<?xml version="1.0" encoding="utf-8"?>
<feed xml:lang="en" xmlns="http://www.w3.org/2005/Atom"><title>Activity for Elk</title><link href="https://sourceforge.net/p/elk/activity/" rel="alternate"/><id>https://sourceforge.net/p/elk/activity/</id><updated>2026-11-08T00:49:25.433000Z</updated><subtitle>Recent activity for Elk</subtitle><entry><title>Long Zhang modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3" rel="alternate"/><published>2026-11-08T00:49:25.433000Z</published><updated>2026-11-08T00:49:25.433000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3</id><summary type="html">Hi Kay, I understand the columns of evecsv(nstsv,nstsv) are used to do linear combination of 1st-variational states. By " first nstfv orbitals are spin-up and the next nstfv orbitals are spin-down", do you mean the columns are ordered so that the first nstfv columns are spin-up states and second nstfv columns are spin-down? Best Regards, Long</summary></entry><entry><title>Lars Nordström posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#ba50" rel="alternate"/><published>2026-11-05T20:47:58.974000Z</published><updated>2026-11-05T20:47:58.974000Z</updated><author><name>Lars Nordström</name><uri>https://sourceforge.net/u/larsnordstrom/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#ba50</id><summary type="html">Collinear magnetism has two meanings. Without SOC there is no ambiguity. Collinear order is of course also a possible magnetic order with SOC, in the meaning that the local magnetic moments are collinear, but it does not mean that the magnetic spin density is collinear on a microscopic level because it cannot be. The flags are for the magnetization desities not the order. I never force the magnetization density to be collinear in presence of SOC and assume it is not common practise. best wishes ...</summary></entry><entry><title>Long Zhang posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#2bd9" rel="alternate"/><published>2026-11-05T16:15:52.894000Z</published><updated>2026-11-05T16:15:52.894000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#2bd9</id><summary type="html">And, it looks collinearity (ncmag=false) is less likely to happen, for spin polarized + SOC. Following the logic in init0, spinpol=true And spinorb=true will make ndmag=3 (thus ncmag=true). ndmag=3 can also be triggered if the quantities bfieldc0 or bfcmt0 become larger than threshhold value. ndmag=1 only happens later in the code through the condition cmagz=true. But cmagz does not default to be true. So, one needs to set cmagz=true in input file. Is forcing collinearity in spin polarized + SOC...</summary></entry><entry><title>Long Zhang modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3" rel="alternate"/><published>2026-11-05T16:13:09.899000Z</published><updated>2026-11-05T16:13:09.899000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3</id><summary type="html">Hi Kay, I understand the columns of evecsv(nstsv,nstsv) are used to do linear combination of 1st-variational states. evecsv( 1:nstfv, j ) yields the up spinor component of state-j. evecsv( nstfv+1:nstsv, j ) yields the down spinor component of state-j. By " first nstfv orbitals are spin-up and the next nstfv orbitals are spin-down", do you mean the columns are ordered so that the first nstfv columns are spin-up states and second nstfv columns are spin-down? Best Regards, Long</summary></entry><entry><title>Long Zhang modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3" rel="alternate"/><published>2026-11-05T09:52:56.012000Z</published><updated>2026-11-05T09:52:56.012000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3</id><summary type="html">Hi Kay, It seems collinearity (ncmag=false) is less likely to happen, for spin polarized and SOC. Following the logic of the code in init0, First, spinpol=true And spinorb=true will make ndmag=3 (thus ncmag=true). ndmag=3 can also be triggered if the quantities bfieldc0 or bfcmt0 become larger than some threshhold value. ndmag=1 only happens later in the code through the condition cmagz=true. But cmagz does not default to be true. So, one needs to set cmagz=true in input file, to have collinear spin...</summary></entry><entry><title>Long Zhang modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3" rel="alternate"/><published>2026-11-05T06:29:22.881000Z</published><updated>2026-11-05T06:29:22.881000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3</id><summary type="html">Hi Kay, It seems collinearity (ncmag=false) is less likely to happen, for spin polarized and SOC. Following the logic of the code in init0, First, spinpol=true And spinorb=true will make ndmag=3 (thus ncmag=true). ndmag=3 can also be triggered if the quantities bfieldc0 or bfcmt0 become larger than some threshhold value. ndmag=1 only happens later in the code through the condition cmagz=true. But cmagz does not default to be true. So, one needs to set cmagz=true in input file, to have collinear spin...</summary></entry><entry><title>Long Zhang modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3" rel="alternate"/><published>2026-11-05T04:11:05.255000Z</published><updated>2026-11-05T04:11:05.255000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3</id><summary type="html">Hi Kay, It seems collinearity (ncmag=false) is less likely to happen, for spin polarized and SOC. Following the logic of the code in init0, First, spinpol=true And spinorb=true will make ndmag=3 (thus ncmag=true). ndmag=3 can also be triggered if the quantities bfieldc0 or bfcmt0 become larger than some threshhold value. ndmag=1 only happens later in the code through the condition cmagz=true. But cmagz does not default to be true. So, one needs to set cmagz=true in input file, to have collinear spin...</summary></entry><entry><title>Long Zhang modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3" rel="alternate"/><published>2026-11-05T02:34:16.470000Z</published><updated>2026-11-05T02:34:16.470000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3</id><summary type="html">Hi Kay, It seems collinearity (ncmag=false) is less likely to happen, for spin polarized and SOC. Following the logic of the code in init0, First, spinpol=true And spinorb=true will make ndmag=3 (thus ncmag=true). ndmag=3 can also be triggered if the quantities bfieldc0 or bfcmt0 become larger than some threshhold value. ndmag=1 only happens later in the code through the condition cmagz=true. But cmagz does not default to be true. So, I'll need to set cmagz=true in input file, to have collinear spin...</summary></entry><entry><title>Long Zhang modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3" rel="alternate"/><published>2026-11-05T02:28:16.967000Z</published><updated>2026-11-05T02:28:16.967000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3</id><summary type="html">Hi Kay, It seems collinearity (ncmag=false) is less likely to happen, for spin polarized and SOC. Following the logic of the code in init0, First, spinpol=true And spinorb=true will make ndmag=3 (thus ncmag=true). ndmag=3 can also be triggered if the quantities bfieldc0 or bfcmt0 become larger than some threshhold value. ndmag=1 only happens later in the code through the condition cmagz=true. But cmagz does not default to be true. So, I'll need to set cmagz=true in input file, to have collinear spin...</summary></entry><entry><title>Long Zhang modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3" rel="alternate"/><published>2026-11-05T02:08:23.398000Z</published><updated>2026-11-05T02:08:23.398000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3</id><summary type="html">Hi Kay, It seems collinearity (ncmag=false) is less likely to happen, for spin polarized and SOC. Following the logic of the code in init0, First, spinpol=true And spinorb=true will make ndmag=3 (thus ncmag=true). ndmag=3 can also be triggered if the quantities bfieldc0 or bfcmt0 become larger than some threshhold value. ndmag=1 only happens later in the code through the condition cmagz=true. But cmagz does not default to be true. So, I'll need to set cmagz=true in input file, to have collinear spin...</summary></entry><entry><title>Long Zhang modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3" rel="alternate"/><published>2026-11-05T01:46:13.409000Z</published><updated>2026-11-05T01:46:13.409000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3</id><summary type="html">Hi Kay, It seems collinearity (boolean ncmag=false) is less likely to happen for spin polarized SOC calculation. Following the logic of the code in init0, First, spinpol=true AND spinorb=true will make ndmag=3 (thus ncmag=true). ndmag=3 will also be triggered if the quantities bfieldc0 or bfcmt0 become larger than some threshhold value. ndmag=1 can happen later in the code only through the condition cmagz=true. But cmagz does not default to be true. So, I'll need to set cmagz=true in input file,...