Can anyone illustrate the outputs for a IMF and SCP phonon calculations.

[Arpan Kundu]

Hello everybody,

I am interested in using independent mode framework (IMF) and self-consistent phonons (SCP) implemented in i-PI. I ran the example of a simple water molecule. I observed that in both cases there are a plethora of output files. I am not interested to calculate free energies which are being printed in i-PI output for an IMF calculation, while for an SCP calculation a separate post-processing script can calculate it for you. I am working on a code that would calculate different properties from these calculations; however, for that purpose, I need to understand the format of those outputs.

For an IMF calculation, I guess the .vfit files contains the interpolated potential data where 1st column is the displacements while the 2nd column is the potential energy. Please correct me if I am wrong. What information is there in the .qvf files. I guess the first column is the actual displacement used while sampling that normal mode, 2nd column is the energy relative to the bottom of the potential well; please correct me if I am wrong, but I cannot guess the third column.

Coming to the SCP calculations; there are a bunch of files with extensions .V0.mode_no, .x.mode_no etc. Can anyone please tell me what information is these files containing and which columns/rows indicate what? 

Thank you very much.

Arpan

[E A Engel]

Hello Arpan,

Regarding the IMF output files:

The .vfit files indeed contain displacements in normal mode space and associated potential energies (after interpolation). Meanwhile the .qvf files contain displacements in normal mode space, associated potential energies, and forces again in the normal mode space, i.e. projected onto the normal mode axes. I hope this helps!

Best wishes,

Edgar

[venkat kapil]

Regarding SCP:

*.V0.${ITER} corresponds to the minimum potential energy of the harmonic trial Hamiltonian at the ${ITER}-th step. $.x.${ITER} contains the positions that were sampled from the quantum canonical Boltzmann distribution of the trail Hamiltonian. Each row contained 3N entries corresponding to the total position of a displaced configuration.

[Yair Litman]

Hi Edgar and Venkat,

Do you think it is possible to add those descriptions in a README file that is generated with the simulations (or elsewhere within the code)?

Best,
Yair

PS: Sorry, if it is already there, but I couldn't find it.