ABINIT Discussion Forums

Dear users/developers, I have a problem in calculating the dielectric tensor of silicon by finite difference of polarization since I get the dielectric tensor depending on the symmetry of k points. I use abinit-8.0.8b. Firstly, I use the following input file. si2_ef.in: ndtset 7 # nsym2 1 nsym3 1 ns...

Dear Jordan Thank you for your reply. I understand how I avoid the error message. It might be better to modify the program mrgddb.x since it could occur that nkpt1=/ nkpt2 & 2*nkpt1=/ nkpt2 when gamma-centered k points are used, where nkpt1 is for q=0 and nkpt2 is for q=/0. By the way, does anyo...

Dear users/developers, I have two questions related to the phonon calculation. I use abinit-7.10.5. (1) Parallel phonon calculation with gamma-centered k points When I run the parallel phonon calculation using 8 cores in order to obtain the phonon dispersion of AlAs crystal, I fail in calculating th...

Dear Mark Torrent Thanks for your reply. I tried to use abinit-7.10.4, however, the result was same. I would like ask you two questions: (1) Are there any versions which are compatible with the electron-positron calculation with paral_kgb=1? In your paer, “Electronic structure investigation of energ...

I calculated the positron lifetime of Al bulk crystal with positron=-1 and ixcpositron=2 as a test of PSN scheme and I found the bad result when setting paral_kgb=1 in both serial and parallel calculations. Although the electronic state calculation was quickly converged, the successive positronic sa...

Dear abinit developers and users, How can I calculate the vibrational modes of a molecule in the response function calculation? I calculated the vibrational modes of a water molecule, however the result was not so good compared to the experimental values, which are 3657, 1595, and 3756 cm^-1. I obta...

When I calculated molecular dynamics of a H2 molecule using the verlet algorithm with ionmov=6 implemented in abinit-7.2.2, I found the sum of the electronic energy and kinetic energy of ions, which were written at the end of each step in a output file, was not preserved. However, when I ignore the ...

In cgwf3.F90, it seems that not only occupied bands but also unoccupied bands, i.e. all bands are orthogonal to the response orbitals. Within density functional perturbation theory, I think only occupied bands should be orthogonal to the response orbitals.

In tddft.F90, rhog is allocated as allocate(rhog(nfftdiel,2),vhartr(nfftdiel)) and forwarded to fourdp and hartre. In these routines, however, rhog is defined as rhog(2,nfft). In tddft.F90, although, that might not affect the calculation because the size of the array does not change. Indeed, I have ...