You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The current routine to get the Hubbard parameters from an hp.x run using Voronoi algorithm is not successful when there is a pair of atoms (say TM-O) that are not actually the nearest neighbours. This corner case makes the routine to avoid to output the V interaction between TM-O, since also the other nearest neighbours interaction would appear in the HUBBARD.dat file. This is not what users would expect in general.
The solution would be to add, for instance, an extra flag in the inputs to keep only the interactions indicated at the very beginning, and filter out the parameters parsed from the HUBBARD.dat. This would/should still be an optional feature, so to avoid to confuse new users that would possibly not specify all the interactions that they wanted (e.g. because they are used to the more "traditional" way by hand on QE, where not all the interactions are specified).
The Vonoi algorithm should then be used only on the very first iteration, or not used at all when this feature is required.
The current routine to get the Hubbard parameters from an
hp.x
run using Voronoi algorithm is not successful when there is a pair of atoms (say TM-O) that are not actually the nearest neighbours. This corner case makes the routine to avoid to output the V interaction between TM-O, since also the other nearest neighbours interaction would appear in the HUBBARD.dat file. This is not what users would expect in general.The solution would be to add, for instance, an extra flag in the inputs to keep only the interactions indicated at the very beginning, and filter out the parameters parsed from the HUBBARD.dat. This would/should still be an optional feature, so to avoid to confuse new users that would possibly not specify all the interactions that they wanted (e.g. because they are used to the more "traditional" way by hand on QE, where not all the interactions are specified).
The Vonoi algorithm should then be used only on the very first iteration, or not used at all when this feature is required.
From a discussion with @azadoks
The text was updated successfully, but these errors were encountered: