<?xml version="1.0" encoding="utf-8"?>
<feed xml:lang="en" xmlns="http://www.w3.org/2005/Atom"><title>Recent posts to Discussion</title><link href="https://sourceforge.net/p/elk/discussion/" rel="alternate"/><link href="https://sourceforge.net/p/elk/discussion/feed.atom" rel="self"/><id>https://sourceforge.net/p/elk/discussion/</id><updated>2026-11-05T20:47:58.561000Z</updated><subtitle>Recent posts to Discussion</subtitle><entry><title>general question about the spin flavor of a spinor wavefunction</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#ba50" rel="alternate"/><published>2026-11-05T20:47:58.561000Z</published><updated>2026-11-05T20:47:58.561000Z</updated><author><name>Lars Nordström</name><uri>https://sourceforge.net/u/larsnordstrom/</uri></author><id>https://sourceforge.net26796ae0d2557d48fa3763a7fe07199ebdfe7495</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;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. &lt;br/&gt;
The flags are for the magnetization desities not the order. &lt;br/&gt;
I never force the magnetization density to be collinear in presence of SOC and assume it is not common practise.&lt;br/&gt;
best wishes&lt;br/&gt;
/Lars&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>general question about the spin flavor of a spinor wavefunction</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#2bd9" rel="alternate"/><published>2026-11-05T16:15:52.400000Z</published><updated>2026-11-05T16:15:52.400000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/</uri></author><id>https://sourceforge.nete77d23fb44a1f4614a67a46f240e47dab95a3215</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;And, it looks collinearity (ncmag=false) is less likely to happen, for spin polarized + SOC. &lt;br/&gt;
Following the logic in init0, &lt;br/&gt;
spinpol=true And spinorb=true will make ndmag=3 (thus ncmag=true). &lt;br/&gt;
ndmag=3 can also be triggered if the quantities bfieldc0 or bfcmt0 become larger than threshhold value. &lt;br/&gt;
ndmag=1 only happens later in the code through the condition cmagz=true. &lt;br/&gt;
But cmagz does not default to be true. &lt;br/&gt;
So, one needs to set cmagz=true in input file. &lt;br/&gt;
Is forcing collinearity in spin polarized + SOC a common practise? &lt;/p&gt;
&lt;p&gt;Thanks again.&lt;/p&gt;
&lt;p&gt;Best,&lt;br/&gt;
Long&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>general question about the spin flavor of a spinor wavefunction</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#8c9d/96f3" rel="alternate"/><published>2026-11-05T01:42:52.318000Z</published><updated>2026-11-05T01:42:52.318000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/</uri></author><id>https://sourceforge.nete8c60aaca87b188900fa388ef0c670877f7bc73e</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Hi Kay, &lt;/p&gt;
&lt;p&gt;It seems collinearity (boolean ncmag=false) is less likely to happen for spin polarized SOC calculation. &lt;br/&gt;
Following the logic of the code in init0, &lt;br/&gt;
First, spinpol=true AND spinorb=true will make ndmag=3 (thus ncmag=true). &lt;br/&gt;
ndmag=3 will also be triggered if the quantities bfieldc0 or bfcmt0 become larger than some threshhold value. &lt;br/&gt;
ndmag=1 can happen later in the code only through the condition cmagz=true. &lt;br/&gt;
But cmagz does not default to be true. &lt;br/&gt;
So, I'll need to set cmagz=true in input file, to have collinear spin polarized SOC calculation. &lt;br/&gt;
Is force collinearity a common practise in this case? Do you know how it affects accuracy? &lt;/p&gt;
&lt;p&gt;Another thing that I would apprecite any comments is, &lt;br/&gt;
suppose the collinear spin polarized SOC calculation is done anyway, &lt;br/&gt;
the j-th column of evecsv(nstsv,nstsv) is used to do linear combination of 1st-variational states, &lt;br/&gt;
evecsv( 1:nstfv, j ) yields the up spinor component of state-j, &lt;br/&gt;
evecsv( nstfv+1:nstsv, j ) yields the down spinor component of state-j. &lt;br/&gt;
As you said they now represent spin-up and spin-down, but I assume I would get states with mixed up and down. &lt;br/&gt;
I am looking at the fix spin moment (FSM) feature of ELK. This is new to me. &lt;br/&gt;
