<!DOCTYPE html>
<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    <p>I am not really a tcsh script expert, but i would introduce an if
      ... else ... instruction into the "run_lapw" script.</p>
    <p><br>
    </p>
    <p>like if (cycle =1 and fortran compiler = ifx) then set
      OMP_NUM_THREADS for lapw0 = 1</p>
    <p>else (cycle > 1) set OMP_NUM_THREADS for lapw0 back to the
      original value in .machines .. </p>
    <p><br>
    </p>
    <p>That would run in every directory and every case and I think it
      is easy to include. But As the script is huge I did not find the
      variables I need to use,  </p>
    <p>any help from the experts who are familiar with the script
      content would be appreciated here and the fastest way.</p>
    <p><br>
    </p>
    <p>Thanks in advance,</p>
    <p>Michael</p>
    <p><br>
    </p>
    <div class="moz-cite-prefix"><br>
    </div>
    <blockquote type="cite"
      cite="mid:3fa9e599-f83a-49f3-87cc-742bae15b08f@ruhr-uni-bochum.de">
      <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
      <div class="msgHead"> <span class="sender pipe"><a
href="https://www.mail-archive.com/search?l=wien@zeus.theochem.tuwien.ac.at&q=from:%22Gavin+Abo%22"
            rel="nofollow" moz-do-not-send="true"><span
              itemprop="author" itemscope=""
              itemtype="http://schema.org/Person"><span itemprop="name">Gavin
                Abo</span></span></a></span> <span class="date"><a
href="https://www.mail-archive.com/search?l=wien@zeus.theochem.tuwien.ac.at&q=date:20250707"
            rel="nofollow" moz-do-not-send="true">Mon, 07 Jul 2025
            05:32:05 -0700</a></span> <br>
      </div>
      <div itemprop="articleBody" class="msgBody"><br>
        *Is there any solution to this problem?* </div>
      <div itemprop="articleBody" class="msgBody"><br>
      </div>
      <div itemprop="articleBody" class="msgBody">As Michael mentioned
        at [1], you could try running OMP_NUM_THREADS=1 for the first
        cycle and then more threads such as  8 (OMP_NUM_THREADS=8) after
        that.</div>
      <div itemprop="articleBody" class="msgBody"><br>
      </div>
      <div itemprop="articleBody" class="msgBody">Perhaps you could make
        a shell script to do that with the file containing something
        like:</div>
      <div itemprop="articleBody" class="msgBody"><br>
      </div>
      <div itemprop="articleBody" class="msgBody">run_lapw -i 1 </div>
      <div itemprop="articleBody" class="msgBody">export
        OMP_NUM_THREADS=1 </div>
      <div itemprop="articleBody" class="msgBody">run_lapw -NI </div>
      <div itemprop="articleBody" class="msgBody">export
        OMP_NUM_THREADS=8 </div>
      <div itemprop="articleBody" class="msgBody"><br>
      </div>
      <div itemprop="articleBody" class="msgBody">[1] <a rel="nofollow"
href="https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg23780.html"
          class="moz-txt-link-freetext" moz-do-not-send="true">https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg23780.html</a><br>
      </div>
      <div itemprop="articleBody" class="msgBody"><br>
      </div>
      <div itemprop="articleBody" class="msgBody">*or else Should I go
        to gfortran/gcc? I have not seen the benchmark timing of ifx and
        gfortran in recent times.*<br>
      </div>
      <div itemprop="articleBody" class="msgBody"><br>
      </div>
      <div itemprop="articleBody" class="msgBody">Currently, I don't
        have any benchmark timing for ifx and gfortran.  It's up to you
        which compilers you try to use.<br>
      </div>
      <div itemprop="articleBody" class="msgBody"><br>
      </div>
      <div itemprop="articleBody" class="msgBody">*Is there a possible
        source of getting old ifort ?* <br>
      </div>
      <div itemprop="articleBody" class="msgBody"><br>
      </div>
      <div itemprop="articleBody" class="msgBody">The old ifort versions
        are likely still downloadable from Intel's website.  However, as
        you should see on the website at [2], a paid for and active
        Priority Support is needed for accessing them in the Intel®
        Registration Center (IRC).</div>
      <div itemprop="articleBody" class="msgBody"><br>
      </div>
      <div itemprop="articleBody" class="msgBody">[2] <a rel="nofollow"
href="https://community.intel.com/t5/Blogs/Tech-Innovation/Tools/A-Historic-Moment-for-The-Intel-Fortran-Compiler-Classic-ifort/post/1614625"
          class="moz-txt-link-freetext" moz-do-not-send="true">https://community.intel.com/t5/Blogs/Tech-Innovation/Tools/A-Historic-Moment-for-The-Intel-Fortran-Compiler-Classic-ifort/post/1614625</a><br>
