the code can be found here:

     https://gitlab.com/ase/ase/blob/master/ase/io/turbomole.py

Jens Jørgen

> Best regards,
> Ivan
>
>
>
> _______________________________________________
> ase-users mailing list
> ase-users at listserv.fysik.dtu.dk
> https://listserv.fysik.dtu.dk/mailman/listinfo/ase-users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.fysik.dtu.dk/pipermail/ase-users/attachments/20160816/7d6ed241/attachment-0001.html>