</summary></entry><entry><title>Long Zhang posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3" rel="alternate"/><published>2026-11-05T01:42:52.985000Z</published><updated>2026-11-05T01:42:52.985000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3</id><summary type="html">Hi Kay, It seems collinearity (boolean ncmag=false) is less likely to happen for spin polarized SOC calculation. Following the logic of the code in init0, First, spinpol=true AND spinorb=true will make ndmag=3 (thus ncmag=true). ndmag=3 will also be triggered if the quantities bfieldc0 or bfcmt0 become larger than some threshhold value. ndmag=1 can happen later in the code only through the condition cmagz=true. But cmagz does not default to be true. So, I'll need to set cmagz=true in input file,...</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=50#8c9d" rel="alternate"/><published>2026-11-04T08:03:34.365000Z</published><updated>2026-11-04T08:03:34.365000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=50#8c9d</id><summary type="html">Hi Long and Miroslav, Just to clarify: for collinear calculations the first nstfv orbitals are spin-up and the next nstfv orbitals are spin-down. For non-collinear calculations the orbitals are not separable into pure up and down but rather general 2-spinor functions. Regards, Kay.</summary></entry><entry><title>Long Zhang modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#4684" rel="alternate"/><published>2026-11-04T04:42:44.103000Z</published><updated>2026-11-04T04:42:44.103000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#4684</id><summary type="html">Hi Miroslav, Thank you for the quick response. This is very helpful! I did think that one spinor wavefunction cannot be exclusively tagged as spin-up or down. AI analyzer confirmed this point. I probably have to think again what can be done and adjust what I want. The thing that confuses me is actually the connection with band energies: The two spinor components belong to one wavefunction, while I assume one spinor wavefunction corresponds to one band energy. But are the band energies distinctly...</summary></entry><entry><title>Long Zhang modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#4684" rel="alternate"/><published>2026-11-04T04:22:18.061000Z</published><updated>2026-11-04T04:22:18.061000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#4684</id><summary type="html">Hi Miroslav, Thank you for the quick response! I did think that one spinor wavefunction cannot be exclusively tagged as spin-up or down. AI analyzer just confirmed this point. But the thing that confuses me is: The two spinor components belong to one wavefunction, while I assume one spinor wavefunction corresponds to one band energy. But are the band energies distinctly either spin-up or spin-down? This is the place doesn't connect in my mind. Best, Long</summary></entry><entry><title>Long Zhang modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#4684" rel="alternate"/><published>2026-11-04T04:18:49.998000Z</published><updated>2026-11-04T04:18:49.998000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#4684</id><summary type="html">Hi Miroslav, Thank you for the quick response! I did think that one spinor wavefunction cannot be exclusively tagged as spin-up or down. AI analyzer just confirmed this point. But the thing that confuses me is: The two spinor components belong to one wavefunction, while I assume one 2nd-variational wavefunction corresponds to one band energy. But are the band energies distinctly either spin-up or spin-down? This is the place doesn't connect in my mind. Best, Long</summary></entry><entry><title>Long Zhang posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#4684" rel="alternate"/><published>2026-11-04T04:05:37.006000Z</published><updated>2026-11-04T04:05:37.006000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#4684</id><summary type="html">Hi Miroslav, Thanks for the quick response! I did think that one spinor wavefunction cannot be exclusively tagged as spin-up or down. AI analyzer just confirmed this point. So, this means I need to adjust what I want to do and what I can do ...... Best, Long</summary></entry><entry><title>Miroslav Iliaš posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#08e7" rel="alternate"/><published>2026-11-04T03:25:17.896000Z</published><updated>2026-11-04T03:25:17.896000Z</updated><author><name>Miroslav Iliaš</name><uri>https://sourceforge.net/u/miroilias/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#08e7</id><summary type="html">Hello Long, I uploaded the fresh ELk code to chatgpt AI analyzer, together with your question, and got this answers , https://chatgpt.com/share/6aeaa4a5-42e0-83eb-8fe9-0e462c0cafa2 . Just continue the code analysis, there is a chance you get clue concerning which of the selected wavefunctions correspond to spin-up states, and which correspond to spin-down states.</summary></entry><entry><title>Long Zhang modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#bd41" rel="alternate"/><published>2026-11-04T03:17:13.127000Z</published><updated>2026-11-04T03:17:13.127000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#bd41</id><summary type="html">Dear ELK users and developers, I am asking for general instructions/suggestions for a particular purpose of mine. This is not about any keyword or usage of ELK. I appreciate any advice and suggestion. I want to extract some of the ground state muffin-tin wavefunctions (e.g. within an energy window or a range of band indices) from a spin polarized and SOC-enabled calculation, for developing and testing a post processing method. I will have to identify which of the selected wavefunctions correspond...</summary></entry><entry><title>Long Zhang modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#bd41" rel="alternate"/><published>2026-11-04T03:15:12.809000Z</published><updated>2026-11-04T03:15:12.809000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#bd41</id><summary type="html">Dear ELK users and developers, I am asking for general instructions/suggestions for a particular purpose of mine. This is not about any keyword or usage of ELK. I appreciate any advice and suggestion. I want to extract some of the ground state muffin-tin wavefunctions (e.g. within an energy window or a range of band indices) from a spin polarized and SOC-enabled calculation, for developing and testing a post processing method. I will have to identify which of the selected wavefunctions correspond...</summary></entry><entry><title>Long Zhang posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#bd41" rel="alternate"/><published>2026-11-04T02:48:54.698000Z</published><updated>2026-11-04T02:48:54.698000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#bd41</id><summary type="html">Dear ELK users and developers, I am asking for general instructions/suggestions for a particular purpose of mine. This is not about any keyword or usage of ELK. I appreciate any advice and suggestion. I want to extract some of the ground state muffin-tin wavefunctions (e.g. within an energy window or a range of band indices) from a spin polarized and SOC-enabled calculation, for developing and testing a post processing method. I will have to identify which of the selected wavefunctions correspond...</summary></entry><entry><title>Miroslav Iliaš posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/3f7447da52/?limit=25#eaba" rel="alternate"/><published>2026-10-18T07:33:32.588000Z</published><updated>2026-10-18T07:33:32.588000Z</updated><author><name>Miroslav Iliaš</name><uri>https://sourceforge.net/u/miroilias/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/3f7447da52/?limit=25#eaba</id><summary type="html">Thanks for the reply. Yes, we want to study superheavy elements (up to Z=120,but also higher Z) with periodic DFT codes, and the all-electron Elk seems to be the good computational tool. The f90 files you provided are from the latest Elk 11.2.3, right ? Would be good to have more verbose output when generating the species files for new SHEs.