Using LibRPA driver with FHI-aims dataset
This tutorial will guide you through the process of setting up and running a calculation using FHI-aims for geometry and SCF computations and LibRPA for RPA energy computation. We will use an H2O molecule as the test case (the setup files can be found here).
1. Prerequisites
Before starting, ensure that you have:
2. FHI-aims Input Files
These files are required to run the initial SCF and geometry optimization using FHI-aims:
control.in
: This file contains control parameters for FHI-aims. Below is an example:xc pbe total_energy_method rpa frequency_points 60 k_grid 1 1 1 output librpa binary
geometry.in
: Contains the geometry of the system. For H2O:lattice_vector 20.00000 0.00000 0.00000 lattice_vector 0.00000 20.00000 0.00000 lattice_vector 0.00000 0.00000 20.00000 atom -0.07712649 0.00000000 1.49704522 O atom 0.82002231 0.00000000 1.86358518 H atom 0.07269418 -0.00000000 0.53972961 H
Run FHI-aims, and after the calculation is completed, the following files will be exported:
stru_out
band_out
KS_eigenvector_<myid>.txt
Cs_data_<myid>.txt
coulomb_mat_<myid>.txt
3. LibRPA Input File
librpa.in
: This file specifies the parameters for the RPA calculation. Example:task = rpa nfreq = 16 binary_input = t use_scalapack_ecrpa = t
4. Running the RPA Calculation with LibRPA
After obtaining the output files from FHI-aims and setting up the parameters in librpa.in
, you can use LibRPA to calculate the RPA correlation energy:
mpirun -np 4 /path/to/LibRPA/build/chi0_main.exe > LibRPA.out
5. LibRPA Output
After the LibRPA calculation is completed, an output file (LibRPA.out
) will be generated, which contains the results of the RPA correlation energy as well as the contributions from each k-point (in Hartree).
RPA correlation energy (Hartree)
| Weighted contribution from each k:
| ( 0, 0, 0): (-0.306259,0)
| Total EcRPA: -0.306258516