Will FSM be helpful to make the spinors less mixed? or it's not related? &lt;/p&gt;
&lt;p&gt;Best Regards, &lt;br/&gt;
Long&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>general question about the spin flavor of a spinor wavefunction</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=50#8c9d" rel="alternate"/><published>2026-11-04T08:03:33.227000Z</published><updated>2026-11-04T08:03:33.227000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/</uri></author><id>https://sourceforge.net8d79f453a7d6e7a74d7c724826ed925de3ae2f2d</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Hi Long and Miroslav,&lt;/p&gt;
&lt;p&gt;Just to clarify: for &lt;strong&gt;collinear&lt;/strong&gt; calculations the first nstfv orbitals are spin-up and the next nstfv orbitals are spin-down.&lt;/p&gt;
&lt;p&gt;For &lt;strong&gt;non-collinear&lt;/strong&gt; calculations the orbitals are not separable into pure up and down but rather general 2-spinor functions.&lt;/p&gt;
&lt;p&gt;Regards,&lt;br/&gt;
Kay.&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>general question about the spin flavor of a spinor wavefunction</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#4684" rel="alternate"/><published>2026-11-04T04:05:35.853000Z</published><updated>2026-11-04T04:05:35.853000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/</uri></author><id>https://sourceforge.net51739da0671ee4862ea2d27c215ba0c3555d8fc2</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Hi Miroslav, &lt;/p&gt;
&lt;p&gt;Thanks for the quick response! &lt;br/&gt;
I did think that one spinor wavefunction cannot be exclusively tagged as spin-up or down. &lt;br/&gt;
AI analyzer just confirmed this point. &lt;br/&gt;
So, this means I need to adjust what I want to do and what I can do ......  &lt;/p&gt;
&lt;p&gt;Best, &lt;br/&gt;
Long&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>general question about the spin flavor of a spinor wavefunction</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#08e7" rel="alternate"/><published>2026-11-04T03:25:14.291000Z</published><updated>2026-11-04T03:25:14.291000Z</updated><author><name>Miroslav Iliaš</name><uri>https://sourceforge.net/u/miroilias/</uri></author><id>https://sourceforge.net2ce54888c5c2124f907ed42797b39d8a3d685fa2</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Hello Long, &lt;/p&gt;
&lt;p&gt;I uploaded the fresh ELk code to chatgpt AI analyzer, together with your question, and got this answers , &lt;a href="https://chatgpt.com/share/6aeaa4a5-42e0-83eb-8fe9-0e462c0cafa2" rel="nofollow"&gt;https://chatgpt.com/share/6aeaa4a5-42e0-83eb-8fe9-0e462c0cafa2&lt;/a&gt; .&lt;/p&gt;
&lt;p&gt;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.  &lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>general question about the spin flavor of a spinor wavefunction</title><link href="https://sourceforge.net/p/elk/discussion/897820/thread/6ac5de2d92/?limit=25#bd41" rel="alternate"/><published>2026-11-04T02:48:54.117000Z</published><updated>2026-11-04T02:48:54.117000Z</updated><author><name>Long Zhang</name><uri>https://sourceforge.net/u/lzhang7/</uri></author><id>https://sourceforge.net6ed953c161cabf1e5234a0f2dfbd9846e78f35d1</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Dear ELK users and developers, &lt;/p&gt;
&lt;p&gt;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.  &lt;/p&gt;
&lt;p&gt;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. &lt;/p&gt;
&lt;p&gt;I will have to identify which of the selected wavefunctions correspond to spin-up states, and which correspond to spin-down states. Since they are sorted according to band energies, ascending or descending, spin-up and spin-down wavefunctions are mixed in order. &lt;/p&gt;
&lt;p&gt;My question is how to judge a given spinor wavefunction is a spin-up state or a spin-down state? &lt;/p&gt;
&lt;p&gt;I tried to find an existing variable/quantity in the code that can help to tag one spinor wavefunction to be spin-up or down, I couldn’t find one. &lt;/p&gt;
&lt;p&gt;Then I was thinking to compare the square of up-spinor component with the square of down-spinor component in the same wavefunction. If one square is close to zero and the other square is close to one, then it indicates the spin flavor of that wavefunction. &lt;br/&gt;
