User Activity

  • Posted a comment on discussion Elk Users on Elk

    OK. We notice that sometimes users overwrite the latest make.inc file with one from an older version. This can cause problems. In your make.inc file you should have the following lines uncommented: LIB_LIBXC = libxcf03.a libxc.a SRC_LIBXC = libxcf03.f90 libxcifc.f90 (Note the f03 convention rather than f90) Is this correct?

  • Posted a comment on discussion Elk Users on Elk

    Could you try: nm libxc.a | grep libxc_free and nm libxcf03.a | grep libxc_free in the elk/src directory and see if libxc_free exists.

  • Posted a comment on discussion Elk Users on Elk

    Hi, I think it's because Ubuntu 26.04 has an old version of Libxc (5.2.3). You'll have to download the latest Libxc source from here and compile it yourself. Regards, Kay.

  • Posted a comment on discussion Elk Developers on Elk

    Dear Miroslav, That's quite easy to do, just use the fractional species block fspecies. Here is an example which finds the optimal unit cell volume of flerovium: tasks 2 ! structural optimisation ! fractional species fspecies -114.0 'Fl' ! high-quality calculation highq .true. ! no conduction band local-orbitals lorbcnd .false. ! use LAPW instead of APW+l.o. nxoapwlo 1 ! optimise the lattice parameters latvopt 1 ! no optimisation of atomic positions atpopt 0 avec 1.0 1.0 0.0 1.0 0.0 1.0 0.0 1.0 1.0...

  • Created a blog post on Elk

    Elk version 11.0.2 released

  • Posted a comment on discussion Elk Users on Elk

    Hi, If reducebf < 1 then bfieldc and bfcmt are automatically reduced to zero for all tasks except for 0,2,28,200,208,350,380,420,421,440. In other words, if reducebf < 1 then you don't need to do anything if you want zero fields (i.e. when they are just for symmetry breaking). If the fields are supposed to be finite (like an external magnetic field) then reducebf should be set to exactly 1 and the RT-TDDFT calculation will include them. Regards, Kay.

  • Posted a comment on discussion Elk Users on Elk

    Hi, You're quite correct: the formula on page 75 is for an earlier version, and also doesn't have the chirp part. I'll fix that for the next release. Thanks and regards, Kay.

  • Posted a comment on discussion Elk Users on Elk

    Hi Chamara, It's best not to enable both nxoapwlo=1 and lorbcnd=.true. as this can lead to instability in the first-variational eigenvalue problem. Also having rgkmax too high can cause similar instabilities. Here's my input file that works for the structure at high compression: tasks 0 xctype 3 vhighq .true. ngridk 18 18 18 stype 0 swidth 0.001 rgkmax 8.5 gmaxvr 32.0 lmaxi 2 nxoapwlo 1 lorbcnd .false. avec 5.058464396264 0.000000000000 0.000000000000 -0.000000000000 5.058464396264 0.000000000000...

View All

Personal Data

Username:
jkdewhurst
Joined:
2008-12-14 20:35:28
Location:
Germany / CEST

Projects

This is a list of open source software projects that J. K. Dewhurst is associated with:

  • Project Logo Elk   Last Updated:

Personal Tools