        K</div>
      <div itemprop="articleBody" class="msgBody">ind Regards, </div>
      <div itemprop="articleBody" class="msgBody">Gavin WIEN2k user</div>
      <p><br>
      </p>
      <div class="moz-cite-prefix">Am 07.07.2025 um 12:16 schrieb
        susanta mohanta:<br>
      </div>
      <blockquote type="cite"
cite="mid:CAE5f87C=Qn2ReKH422EVKKuyJe_7=koYnry0a7u-dTp3pak5cA@mail.gmail.com">
        <meta http-equiv="content-type"
          content="text/html; charset=UTF-8">
        <div dir="ltr">Dear Prof. Michael Fechtelkord, thank you for
          your suggestion. I have set lapw0:1 and lapw1:8; lapw2:8 in
          .machines and running since lapw0 does not take much time to
          complete compared to others but probably your suggestion is
          better. Could you please write some patch to do this
          automatically, I mean one does not need to stop after one
          cycle.
          <div><br>
          </div>
          <div>with warm regards</div>
          <div>Susanta</div>
        </div>
        <br>
        <div class="gmail_quote gmail_quote_container">
          <div dir="ltr" class="gmail_attr">On Mon, Jul 7, 2025 at
            3:32 PM Michael Fechtelkord <<a
              href="mailto:Michael.Fechtelkord@ruhr-uni-bochum.de"
              moz-do-not-send="true" class="moz-txt-link-freetext">Michael.Fechtelkord@ruhr-uni-bochum.de</a>>
            wrote:<br>
          </div>
          <blockquote class="gmail_quote"
style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
            <div bgcolor="#FFFFFF">
              <p>Hello Susanta,</p>
              <p><br>
              </p>
              <p>set  omp_lapw0:1 in .machines ONLY for the first cycle.</p>
              <p><br>
              </p>
              <p>you can do calculate only one cycle by either by using
                the "-i 1" in run_lapw or by using "touch .stop"</p>
              <div>After running the first cycle you can set
                back  omp_lapw0:8 in .machines. Then it uses all 8
                threads again and should not crash again. Only the first
                cycle is senstive to omp threads more than one.he
                continiu</div>
              <div><br>
              </div>
              <div>Do not forget to use the -NI flag in the continuing
                cycles to avoid that the Broyden files are deleted.</div>
              <div><br>
              </div>
              <div>Best regards,</div>
              <div>Michael</div>
              <div><br>
              </div>
              <div><br>
              </div>
              <div>Am 07.07.2025 um 10:58 schrieb susanta mohanta:<br>
              </div>
              <blockquote type="cite">
                <div dir="ltr">
                  <div>Dear Prof. Abo, I have set <font color="#0000ff">OMP_NUM_THREADS=1</font>:
                    and it's running without problem. But by doing so,
                    it is using only one core out of 8. As mentioned by
                    MIcheal, it only runs without error if one
                    sets omp_lapw0:1 in .machine file and always crashes
                    if I increase more than 1. Is there any solution to
                    this problem? or else Should I go to gfortran/gcc? I
                    have not seen the benchmark timing of ifx and
                    gfortran in recent times.  Is there a possible
                    source of getting old ifort ?</div>
                  <div><br>
                  </div>
                  <div>Thank you for your kind help. Waiting for your
                    reply</div>
                  <div>with warm regards</div>
                  <div>Susanta</div>
                </div>
                <br>
                <div class="gmail_quote">
                  <div dir="ltr" class="gmail_attr">On Mon, Jul 7, 2025
                    at 1:45 PM Gavin Abo <<a
                      href="mailto:gabo13279@gmail.com" target="_blank"
                      moz-do-not-send="true"
                      class="moz-txt-link-freetext">gabo13279@gmail.com</a>>
                    wrote:<br>
                  </div>
                  <blockquote class="gmail_quote"
style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
                    <div>
                      <p>When you ran ./userconfig_lapw during
                        installation, if you set OMP_NUM_THREADS ≥ 2, a
                        solution might be to set OMP_NUM_THREADS to 1.