Is this the correct way to judge the spin flavor? Is this the only way? &lt;/p&gt;
&lt;p&gt;Again, appreciate any advice and suggestion.  &lt;/p&gt;
&lt;p&gt;Best Regards, &lt;br/&gt;
Long&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>problem with generating Cn species file</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/3f7447da52/?limit=25#eaba" rel="alternate"/><published>2026-10-18T07:33:31.668000Z</published><updated>2026-10-18T07:33:31.668000Z</updated><author><name>Miroslav Iliaš</name><uri>https://sourceforge.net/u/miroilias/</uri></author><id>https://sourceforge.netd35fea420a4f644e92b04ce53350c8cd63e782e4</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;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. &lt;/p&gt;
&lt;p&gt;The f90 files you provided are from the latest Elk  11.2.3, right ?&lt;/p&gt;
&lt;p&gt;Would be good to have  more verbose output when generating  the species files for new SHEs.&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>problem with generating Cn species file</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/3f7447da52/?limit=25#a21e" rel="alternate"/><published>2026-10-18T00:18:27.594000Z</published><updated>2026-10-18T00:18:27.594000Z</updated><author><name>Peter Elliott</name><uri>https://sourceforge.net/u/elliottp/</uri></author><id>https://sourceforge.netdb8427622c08f0ebdff3495fe90119c3156fd791</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;From your output it doesn't look like it crashed, it just gave a warning.&lt;/p&gt;
&lt;p&gt;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.&lt;/p&gt;
&lt;p&gt;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 updates nrmt after each pass. &lt;br/&gt;
Still, if it didn't give further warnings, the smaller mesh was probably fine.  You can always edit the species file manually if you want to change nrmt. &lt;/p&gt;
&lt;p&gt;If the occupations are known, it's maybe safer to use a species block instead of fspecies, as in the example input file attached.  &lt;/p&gt;
&lt;p&gt;I have zero experience with these super-heavy atoms, but I imagine they are very tricky, especially with a deep ecvcut value, which promotes a lot of semi-core states to the valence calculation, but maybe that's necessary.&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>How to Perform a One-Shot (G0W0) Calculation in the Elk Code</title><link href="https://sourceforge.net/p/elk/discussion/897822/thread/73e0f879c3/?limit=50#4744" rel="alternate"/><published>2026-10-14T12:28:28.113000Z</published><updated>2026-10-14T12:28:28.113000Z</updated><author><name>J. K. Dewhurst</name><uri>https://sourceforge.net/u/jkdewhurst/</uri></author><id>https://sourceforge.nete039a6f8512a67d91b02604a04e4eb477b2714b7</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Hi,&lt;/p&gt;
&lt;p&gt;The GW calculation in Elk is one-shot G0W0 in the Matsubara frequency representation.&lt;/p&gt;
&lt;p&gt;It is very important to converge all the relevant parameters for the GW calculation.&lt;/p&gt;
&lt;p&gt;Here is an example for Ni which is properly converged:&lt;/p&gt;
&lt;div class="codehilite"&gt;&lt;pre&gt;&lt;span&gt;&lt;/span&gt;&lt;code&gt;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 100 0
 -2.5 2.0

! large number of empty states required
nempty
  64

ngridk
  16 16 16

ngridq
  8 8 8

avec
  1.0  1.0  0.0
  1.0  0.0  1.0
  0.0  1.0  1.0

scale
  3.33

atoms
  1                                   : nspecies
  'Ni.in'                             : spfname
  1                                   : natoms
  0.0  0.0  0.0    0.0  0.0  0.0      : atposl, bfcmt
&lt;/code&gt;&lt;/pre&gt;&lt;/div&gt;

&lt;p&gt;I've attached the GW spectral function. The low spectral weight values near the Fermi energy are a known problem with the GW method (see: I. V. Solovyev, Phys. Rev. B 71, 045103 (2005)) however the peak positions appear to be fine.&lt;/p&gt;
&lt;p&gt;Note that these calculations can be very expensive if there are a large number of states, k-points and Matsubara frequencies required. This calculation required about 20 hours on 2304 cores.&lt;/p&gt;
&lt;p&gt;Regards,&lt;br/&gt;
Kay.&lt;/p&gt;&lt;/div&gt;</summary></entry></feed>