ABINIT Discussion Forums

Hello,

Could you help me to understand why ETOT is jumping so much in phonon calculations? I just run trf2_x. example and in the first run
ETOT ~ -9.7 Ha but then it jumps to large positive values for different q-vectors.

Thank you,

Jan Gryko

Because the total energy (ETOT) has different meanings in the different calculations. In the first run, ETOT is the total ground state energy. In the phonon runs, what is being calculated (and represented by ETOT) is the second derivative of the energy with respect to the perturbation being done to the system. Thus there's no reason they should be similar in these two very different contexts.

Thank you very much for quick answer. Perfectly clear.

JG

In phonon calculations, the ETOT quantity (which is, as pointed out by Prof. Zwanziger, is the second derivative with respect to the perturbation) converges very quickly. But the square of density/potential residues converges much more slowly. Question: Is it OK to get convergence in ETOT only, since what we need are second derivatives? If so, can I use toldfe instead of tolvrs?

Thank you,
JG

Well, I guess you could try it this way but it is not recommended--much safer to use tolvrs with moderate convergence, say 1D-8 or 1D-10.