</summary></entry><entry><title>Peter Elliott posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/3f7447da52/?limit=25#a21e" rel="alternate"/><published>2026-10-18T00:18:28.523000Z</published><updated>2026-10-18T00:18:28.523000Z</updated><author><name>Peter Elliott</name><uri>https://sourceforge.net/u/elliottp/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/3f7447da52/?limit=25#a21e</id><summary type="html">From your output it doesn't look like it crashed, it just gave a warning. This warning only comes from the first pass, which uses a bad guess for the occupations, but after that it finds the right occupations and behaves as it should. So the generated species file is more or less fine. There is a subtlety that it since it didn't occupy the n=7 shell in the first pass, it fixes the radial mesh smaller than it ideally would. Probably it would be safer to use the attached genfspecies.f90 file, that...</summary></entry><entry><title>Miroslav Iliaš posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/3f7447da52/?limit=25#b663" rel="alternate"/><published>2026-10-17T21:44:34.063000Z</published><updated>2026-10-17T21:44:34.063000Z</updated><author><name>Miroslav Iliaš</name><uri>https://sourceforge.net/u/miroilias/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/3f7447da52/?limit=25#b663</id><summary type="html">Greetings, dear Elk developers, I am trying to generate the Cn (Z=112) species file, but it crashes - see below. Input, output files are attached. Elk is installed within the conda channel. Best, Miro (elk_env) miroi@MIRO :~/work/projects/open-collection/theoretical_chemistry/software/elk-ase/runs/Cn/simple_run/generate_Cn_species_file/.run_elk.bash ┌─────────────────────────────────┐ │ Elk code version 10.2.4 started │ └─────────────────────────────────┘ Warning(atom): maximum iterations exceeded...</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/73e0f879c3/?limit=50#4744" rel="alternate"/><published>2026-10-14T12:28:29.308000Z</published><updated>2026-10-14T12:28:29.308000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/73e0f879c3/?limit=50#4744</id><summary type="html">Hi, The GW calculation in Elk is one-shot G0W0 in the Matsubara frequency representation. It is very important to converge all the relevant parameters for the GW calculation. Here is an example for Ni which is properly converged: tasks 0 10 600 610 630 spinpol .true. bfieldc 0.0 0.0 0.1 reducebf 0.85 highq .true. ! Matsubara frequency cut-off wmaxgw 15.0 ! response function G-vector cut-off gmaxrf 3.5 ! temperature of many-body system (determines Matsubara frequency spacing) tempk 1000.0 wplot 3000...</summary></entry><entry><title>Marks posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/73e0f879c3/?limit=25#e4d8" rel="alternate"/><published>2026-10-12T09:03:43.546000Z</published><updated>2026-10-12T09:03:43.546000Z</updated><author><name>Marks</name><uri>https://sourceforge.net/u/marks28/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/73e0f879c3/?limit=25#e4d8</id><summary type="html">Dear Elk developers, I am interested in performing a GW calculation using the Elk code, and I would like to clarify how the GW implementation works and how to perform a one-shot (G0W0) calculation. I have been going through the Elk user guide and found the following GW-related tasks: 600 - Output the GW self-energy matrix elements. 610 - Generates the GW spectral function. 620 - Generates the GW band structure, i.e. the k-dependent spectral function. 630 - Writes the GW Fermi energy to GWEFERMI.OUT....</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/c423eed840/?limit=50#fc80" rel="alternate"/><published>2026-09-14T12:47:58.515000Z</published><updated>2026-09-14T12:47:58.515000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/c423eed840/?limit=50#fc80</id><summary type="html">Further update: I've contacted Susi Lehtola and he requests that you file a ticket on the Libxc Gitlab tracker with a reproducer.</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/c423eed840/?limit=50#17f4" rel="alternate"/><published>2026-09-14T11:23:08.029000Z</published><updated>2026-09-14T11:23:08.029000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/c423eed840/?limit=50#17f4</id><summary type="html">Update: I've now reproduced the issue by using Libxc version 7.1.2. This version does not fail gracefully with negative sigma for r2SCAN like version 7.0.0. This should be fixed in Libxc but we may have make a work-around in Elk for the time being. Thanks and regards, Kay.</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/c423eed840/?limit=50#3135" rel="alternate"/><published>2026-09-14T09:35:12.724000Z</published><updated>2026-09-14T09:35:12.724000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/c423eed840/?limit=50#3135</id><summary type="html">Hi Ricardo, I ran your example and it works fine for me. I've attached my INFO.OUT file. Perhaps your stack size was insufficient. Try using the script elk/elk.sh to run the code: #!/bin/bash export OMP_NUM_THREADS=32 export OMP_PROC_BIND=false export OMP_STACKSIZE=512M ulimit -Ss 524288 ~/elk/src/elk Best wishes, Kay.</summary></entry><entry><title>Ricardo Ribeiro posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/c423eed840/?limit=25#5216" rel="alternate"/><published>2026-09-14T09:21:41.499000Z</published><updated>2026-09-14T09:21:41.499000Z</updated><author><name>Ricardo Ribeiro</name><uri>https://sourceforge.net/u/ricardoribeiro/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/c423eed840/?limit=25#5216</id><summary type="html">Elk version: 11.2.3 (stock, unmodified) Libxc version: 7.1.2 Platform: Linux x86-64, OpenMPI 5 + OpenMP 1. Summary With any Libxc meta-GGA correlation functional, a symmetrised collinear antiferromagnet whose lattice vectors are orthogonal produces NaN in the exchange-correlation potential from SCF loop 1, and hence a NaN total energy. The cause is a catastrophic cancellation in the total-density gradient invariant that Libxc reconstructs from the three spin-resolved invariants Elk passes to it....</summary></entry><entry><title>J. K. Dewhurst created a blog post</title><link href="https://sourceforge.net/p/elk/news/2026/08/elk-version-1123-released/" rel="alternate"/><published>2026-08-28T10:29:48.406000Z</published><updated>2026-08-28T10:29:48.406000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/news/2026/08/elk-version-1123-released/</id><summary type="html">Elk version 11.2.3 released</summary></entry><entry><title>Elk released /elk-11.2.3.tgz</title><link href="https://sourceforge.net/projects/elk/files/elk-11.2.3.tgz/download" rel="alternate"/><published>2026-08-28T10:27:02.851000Z</published><updated>2026-08-28T10:27:02.851000Z</updated><author><name>Elk</name><uri>https://sourceforge.net/p/elk/profile/</uri></author><id>https://sourceforge.net/projects/elk/files/elk-11.2.3.tgz/download</id><summary type="html"/></entry><entry><title>Long Yu modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6443bf6c/?limit=25#ac58/2cc9" rel="alternate"/><published>2026-08-21T15:34:58.763000Z</published><updated>2026-08-21T15:34:58.763000Z</updated><author><name>Long Yu</name><uri>https://sourceforge.net/u/chichenyulong/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6443bf6c/?limit=25#ac58/2cc9</id><summary type="html">Dear Kay, A more sophisitcated way would be to use TDDFT with a constantly increasing vector potential, E=-1/c dA(t)/dt. Has this function been implemented now? Thank you. Regards, Long</summary></entry><entry><title>Long Yu posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6443bf6c/?limit=25#ac58/2cc9" rel="alternate"/><published>2026-08-21T15:33:49.468000Z</published><updated>2026-08-21T15:33:49.468000Z</updated><author><name>Long Yu</name><uri>https://sourceforge.net/u/chichenyulong/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/6443bf6c/?limit=25#ac58/2cc9</id><summary type="html">Dear Kay, A more sophisitcated way would be to use TDDFT with a constantly increasing vector potential, E=-1/c dA(t)/dt. Has this function been implemented now? Regards, Long</summary></entry><entry><title>Zixuan Li posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/f03c738c3b/?limit=25#9518" rel="alternate"/><published>2026-08-15T01:36:20.596000Z</published><updated>2026-08-15T01:36:20.596000Z</updated><author><name>Zixuan Li</name><uri>https://sourceforge.net/u/zxli-phy/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/f03c738c3b/?limit=25#9518</id><summary type="html">Dear all I've tested the SCF and TDDFT successive two-step calculation with or without setting the reducek tag in the SCF part and the results of them are different. Then I've checked the INFO.OUT produced by the two calculation, it shows that the kpoint reduction seems to be closed automatically in the TDDFT process regardless the reducek tag is set or not. So, reasonably, I prefer that setting 'reducek = 0' in the SCF part may be the right one. By the way, I've search the forum, and discovery that...