Dear all, I am performing my first hybrid functional calculations with PWSCF. I was trying to simulate a quite large system (24 at/unit cell volume=2082.0010 a.u.^3) on a large machine (CINECA-Fermi). The scf appears quite slow so I was trying to test the convergence of the grid density for the Fock operator, to see if I can be any faster without losing too much accuracy. Following the EXX examples I set in the &system section
&system ................ ecutwfc = 70, ecutrho = 280, occupations = 'smearing', degauss = 0.01, input_dft='pbe0', nqx1 = 1, nqx2 = 1, nqx3 = 1, / and for the regular Monkhorst-Pack grid: / K_POINTS (automatic) 3 3 3 1 1 1 In the output file I obtain a Monkhorst-Pack of 42 k-points, which is fine. Anyway I was expecting to get something like EXX: grid of k+q point setup nkqs = 1 since, as far as I have understood, setting all nqxs to 1 amounts to performing a q=0 calculation I get, instead: EXX: grid of k+q point setup nkqs = 42 that is the grid for the Fock operator is as dense as the regular Monkhorst-Pack grid. Does all this correspond to an expected behavior of pwscf? I thank all of you for your attention, Best Regards, Roberto Gaspari, Italian Institute of Technology, Concept Lab
_______________________________________________ Pw_forum mailing list [email protected] http://pwscf.org/mailman/listinfo/pw_forum