                        Unfortunately, there is a known lapw0 issue with
                        using 2 or more OMP_NUM_THREADS when compiling
                        WIEN2k with the ifx compiler at this time as
                        mentioned in a previous post to this mailing
                        list at [1].</p>
                      <p>You could check your ~/.bashrc to see what you
                        had ./userconfig_lapw generate there for
                        OMP_NUM_THREADS.</p>
                      <p>For example, my ~/.bashrc file has a line with
                        <font color="#0000ff">OMP_NUM_THREADS=1</font>:<br>
                      </p>
                      <p>username@computername:~$ grep OMP_NUM_THREADS
                        ~/.bashrc <br>
                        if [ "$OMP_NUM_THREADS" = "" ]; then export <font
                          color="#0000ff">OMP_NUM_THREADS=1</font>; fi<br>
                      </p>
                      <p>So, if have a value other than 1, you could try
                        editing ~/.bashrc and change it to be <font
                          color="#0000ff">OMP_NUM_THREADS=1</font>.<br>
                      </p>
                      <p>If that doesn't remove the lapw0 error, there
                        may be some other runtime issue with the lapw0
                        compiled with ifx for your particular computer
                        system.  In which case, you might have to use
                        instead other compilers (e.g., gfortran/gcc [2])
                        to get around the issue.</p>
                      [1] <a
href="https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg23752.html"
                        target="_blank" moz-do-not-send="true"
                        class="moz-txt-link-freetext">https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg23752.html</a><br>
                      [2] <a
href="https://github.com/gsabo/WIEN2k-Docs/blob/main/WIEN2k24.1_Ubuntu24.04.1_Install_with_gfortran.pdf"
                        target="_blank" moz-do-not-send="true"
                        class="moz-txt-link-freetext">https://github.com/gsabo/WIEN2k-Docs/blob/main/WIEN2k24.1_Ubuntu24.04.1_Install_with_gfortran.pdf</a><br>
                      <p>Kind Regards,</p>
                      Gavin<br>
                      WIEN2k usereb Page: <a
                        class="moz-txt-link-freetext"
href="https://www.ruhr-uni-bochum.de/kristallographie/kc/mitarbeiter/fechtelkord/"
                        moz-do-not-send="true">https://www.ruhr-uni-bochum.de/kristallographie/kc/mitarbeiter/fechtelkord/</a></div>
                  </blockquote>
                </div>
              </blockquote>
            </div>
          </blockquote>
        </div>
      </blockquote>
    </blockquote>
    <pre class="moz-signature" cols="72">-- 
Dr. Michael Fechtelkord

Institut für Geologie, Mineralogie und Geophysik
Ruhr-Universität Bochum
Universitätsstr. 150
D-44780 Bochum

Phone: +49 (234) 32-24380
Fax:  +49 (234) 32-04380
Email: <a class="moz-txt-link-abbreviated moz-txt-link-freetext"
    href="mailto:Michael.Fechtelkord@ruhr-uni-bochum.de">Michael.Fechtelkord@ruhr-uni-bochum.de</a>
Web Page: <a class="moz-txt-link-freetext"
href="https://www.ruhr-uni-bochum.de/kristallographie/kc/mitarbeiter/fechtelkord/">https://www.ruhr-uni-bochum.de/kristallographie/kc/mitarbeiter/fechtelkord/</a>
</pre>
  </body>
</html>