</summary></entry><entry><title>Jerry Li modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/7723f2ebc9/?limit=25#a6a0" rel="alternate"/><published>2026-08-09T04:21:00.117000Z</published><updated>2026-08-09T04:21:00.117000Z</updated><author><name>Jerry Li</name><uri>https://sourceforge.net/u/jerryxinjie/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/7723f2ebc9/?limit=25#a6a0</id><summary type="html">Dear All, Cupric oxide (CuO) is a p-type semiconductor with an indirect band gap reported to be in the range of 1.2–1.9 eV. It has strong electron correlations, and antiferromagnetic ground state. I have tried both DFT+U and hybrid-functional approaches, but neither method is able to reproduce the experimental band gap correctly, as shown in the attached screenshot. The ELK input files used for these calculations are also attached for reference. I would greatly appreciate any suggestions or insights...</summary></entry><entry><title>Jerry Li modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/7723f2ebc9/?limit=25#a6a0" rel="alternate"/><published>2026-08-09T03:52:29.137000Z</published><updated>2026-08-09T03:52:29.137000Z</updated><author><name>Jerry Li</name><uri>https://sourceforge.net/u/jerryxinjie/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/7723f2ebc9/?limit=25#a6a0</id><summary type="html">Dear All, Cupric oxide (CuO) is a p-type semiconductor with an indirect band gap reported to be in the range of 1.2–1.9 eV. I have tried both DFT+U and hybrid-functional approaches, but neither method is able to reproduce the experimental band gap correctly, as shown in the attached screenshot. The ELK input files used for these calculations are also attached for reference. I would greatly appreciate any suggestions or insights regarding the possible reasons for this discrepancy and recommendations...</summary></entry><entry><title>Jerry Li posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/7723f2ebc9/?limit=25#a6a0" rel="alternate"/><published>2026-08-09T03:50:41.078000Z</published><updated>2026-08-09T03:50:41.078000Z</updated><author><name>Jerry Li</name><uri>https://sourceforge.net/u/jerryxinjie/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/7723f2ebc9/?limit=25#a6a0</id><summary type="html">Dear All, Cupric oxide (CuO) is a p-type semiconductor with an indirect band gap reported to be in the range of 1.2–1.9 eV. I have tried both DFT+U and hybrid-functional approaches, but neither method is able to reproduce the experimental band gap correctly, as shown in the attached screenshot. The ELK input files used for these calculations are also attached for reference. I would greatly appreciate any suggestions or insights regarding the possible reasons for this discrepancy and recommendations...</summary></entry><entry><title>marcindulak modified a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/9fe1485f0d/?limit=25#882c" rel="alternate"/><published>2026-07-19T05:20:26.932000Z</published><updated>2026-07-19T05:20:26.932000Z</updated><author><name>marcindulak</name><uri>https://sourceforge.net/u/marcindulak/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/9fe1485f0d/?limit=25#882c</id><summary type="html">I'm sorry for the noise. I noticed my elk was not latest. elk-11.0.2 with LIB_LIBXC = -L%{_libdir} -lxcf03 -lxc and SRC_LIBXC = libxcifc.f90 builds with f03 libxc-7.1.0.</summary></entry><entry><title>marcindulak posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/9fe1485f0d/?limit=25#882c" rel="alternate"/><published>2026-07-19T05:05:29.972000Z</published><updated>2026-07-19T05:05:29.972000Z</updated><author><name>marcindulak</name><uri>https://sourceforge.net/u/marcindulak/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/9fe1485f0d/?limit=25#882c</id><summary type="html">I'm sorry for the noise. I noticed my elk was not latest. elk-11.0.2 with LIB_LIBXC = -L%{_libdir} -lxc and SRC_LIBXC = libxcifc.f90 builds with f03 libxc-7.1.0.</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/9fe1485f0d/?limit=50#16b1" rel="alternate"/><published>2026-07-18T19:42:16.336000Z</published><updated>2026-07-18T19:42:16.336000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/9fe1485f0d/?limit=50#16b1</id><summary type="html">The f03 interface.</summary></entry><entry><title>marcindulak posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/9fe1485f0d/?limit=25#8fc6" rel="alternate"/><published>2026-07-18T19:36:46.045000Z</published><updated>2026-07-18T19:36:46.045000Z</updated><author><name>marcindulak</name><uri>https://sourceforge.net/u/marcindulak/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/9fe1485f0d/?limit=25#8fc6</id><summary type="html">OK, so it's related to packaging libxc, as hinted above. I don't find libxc f90 modules packaged on Fedora. Are you using f90 or f03 interface? rpm -ql libxc-devel | grep modules /usr/lib64/gfortran/modules/xc_f03_funcs_m.mod /usr/lib64/gfortran/modules/xc_f03_lib_m.mod</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/9fe1485f0d/?limit=50#1596" rel="alternate"/><published>2026-07-18T19:14:16.395000Z</published><updated>2026-07-18T19:14:16.395000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/9fe1485f0d/?limit=50#1596</id><summary type="html">Hi Marcin, Libxc 7.0.0 has worked fine with Elk for over a year now (since version 10.5.16). Regards, Kay.</summary></entry><entry><title>marcindulak posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/9fe1485f0d/?limit=25#86da" rel="alternate"/><published>2026-07-18T18:32:34.300000Z</published><updated>2026-07-18T18:32:34.300000Z</updated><author><name>marcindulak</name><uri>https://sourceforge.net/u/marcindulak/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/9fe1485f0d/?limit=25#86da</id><summary type="html">Hi, What's the status of the libxc7 support in elk? I see there is a new https://gitlab.com/libxc/libxc/-/tags/7.1.0, but I see no mention of xc_f03 in the elk source: grep xc_f03 -r elk-10.1.12 The patch to switch to f03 https://src.fedoraproject.org/rpms/elk/blob/962f050def3391e3b2ba943d9148f262afc64736/f/elk-9.2.12-libxc7.patch is rather large, and does not apply anymore.</summary></entry><entry><title>CY ZHUNG modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=25&amp;page=1#177b" rel="alternate"/><published>2026-07-07T04:07:20.576000Z</published><updated>2026-07-07T04:07:20.576000Z</updated><author><name>CY ZHUNG</name><uri>https://sourceforge.net/u/a-z/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=25&amp;page=1#177b</id><summary type="html">Yes. but that is had done. i do learn to use three DFT tool suits like QE,CASTEP. Read the make.inc and do some modification is an easy work. Any more suggestion?</summary></entry><entry><title>CY ZHUNG posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=25&amp;page=1#177b" rel="alternate"/><published>2026-07-06T20:48:35.928000Z</published><updated>2026-07-06T20:48:35.928000Z</updated><author><name>CY ZHUNG</name><uri>https://sourceforge.net/u/a-z/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=25&amp;page=1#177b</id><summary type="html">Yes. but that is had done. i do learn to use three DFT tools suit like QE,CASTEP. Read the make.inc and do some modification is an easy work. Any more suggestion?</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=50#1701" rel="alternate"/><published>2026-07-06T08:44:36.425000Z</published><updated>2026-07-06T08:44:36.425000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=50#1701</id><summary type="html">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?</summary></entry><entry><title>CY ZHUNG modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=25&amp;page=1#c898" rel="alternate"/><published>2026-07-06T01:03:15.418000Z</published><updated>2026-07-06T01:03:15.418000Z</updated><author><name>CY ZHUNG</name><uri>https://sourceforge.net/u/a-z/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=25&amp;page=1#c898</id><summary type="html">"nm libxc.a | grep libxc_free" gets the flags but "nm libxcf03.a | grep libxc_free" gets nothing on Ubuntu 26.04. U libxc_free_flags U libxc_free_flags 0000000000000690 T libxc_free_flags The flags are almost the same on my archlinux laptop except the last digtal number in "0000……*** T libxc_free_flages"</summary></entry><entry><title>CY ZHUNG posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=25&amp;page=1#c898" rel="alternate"/><published>2026-07-06T01:02:55.874000Z</published><updated>2026-07-06T01:02:55.874000Z</updated><author><name>CY ZHUNG</name><uri>https://sourceforge.net/u/a-z/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=25&amp;page=1#c898</id><summary type="html">"nm libxc.a | grep libxc_free" get the flags but "nm libxcf03.a | grep libxc_free" get nothing on Ubuntu 26.04. U libxc_free_flags U libxc_free_flags 0000000000000690 T libxc_free_flags The flags are almost the same on my archlinux laptop except the last digtal number in "0000……*** T libxc_free_flages"</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=50#c9f4" rel="alternate"/><published>2026-07-05T21:07:03.327000Z</published><updated>2026-07-05T21:07:03.327000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=50#c9f4</id><summary type="html">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.