Dear All,
Thank you all for your suggestions. I am no longer getting the error after increasing the energy cutoff of the wavefunctions to 250.0 Ry, decreasing the smearing to 0.01, and resetting the diagonalization back to davidson. I have experimented with varying the starting magnetizations and it consistently converges to a non-magnetic state. However, M. Kan et al. showed that a ferromagnetic state should exist via DFT (J. Phys. Chem. Lett., 2013, 4 (20), pp 3382–3386). Is this lack of magnetic states in my investigations more likely a result of not yet discovering a suitable starting magnetization, or could there be another reason such as choice of pseudopotentials? Thanks again, Lance Kavalsky MASc student Dept. of Materials Science and Engineering University of Toronto ________________________________ From: [email protected] <[email protected]> on behalf of Lorenzo Paulatto <[email protected]> Sent: Saturday, September 2, 2017 3:48:11 AM To: [email protected] Subject: Re: [Pw_forum] Problems Computing Cholesky for VC-Relax of MnO2 On 01/09/17 23:04, Lance Kavalsky wrote: nspin=2, lda_plus_u=.TRUE. hubbard_U(1)= 3.9 starting_magnetization(1)=0.5 starting_magnetization(2)=-0.5 I would be that it is the GGA+U correction that is causing problems, not the position of the atoms. Does your calculation converge without the +U? Does it converge wit ha very small +U value? I also think, your smearing is a bit large for the number of k-points you have, this should not be a problem for convergence, but can affect the accuracy of your results. diagonalization = 'cg' Don't use CG, it is slow and does not converge "better" than davidson. Why does everyone use CG diagonalization first thing when they have the slightest problem? (It is a serious question, if it is in some tutorial somewhere that tutorial needs to be corrected!) -- Dr. Lorenzo Paulatto IdR @ IMPMC -- CNRS & Université Paris 6 phone: +33 (0)1 442 79822 / skype: paulatz www: http://www-int.impmc.upmc.fr/~paulatto/ mail: 23-24/423 Boîte courrier 115, 4 place Jussieu 75252 Paris Cédex 05 --> -->
_______________________________________________ Pw_forum mailing list [email protected] http://pwscf.org/mailman/listinfo/pw_forum
