dear PWSCF users, On Mon, 10 Mar 2003, zyli wrote:
> > It's said that the PWSCF can determine the Hubbard U > parameter in LDA+U calculations automaticaly. But in the > INPUT.PW I also find that I need to give Hubbard_U as a > parameter, why? > Where is it said that PWSCF can determine the Hubbard U parameter AUTOMATICALLY ? It is not true. Please let me know where this is stated so we can correct this error. PWSCF can perform LDA+U calculations using an input value for U specified using the Hubbard_U variables in &system namelist. PWSCF can compute a value for U from the variation of the localized state occupancy due to a localized perturbing potential specified through the Hubbard_alpha variables in &system namelist. Details about the idea are contained in Matteo Cococcioni's PhD thesis and in a paper we are currently writing. Defining automatic this procedure is, to say the least, an exageration. Stefano de Gironcoli