</summary></entry><entry><title>CY ZHUNG modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=25&amp;page=1#af24/42ee" rel="alternate"/><published>2026-07-05T09:24:16.320000Z</published><updated>2026-07-05T09:24:16.320000Z</updated><author><name>CY ZHUNG</name><uri>https://sourceforge.net/u/a-z/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=25&amp;page=1#af24/42ee</id><summary type="html">I have checked. gnu libxc has not been installed as the attachment shows. Libxc was compiled with cmake and the lib files have been cp to folder elk-./src. Do you have any other suggestion？ attachments: sudo apt remove libxc-dev [sudo: authenticate] Password: Package 'libxc-dev' is not installed, so not removed</summary></entry><entry><title>CY ZHUNG posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=25&amp;page=1#af24/42ee" rel="alternate"/><published>2026-07-05T09:23:15.730000Z</published><updated>2026-07-05T09:23:15.730000Z</updated><author><name>CY ZHUNG</name><uri>https://sourceforge.net/u/a-z/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=25&amp;page=1#af24/42ee</id><summary type="html">I have checked. gnu libxc has not been installed as the attachment shows. Libxc was compiled with cmake and the lib files have been cp to folder elk-./src. Do you have any other suggestion？</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=50#af24" rel="alternate"/><published>2026-07-05T08:44:10.843000Z</published><updated>2026-07-05T08:44:10.843000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=50#af24</id><summary type="html">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.</summary></entry><entry><title>CY ZHUNG posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=25#75a1" rel="alternate"/><published>2026-07-04T11:44:10.665000Z</published><updated>2026-07-04T11:44:10.665000Z</updated><author><name>CY ZHUNG</name><uri>https://sourceforge.net/u/a-z/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/bb25c53e70/?limit=25#75a1</id><summary type="html">Dears. ELK-11.02 seems could not read libxcfc03.o on Ubuntu 26.04 LTS. Both intel oneapi and gnu compiler+mkl have the problem when compiling. Could you any help? Attachment from gnu compiler+mkl f03.a libxc.a libwannier.a /usr/bin/x86_64-linux-gnu-ld.bfd: libxcf03.o: in function __xc_f03_lib_m_MOD_xc_f03_functional_get_name': libxcf03.f90:(.text+0xa1e): undefined reference tolibxc_free' /usr/bin/x86_64-linux-gnu-ld.bfd: libxcf03.f90:(.text+0xa9f): undefined reference to `libxc_free' collect2: error:...</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/7852f803f7/?limit=50#cb7c" rel="alternate"/><published>2026-06-30T19:17:57.840000Z</published><updated>2026-06-30T19:17:57.840000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/7852f803f7/?limit=50#cb7c</id><summary type="html">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...</summary></entry><entry><title>Miroslav Iliaš posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/7852f803f7/?limit=25#e081" rel="alternate"/><published>2026-06-30T17:09:24.715000Z</published><updated>2026-06-30T17:09:24.715000Z</updated><author><name>Miroslav Iliaš</name><uri>https://sourceforge.net/u/miroilias/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/7852f803f7/?limit=25#e081</id><summary type="html">Dear Elk-developers, I see elk-11.0.2/species/* are up to Z=104 (element Rf). Would it be please possible to have constructed species up to Z=120 ? The pdf manual has a paragraph called "4.4.2 Species files", but without detailed instructions... Any help is appreciated.</summary></entry><entry><title>Jerry Li posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/84dabadc13/?limit=25#a857/327d/831d" rel="alternate"/><published>2026-06-25T16:21:19.044000Z</published><updated>2026-06-25T16:21:19.044000Z</updated><author><name>Jerry Li</name><uri>https://sourceforge.net/u/jerryxinjie/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/84dabadc13/?limit=25#a857/327d/831d</id><summary type="html">128 threads</summary></entry><entry><title>Kai Luo posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/84dabadc13/?limit=25#a857/327d" rel="alternate"/><published>2026-06-14T07:40:50.037000Z</published><updated>2026-06-14T07:40:50.037000Z</updated><author><name>Kai Luo</name><uri>https://sourceforge.net/u/luokai1103/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/84dabadc13/?limit=25#a857/327d</id><summary type="html">Thanks for testing. How many threads did you use? I tried 8 threads runs fine but could not run beyond 20 outer macro iterations.</summary></entry><entry><title>Jerry Li modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/84dabadc13/?limit=25#a857" rel="alternate"/><published>2026-06-12T15:05:03.957000Z</published><updated>2026-06-12T15:05:03.957000Z</updated><author><name>Jerry Li</name><uri>https://sourceforge.net/u/jerryxinjie/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/84dabadc13/?limit=25#a857</id><summary type="html">I tested the RDMDFT/NiO example with Latest version: 11.0.2. No problem at all!</summary></entry><entry><title>Kai Luo posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/84dabadc13/?limit=25#a857/9816" rel="alternate"/><published>2026-06-12T11:37:16.915000Z</published><updated>2026-06-12T11:37:16.915000Z</updated><author><name>Kai Luo</name><uri>https://sourceforge.net/u/luokai1103/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/84dabadc13/?limit=25#a857/9816</id><summary type="html">I haven't tried it, but I guess rdmft part is not changed at all.</summary></entry><entry><title>Jerry Li posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/84dabadc13/?limit=25#a857" rel="alternate"/><published>2026-06-11T20:42:29.207000Z</published><updated>2026-06-11T20:42:29.207000Z</updated><author><name>Jerry Li</name><uri>https://sourceforge.net/u/jerryxinjie/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/84dabadc13/?limit=25#a857</id><summary type="html">No problem with Latest version: 11.0.2</summary></entry><entry><title>Kai Luo posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/84dabadc13/?limit=25#9883" rel="alternate"/><published>2026-06-10T02:35:04.925000Z</published><updated>2026-06-10T02:35:04.925000Z</updated><author><name>Kai Luo</name><uri>https://sourceforge.net/u/luokai1103/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/84dabadc13/?limit=25#9883</id><summary type="html">Dear All, I have recently tried to use RDMFT calculation with version 10.9.5. Using the example input for NiO, I encountered the following error. I used 1 rank with 16 threads and it finishes the Kohn-Sham task and gives two RDMC_ENERGY.OUT then crashed. Could anyone give suggestions on where it goes wrong? Thanks! 1 -1540.783177 2 -1538.946974 forrtl: severe (174): SIGSEGV, segmentation fault occurred Image PC Routine Line Source libpthread-2.28.s 00001502A6AC5990 Unknown Unknown Unknown elk 0000000001701221...</summary></entry><entry><title>J. K. Dewhurst created a blog post</title><link href="https://sourceforge.net/p/elk/news/2026/06/elk-version-1102-released/" rel="alternate"/><published>2026-06-05T16:53:41.412000Z</published><updated>2026-06-05T16:53:41.412000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/news/2026/06/elk-version-1102-released/</id><summary type="html">Elk version 11.0.2 released</summary></entry><entry><title>Elk released /elk-11.0.2.tgz</title><link href="https://sourceforge.net/projects/elk/files/elk-11.0.2.tgz/download" rel="alternate"/><published>2026-06-05T14:21:03.368000Z</published><updated>2026-06-05T14:21:03.368000Z</updated><author><name>Elk</name><uri>https://sourceforge.net/p/elk/profile/</uri></author><id>https://sourceforge.net/projects/elk/files/elk-11.0.2.tgz/download</id><summary type="html"/></entry><entry><title>Abderrahmane Reggad  posted a comment on discussion Elk Developers</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/628610bf3e/?limit=25#f0ec" rel="alternate"/><published>2026-06-03T18:39:23.552000Z</published><updated>2026-06-03T18:39:23.552000Z</updated><author><name>Abderrahmane Reggad </name><uri>https://sourceforge.net/u/reggad/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897822/thread/628610bf3e/?limit=25#f0ec</id><summary type="html">Dear Elk Developers, I am a teacher and researcher and I have created two modified versions of Elk 3.3.17: elkpdg - a pedagogical copy designed to teach ab initio calculations step by step, using 3 simplified input files. elkader - a research-oriented modification with 3 input files and two Python scripts: one that merges the 3 input files into a standard elk.in, and another that converts all numerical parameters into human-readable form. The algorithms and input file structure are my own design....