On Sat, Apr 18, 2020 at 1:02 PM Mladenovic Marko <[email protected]> wrote:
I have seen that the same question was posed several times on the mailing > list but no clear solution was provided. > clear solutions are (typically but not always) provided only to clear problems > Does anyone have an idea how this can be solved? > of course, no. "Segmentation fault" = "something went horribly wrong". Typically, too much memory required, but there may be many other reasons, including code or compiler bugs. > Does it have to do with the QE version, number of nodes...? > Hard to say since you did not provide any such information. In Doc/release-notes there are notes like the following one: Problems fixed in version 6.4 (+ = in qe-6.3-backports as well) : [...] + projwfc.x not working with new xml format in noncolinear/spinorbit case Inputs that I use for PDOS and SCF calculations are enclosed. > the one for scf is incomplete so the problem cannot be reproduced. Even if it was complete, it's a big run. With 1300 bands, a few k-points, and writing to file the overlap matrices (lwrite_overlaps=.true.), you may easily end up using a surprising amount of memory Paolo > > Thanks! > > Marko Mladenovic, PhD > Postdoctoral researcher > École polytechnique fédérale de Lausanne > SB - ISIC - LCBC > web: https://people.epfl.ch/marko.mladenovic > _______________________________________________ > Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso) > users mailing list [email protected] > https://lists.quantum-espresso.org/mailman/listinfo/users -- Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche, Univ. Udine, via delle Scienze 208, 33100 Udine, Italy Phone +39-0432-558216, fax +39-0432-558222
_______________________________________________ Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso) users mailing list [email protected] https://lists.quantum-espresso.org/mailman/listinfo/users
