I’m trying to use a fitfc command.
After running that, I got too many number of perturbations.
For example, from p-<dr>_<er>1 to p-<dr><er>_40.
I think it will take too much time to handle all of those perturbations.
So I tried to reduce the number of perturbations by changing -fr, -er, and -dr.
But I failed to reduce.
Could you please tell me how to reduce them?
Unfortunately, the number of perturbations is determined by the (low) symmetry of your structure.
This is truly the minimum number of symmetrically distinct perturbations your structure has.
However, here is how you can reduce the computational burden.
first only run the "p+" perturbations, the "p-" just add stability to the fit.
use small supercells, i.e. make -er=… small
keep in mind that each calculation is a static one, no relaxations, so it is actually quite fast.
you can reduce the energy cutoffs, k-points etc. because vibrational effects do tend to have a smaller effect of the free energy than the T=0 energy.
once you have many but not all perturbations done, you can try the -sfc=1 or -sfc=2 options, which uses a simpler force constant model that may be determined by fewer perturbations.