</summary></entry><entry><title>MUHTASIM MINHAZ posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/86d9bfcee2/?limit=25#a092" rel="alternate"/><published>2026-05-17T08:04:31.034000Z</published><updated>2026-05-17T08:04:31.034000Z</updated><author><name>MUHTASIM MINHAZ</name><uri>https://sourceforge.net/u/minhazsamin/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/86d9bfcee2/?limit=25#a092</id><summary type="html">Dear ELK developers and users, I am calculating the spontaneous electric polarization of NiI₂, a spin-spiral multiferroic, using ELK version 10.9.5. My setup uses spinsprl .true. with a spin-spiral q-vector, DFT+U on Ni (FLL, U=4.2 eV, J=0.8 eV), GGA, spin-orbit coupling, and a 1×1×1 primitive cell. After reading the full task list and grepping the source code, I found that ELK does not appear to have a standalone task for computing spontaneous Berry phase polarization. The subroutine polar() in...</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/e09ece96e8/?limit=50#bb47" rel="alternate"/><published>2026-05-04T20:57:07.653000Z</published><updated>2026-05-04T20:57:07.653000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/e09ece96e8/?limit=50#bb47</id><summary type="html">Hi, If reducebf &lt; 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 &lt; 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.</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/41651d4fca/?limit=50#7c05" rel="alternate"/><published>2026-05-04T20:48:35.718000Z</published><updated>2026-05-04T20:48:35.718000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/41651d4fca/?limit=50#7c05</id><summary type="html">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.</summary></entry><entry><title>inho lee posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/41651d4fca/?limit=25#83ff" rel="alternate"/><published>2026-05-03T23:17:00.307000Z</published><updated>2026-05-03T23:17:00.307000Z</updated><author><name>inho lee</name><uri>https://sourceforge.net/u/inholee/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/41651d4fca/?limit=25#83ff</id><summary type="html">Dear developers, In the manual the Elk code manual, the formulas of vectorpotential in[5.99 pulse] (page 34) and in [7.48 genafieldt] (page 75) looks inconsistent. I am not sure whether the A_0 is divided by (sigma)(sqrt(2pi)) or not. When I change FWHM of pulse, [AFIELD.OUT] did not changed the amplitude, therefore, I think A-0 is not divided by (sigma)(sqrt(2pi)). Would you check the definition of A(t) is consistent in the code?, Also, would you check the peak laser pulse intensity written in [TD-INFO.OUT]...</summary></entry><entry><title>inho lee posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/e09ece96e8/?limit=25#c51a" rel="alternate"/><published>2026-05-03T21:31:12.830000Z</published><updated>2026-05-03T21:31:12.830000Z</updated><author><name>inho lee</name><uri>https://sourceforge.net/u/inholee/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/e09ece96e8/?limit=25#c51a</id><summary type="html">Dear All, I would like to ask about [bfieldc] and [bfcmt] variables. I am investigating RT-TDDFT calculation of ferromagnetic CrI3 monolayer, with spin orbit coupling I used small bfieldc value for ground state calculation [tasks=0]. My questions is, Are bfieldc or bfcmt included in RT-TDDFT calculation? (using tasks=460) (I think that the equation1 in https://pubs.acs.org/doi/pdf/10.1021/acs.nanolett.7b05118 means external B-field is included) If so, do I have to remove bfieldc or bfcmt variables...</summary></entry><entry><title>Chamara Somarathna modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/51ab6972d5/?limit=25#2ae0/a152" rel="alternate"/><published>2026-04-30T15:16:37.907000Z</published><updated>2026-04-30T15:16:37.907000Z</updated><author><name>Chamara Somarathna</name><uri>https://sourceforge.net/u/white-leo/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/51ab6972d5/?limit=25#2ae0/a152</id><summary type="html">Hi Kay, I really appreciate you helping me out a second time. I’m incredibly grateful for your time and expertise With your suggestions of setting lorbcnd=.false. and reduced rgkmax, I was able to get rid of, Warning(rhonorm): total charge density incorrect and Warning(occupy): minimum eigenvalue less than minimum linearisation energy But I still get the warning, Warning(linengy): could not find 1 linearisation energies in all of the sc loops. Did you also see this and which version of Elk do you...</summary></entry><entry><title>Chamara Somarathna posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/51ab6972d5/?limit=25#2ae0/a152" rel="alternate"/><published>2026-04-30T15:15:40.606000Z</published><updated>2026-04-30T15:15:40.606000Z</updated><author><name>Chamara Somarathna</name><uri>https://sourceforge.net/u/white-leo/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/51ab6972d5/?limit=25#2ae0/a152</id><summary type="html">Hi Kay, I really appreciate you helping me out a second time. I’m incredibly grateful for your time and expertise With your suggestions of setting lorbcnd=.false. and reduced rgkmax, I was able to get rid of, Warning(rhonorm): total charge density incorrect and Warning(occupy): minimum eigenvalue less than minimum linearisation energy But I still get the warning, Warning(linengy): could not find 1 linearisation energies in all of the sc loops. Did you also see this and which version of Elk do you...</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/51ab6972d5/?limit=50#2ae0" rel="alternate"/><published>2026-04-30T07:45:48.541000Z</published><updated>2026-04-30T07:45:48.541000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/51ab6972d5/?limit=50#2ae0</id><summary type="html">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...</summary></entry><entry><title>Chamara Somarathna posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/51ab6972d5/?limit=25#f2da" rel="alternate"/><published>2026-04-26T15:10:05.084000Z</published><updated>2026-04-26T15:10:05.084000Z</updated><author><name>Chamara Somarathna</name><uri>https://sourceforge.net/u/white-leo/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/51ab6972d5/?limit=25#f2da</id><summary type="html">Hi all, Elk version used: 10.8.16 I am a new Elk user and your help is very much appriciated. I try to do a ground-state calculation (task 0) for the B1 crystal stucture of MgO at very high compression (~75% compression). But I keep get the following errores in all the sc loops, Warning(rhonorm): total charge density incorrect for s.c. loop 13 Calculated : 79.86817685 Required : 80.00000000 and Warning(linengy): could not find 1 linearisation energies in s.c. loop 11 Warning(occupy): minimum eigenvalue...</summary></entry><entry><title>Long Zhang modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=25#a7b2" rel="alternate"/><published>2026-04-24T20:17:08.490000Z</published><updated>2026-04-24T20:17:08.490000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=25#a7b2</id><summary type="html">Dear Michael and Kay, Thank you both for all of your inputs! I think I have been brought to the point where the rest of the job is to fine tune the calculation. I'll bring up any other issue, if any. Apologies for the delayed response. Best, Long</summary></entry><entry><title>Long Zhang posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=25#a7b2" rel="alternate"/><published>2026-04-24T16:49:51.105000Z</published><updated>2026-04-24T16:49:51.105000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=25#a7b2</id><summary type="html">Dear Michael and Kay, Thank you both for all of your inputs! I think I have been brought to the point where the rest of my job is to fine tune the calculation. I'll bring up any other issue, if any. Sorry for my delayed response. Best, Long</summary></entry><entry><title>Chamara Somarathna posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/e639119b8e/?limit=25#c4b7/c11f" rel="alternate"/><published>2026-04-22T15:46:21.267000Z</published><updated>2026-04-22T15:46:21.267000Z</updated><author><name>Chamara Somarathna</name><uri>https://sourceforge.net/u/white-leo/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/e639119b8e/?limit=25#c4b7/c11f</id><summary type="html">Hi Kay, Thank you very much for your very detailed reply with supportive calculations. This is helping me to go farward solving an issue I had for a week. I will do the batch calculation as you pointed out. From the task 440, I was calculating the pressure in the wrong way. As you pointed out, the stress value given by Elk is ∂E/∂t, but I thought it was the ∂E/∂V. So again, thank you verymuch for your tremendous help. Best, Chamara</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/e639119b8e/?limit=50#c4b7" rel="alternate"/><published>2026-04-22T12:52:27.937000Z</published><updated>2026-04-22T12:52:27.937000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/e639119b8e/?limit=50#c4b7</id><summary type="html">(continued) You can also determine the pressure with task=440 although this is not quite as accurate as fitting to an EOS. Here is the input file I used: tasks 440 wrtdisk .false. deltast 0.001 xctype 3 vhighq .true. lorbcnd .false. nxoapwlo 1 rgkmax 10.0 gmaxvr 32.0 lmaxi 2 ngridk 24 24 24 avec 7.818401918226 0.000000000000 0.000000000000 0.000000000000 7.818401918226 0.000000000000 0.000000000000 0.000000000000 7.818401918226 atoms 2 : nspecies 'Mg.in' : spfname 4 : natoms 0.0 0.0 0.0 0.0 0.5 0.5...</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/e639119b8e/?limit=50#22f9" rel="alternate"/><published>2026-04-22T12:29:06.967000Z</published><updated>2026-04-22T12:29:06.967000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/e639119b8e/?limit=50#22f9</id><summary type="html">Hi Chamara, This type of calculation has to be converged very well to get reliable pressures. The first and most reliable way of performing the calculation is to calculate an energy-vs-volume curve and fit it to an equation of state (EOS). Here is the input file I used for this: tasks 0 1 1 1 1 1 1 1 1 1 1 batch .true. wrtdisk .false. trmt0 .false. xctype 3 vhighq .true. lorbcnd .false. nxoapwlo 1 rgkmax 10.0 gmaxvr 32.0 lmaxi 2 ngridk 24 24 24 avec 7.818401918226 0.000000000000 0.000000000000 0.01...</summary></entry><entry><title>Chamara Somarathna posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/e639119b8e/?limit=25#f1f2" rel="alternate"/><published>2026-04-21T14:59:02.078000Z</published><updated>2026-04-21T14:59:02.078000Z</updated><author><name>Chamara Somarathna</name><uri>https://sourceforge.net/u/white-leo/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/e639119b8e/?limit=25#f1f2</id><summary type="html">Hi, I am new to Elk and trying to compare the VASP pseudopotentials pressure vs Elk all-electron pressures. I started with a simple B1 unit-cell of MgO which gives me about 1 GPa pressure in VASP. However, when I use Elk for the same geometry (converted to Bohr), the STRESS.OUT gives me about 1900 GPa which is not correct. I converted the Elk pressure using 1 Hartree/Bohr^3 = 29421.03 GPa. I would greatly appreciate it if someone could help me to understand this inconsistency. Here is my elk input...</summary></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=50#6840" rel="alternate"/><published>2026-04-13T20:32:09.334000Z</published><updated>2026-04-13T20:32:09.334000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=50#6840</id><summary type="html">Dear Long and Michael, This is quite difficult to converge. I think Michael is correct using mixtype=1 and rgkmax=7.5. I use task=28 to calculate the MAE. This is because it also rotates the spherical covering used for the spherical harmonic transform which eliminates some error. Here is my input file: tasks 28 npmae -1 mixsave .true. wrtdisk .false. epsengy 1.e-6 epspot 1.e-8 spinorb .true. bfieldc 0.0 0.0 0.5 reducebf 0.85 ngridk 24 24 24 rgkmax 7.5 gmaxvr 18.0 lmaxapw 9 lmaxo 8 xctype 20 mixtype...</summary></entry><entry><title>mfechner posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=25#2e29" rel="alternate"/><published>2026-04-10T14:11:33.142000Z</published><updated>2026-04-10T14:11:33.142000Z</updated><author><name>mfechner</name><uri>https://sourceforge.net/u/mfechner/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=25#2e29</id><summary type="html">excuse the delay here is my input see the reducebf ! tasks 0 maxscl 200 nwrite 50 epsengy 1.e-6 epspot 1.e-8 spinpol .true. spinorb .true. sppath '' ngridk 8 8 8 rgkmax 7.5 gmaxvr 18.0 lmaxapw 9 lmaxo 8 xctype 20 0 0 mixtype 1 reducebf 0.95 dft+u 1 1 : dftu, inpdftu 1 3 0.294 0.000 : is, l, U, J !! U on Tb-f, 0.294Ha = 8eV !rotavec ! 0 1 0 270 !! rotation of 270 degrees around +y makes a1 on +z avec -0.1716939763E-14 0.000000000 9.346584740 0.8584698813E-15 8.094379823 -4.673292370 -7.540006694 0.000000000...</summary></entry><entry><title>J. K. Dewhurst created a blog post</title><link href="https://sourceforge.net/p/elk/news/2026/04/elk-version-1095-released/" rel="alternate"/><published>2026-04-09T14:22:04.952000Z</published><updated>2026-04-09T14:22:04.952000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/news/2026/04/elk-version-1095-released/</id><summary type="html">Elk version 10.9.5 released</summary></entry><entry><title>Elk released /elk-10.9.5.tgz</title><link href="https://sourceforge.net/projects/elk/files/elk-10.9.5.tgz/download" rel="alternate"/><published>2026-04-09T14:20:04.449000Z</published><updated>2026-04-09T14:20:04.449000Z</updated><author><name>Elk</name><uri>https://sourceforge.net/p/elk/profile/</uri></author><id>https://sourceforge.net/projects/elk/files/elk-10.9.5.tgz/download</id><summary type="html"/></entry><entry><title>Long Zhang modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=25#5cc5" rel="alternate"/><published>2026-04-01T11:20:33.117000Z</published><updated>2026-04-01T11:20:33.117000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=25#5cc5</id><summary type="html">Dear Michael, I did not get that Warning about linearisation energy when I used rotavec to rotate. But I failed to reproduce your success when I directly use the rotated structure copied from GEOMETRY.OUT. I also tried rgkmax 7 and mixtype 1, still no luck. Here is the elk.in I used, with structure copied from GEOMETRY.OUT. tasks 0 maxscl 200 nwrite 50 epsengy 1.e-6 epspot 1.e-8 spinpol .true. spinorb .true. reducebf 0.95 sppath './species/' ngridk 24 24 24 rgkmax 8.0 gmaxvr 18.0 lmaxapw 9 lmaxo...</summary></entry><entry><title>Long Zhang posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=25#5cc5" rel="alternate"/><published>2026-04-01T11:19:23.243000Z</published><updated>2026-04-01T11:19:23.243000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=25#5cc5</id><summary type="html">Dear Michael, I did not get that Warning about linearisation energy when I used rotavec to rotate. But I failed to reproduce your success when I used the rotated structure copied from GEOMETRY.OUT. I also tried rgkmax 7 and mixtype 1, still no luck. Here is the elk.in I used, with structure copied from GEOMETRY.OUT. tasks 0 maxscl 200 nwrite 50 epsengy 1.e-6 epspot 1.e-8 spinpol .true. spinorb .true. reducebf 0.95 sppath './species/' ngridk 24 24 24 rgkmax 8.0 gmaxvr 18.0 lmaxapw 9 lmaxo 8 xctype...</summary></entry><entry><title>mfechner posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=25#025e" rel="alternate"/><published>2026-03-31T10:24:11.809000Z</published><updated>2026-03-31T10:24:11.809000Z</updated><author><name>mfechner</name><uri>https://sourceforge.net/u/mfechner/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=25#025e</id><summary type="html">two more small things to recommend: you should switch on reducebf, which decreases the applied magnetic field as you want to have at the end a computation without applied field. A value of 0.95 is good in most cases. In addition switching to linear mixing sometimes keep things also more stabe!</summary></entry><entry><title>mfechner posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=25#b0e9" rel="alternate"/><published>2026-03-31T10:22:23.775000Z</published><updated>2026-03-31T10:22:23.775000Z</updated><author><name>mfechner</name><uri>https://sourceforge.net/u/mfechner/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=25#b0e9</id><summary type="html">Dear Long, I tried your case and can indeed confirm the issue. However, could you check whether, when you rotate, there are any messages in the standard output files such as: "Warning(occupy): minimum eigenvalue less than minimum linearisation energy : -44.32280382 -4.383400000 for s.c. loop 137" This suggests that there is an issue when you rotate. Instead of rotating the structure using Elk, I tried directly using the rotated structure. (You can find it in GEOMETRY.OUT and directly copy it into...</summary></entry><entry><title>Long Zhang posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=25#f988" rel="alternate"/><published>2026-03-30T20:34:12.926000Z</published><updated>2026-03-30T20:34:12.926000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/7da011c8dd/?limit=25#f988</id><summary type="html">Dear ELK users and developers, I am studying magnetic TbCo5 using dft+u, with the current version 10.8.16 of ELK. The elk.in is copied at the end of this message. I am basically doing the MAE thing for this 4f SOC system, but with bfcmt instead of bfieldc used in the given mae example, because I want to get the afm in unit cell state. The lattice vectors are like this: a1 and a2 in xy-plane and 120 degrees away, a3 on z-axis. |a1|=|a2|=a, |a3|=c. I managed to converge it when I set magnetic direction...</summary></entry><entry><title>Markus modified a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/f5997e6dc4/?limit=25#6545" rel="alternate"/><published>2026-03-27T09:59:42.028000Z</published><updated>2026-03-27T09:59:42.028000Z</updated><author><name>Markus</name><uri>https://sourceforge.net/u/markusmeinert/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/f5997e6dc4/?limit=25#6545</id><summary type="html">Yes, there is an example for CaO in the examples folder. You need the core states as valence, i.e. you have to model them with local orbitals. There is a special species file for Ca in the example which shows you how to do it for the 2p states. These calculations are extremely expensive and need lots of RAM. Carefully optimize the initial and final states to keep the matrix size as small as possible and start with a very small number of k-points. XMCD is also possible, compare with the XMCD example....</summary></entry><entry><title>Markus posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/f5997e6dc4/?limit=25#6545" rel="alternate"/><published>2026-03-27T09:58:16.575000Z</published><updated>2026-03-27T09:58:16.575000Z</updated><author><name>Markus</name><uri>https://sourceforge.net/u/markusmeinert/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/f5997e6dc4/?limit=25#6545</id><summary type="html">Yes, there is an example for CaO in the examples folder. You need the core states as valence, i.e. you have to model them with local orbitals. There is a special species file for Ca in the example which shows you how to do it for the 2p states. These calculations are extremely expensive and need lots of RAM. Carefully optimize the initial and final states to keep the matrix size as small as possible and start with a very small number of k-points. Best wishes, Markus</summary></entry><entry><title>Krishnendu Mukherjee posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/95c0274188/?limit=25#a076" rel="alternate"/><published>2026-03-26T14:58:36.907000Z</published><updated>2026-03-26T14:58:36.907000Z</updated><author><name>Krishnendu Mukherjee</name><uri>https://sourceforge.net/u/krishnendu/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/95c0274188/?limit=25#a076</id><summary type="html">Dear All, Has anybody modelled the coherence length of superconductor using elk code. Any guidance will be greatly appreciated. Greetings, Krishnendu</summary></entry><entry><title>Jerry Li posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/f5997e6dc4/?limit=25#eb9d" rel="alternate"/><published>2026-03-26T04:04:29.266000Z</published><updated>2026-03-26T04:04:29.266000Z</updated><author><name>Jerry Li</name><uri>https://sourceforge.net/u/jerryxinjie/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/f5997e6dc4/?limit=25#eb9d</id><summary type="html">Hello everyone, I am interested in computing X-ray Magnetic Circular Dichroism (XMCD) and X-ray Absorption Near Edge Structure (XANES) for magnetic semiconductors. Has anyone performed Elk calculations of X-ray absorption spectra (XAS) using the Bethe–Salpeter equation (BSE)? I would greatly appreciate it if you could share an example of your input files. Thank you in advance!</summary></entry><entry><title>Hanley Coath posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/b27919ed73/?limit=25#688a/3ce2" rel="alternate"/><published>2026-03-24T17:09:23.947000Z</published><updated>2026-03-24T17:09:23.947000Z</updated><author><name>Hanley Coath</name><uri>https://sourceforge.net/u/hcoath/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/b27919ed73/?limit=25#688a/3ce2</id><summary type="html">Hello Markus, Thanks for pointing me towards the code, it was very useful.</summary></entry><entry><title>Markus posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/b27919ed73/?limit=25#688a" rel="alternate"/><published>2026-03-24T16:33:49.450000Z</published><updated>2026-03-24T16:33:49.450000Z</updated><author><name>Markus</name><uri>https://sourceforge.net/u/markusmeinert/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/b27919ed73/?limit=25#688a</id><summary type="html">Hi Hanley, you can find it in the code, start from occupy.f90: https://elk.sourceforge.io/doxygen/html/occupy_8f90_source.html Then look for stheta.f90 and stheta_fd.f90: https://elk.sourceforge.io/doxygen/html/stheta_8f90_source.html and https://elk.sourceforge.io/doxygen/html/stheta__fd_8f90_source.html The argument of the standard Fermi-Dirac distribution is (efermi-e)/swidth, see line 64 of occupy.f90. This is the distribution, then the approximation to the smooth delta function is specified...</summary></entry><entry><title>Hanley Coath posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/b27919ed73/?limit=25#3146" rel="alternate"/><published>2026-03-24T12:06:13.790000Z</published><updated>2026-03-24T12:06:13.790000Z</updated><author><name>Hanley Coath</name><uri>https://sourceforge.net/u/hcoath/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/b27919ed73/?limit=25#3146</id><summary type="html">In the elk manual, it says that the default function used when calculating the smooth approximation to the Dirac-Delta function (swidth), uses a Fermi-Dirac distribution. What is the exact form of the distribution used in the elk code? It doesn't seem to be referenced in the manual. Is it the standard Fermi-Dirac distribution? ie. f(z) = 1 / (1 + exp z−μ/kBT) or is it some other approximation of this function (such as: https://link.springer.com/article/10.1007/s11075-010-9397-6). If anyone knows,...</summary></entry><entry><title>J. K. Dewhurst created a blog post</title><link href="https://sourceforge.net/p/elk/news/2026/02/elk-version-10816-released/" rel="alternate"/><published>2026-02-23T14:46:43.393000Z</published><updated>2026-02-23T14:46:43.393000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/news/2026/02/elk-version-10816-released/</id><summary type="html">Elk version 10.8.16 released</summary></entry><entry><title>Elk released /elk-10.8.16.tgz</title><link href="https://sourceforge.net/projects/elk/files/elk-10.8.16.tgz/download" rel="alternate"/><published>2026-02-23T14:45:03.438000Z</published><updated>2026-02-23T14:45:03.438000Z</updated><author><name>Elk</name><uri>https://sourceforge.net/p/elk/profile/</uri></author><id>https://sourceforge.net/projects/elk/files/elk-10.8.16.tgz/download</id><summary type="html"/></entry><entry><title>J. K. Dewhurst posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/09b7ccc873/?limit=50#a8e2" rel="alternate"/><published>2026-02-11T23:10:41.081000Z</published><updated>2026-02-11T23:10:41.081000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/09b7ccc873/?limit=50#a8e2</id><summary type="html">Hi All, Don't forget to use primcell=.true. after you've set up your magnetic fields. Elk will then find the smallest unit cell which produces the same crystal. For example tasks 0 primcell .true. spinpol .true. ngridk 2 2 2 avec 9.230000000 0.000000000 0.000000000 0.000000000 9.230000000 0.000000000 0.000000000 0.000000000 18.46000000 atoms 2 : nspecies 'U.in' : spfname 8 : natoms; atposl, bfcmt below 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.10000000 0.00000000 0.00000000 0.50000000...</summary></entry><entry><title>Hanley Coath posted a comment on discussion Elk Users</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/09b7ccc873/?limit=25#75eb" rel="alternate"/><published>2026-02-11T15:40:20.413000Z</published><updated>2026-02-11T15:40:20.413000Z</updated><author><name>Hanley Coath</name><uri>https://sourceforge.net/u/hcoath/profile/</uri></author><id>https://sourceforge.net/p/elk/discussion/897820/thread/09b7ccc873/?limit=25#75eb</id><summary type="html">Hi Markus, thanks for the response. I had literally just figured this out this morning when I opened my structure file with xcrysden, and was quite surprised to see a UN2 structure. Definitely a quick fix. Cheers</summary></entry></feed>