<div dir="ltr"><div>Dear Aron,<br><br></div>thanks for your kind reply. <br>Still look forward to receiving any help on the error from other users.<br></div><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Jun 29, 2015 at 1:22 PM, Aron Szabo <span dir="ltr"><<a href="mailto:szaboa@iis.ee.ethz.ch" target="_blank">szaboa@iis.ee.ethz.ch</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
  
    
  
  <div bgcolor="#FFFFFF" text="#000000">
    <div>Dear Seyed,<br>
      <br>
      I've managed to get nicely converged Wannier functions for
      phosphorene by using sp3 projections on each 4 atoms, therefore
      including 6 conduction bands in the calculation too with the
      disentanglement method. The largest spread was 2 A^2. In my
      experience it should not be much larger than 3 A^2 for most
      semiconductors. And the imaginary part of the Hamiltonian should
      be practically 0 everywhere (or in the order of magnitude of
      1e-6). I was using 21x21x1 k points. I did not check with an even
      mesh, but sometimes it matters a lot for the wannierization
      whether you use even or odd mesh. I've used VASP, and I'm not an
      expert with espresso, and I did not look into you input and output
      files. Just wanted share my experience with phosphorene.<br>
      <br>
      Best,<br>
      Aron<div><div class="h5"><br>
      <br>
      On 06/29/2015 09:38 AM, Seyed Mojtaba Rezaei Sani wrote:<br>
    </div></div></div>
    <blockquote type="cite"><div><div class="h5">
      <div dir="ltr">
        <div>
          <div>
            <div>
              <div>Dear Nicola,<br>
                <br>
              </div>
              Thanks. Phosphorene has four P atoms in its unit cell so
              considering 10 valence bands and therefore 10 WFs, we
              can't fit a projection of s orbital on every atom and bond
              to this number. <br>
            </div>
            But I have found new thing. Instead of increasing the number
            of kpoints in wannier input, I chose a coarser grid which
            seems to be fine enough to provide converged WFs with both
            versions of codes. The resulted band structure completely
            matches the PWscf one but there is an error in transport
            section implies that there are 0 unit cells and Wfs inside
            the principal layer:<br>
            <br>
 *---------------------------------------------------------------------------*<br>
             |                             
            TRANSPORT                                    |<br>
 *---------------------------------------------------------------------------*<br>
            <br>
            <br>
             Calculation of Quantum Conductance and DoS: bulk mode<br>
            <br>
             Maximum imaginary part of the real-space Hamiltonian:    
            0.189159<br>
            <br>
 ------------------------------------------------------------------------------<br>
                 Maximum real part of the real-space Hamiltonian at each
            lattice point<br>
                    
            --------------------------------------------------------------<br>
                        Lattice point R           Max |H_ij(R)|<br>
                              -1                     0.000000<br>
                               0                     7.773991<br>
                               1                     0.000000<br>
                    
            --------------------------------------------------------------<br>
            <br>
             Number of unit cells inside the principal layer:     0<br>
             Number of Wannier Functions inside the principal layer:    
            0<br>
            <br>
             Calculating quantum conductance and density of states...<br>
             ERROR:  IN ZGESV IN tran_transfer, INFO=          -4<br>
             Exiting.......<br>
             tran_transfer: problem in ZGESV 1<br>
            <br>
          </div>
          <br>
          <br>
          in/output files are attached. Any help would be greatly
          appreciated.<br>
        </div>
        I have a stupid question too, What are the satisfying values for
        the imaginary part and spread of WFs?<br>
        <br>
        <div><br>
        </div>
      </div>
      <div class="gmail_extra"><br>
        <div class="gmail_quote">On Thu, Jun 25, 2015 at 11:47 PM,
          Nicola Marzari <span dir="ltr"><<a href="mailto:nicola.marzari@epfl.ch" target="_blank">nicola.marzari@epfl.ch</a>></span>
          wrote:<br>
          <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><br>
            <br>
            Dear Seyed,<br>
            <br>
            <br>
            any news on this?<span><font color="#888888"><br>
                <br>
                                        nicola</font></span>
            <div>
              <div><br>
                <br>
                <br>
                <br>
                On 15/06/2015 14:17, Nicola Marzari wrote:<br>
                <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
                  <br>
                  Thanks.<br>
                  <br>
                  At first sight the two espresso codes give the same
                  eigenvalues in the<br>
                  nscf,<br>
                  but being run with two different parallelization
                  choices, the<br>
                  phases of the wannier functions will all be different,
                  so<br>
                  it's like dealing with two independent calculations.
                  So the question is<br>
                  for Wannier - to see if you have converged to the
                  correct localized state,<br>
                  or not. Usually, if the projections are correct, it
                  shouldn't matter, but<br>
                  maybe in this case they are not correct, and hten the
                  codes randomly<br>
                  converges, or not.<br>
                  <br>
                  One of the outputs records a large imaginary part -
                  that's a bad sign.<br>
                  The other has two much less localized wavefunctions -
                  maybe good, maybe<br>
                  bad.<br>
                  <br>
                  Can you try with projections that have a s orbital at
                  every atom, and<br>
                  a s orbital at every bond? Those should work, and it
                  would be nice<br>
                  to see that both codes give the same results.<br>
                  <br>
                  Otherwise, we'll need to dig deeper.<br>
                  <br>
                               nicola<br>
                  <br>
                  <br>
                  <br>
                  <br>
                  On 15/06/2015 13:41, Seyed Mojtaba Rezaei Sani wrote:<br>
                  <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
                    Hi Nicola<br>
                    <br>
                    Thanks a lot for your kind response. Surely I did
                    checked outputs before<br>
                    too. There was no considerable difference between
                    them and also there is<br>
                    no unconverged band message. So I can say:<br>
                    <br>
                    Re 1) Every thing works perfectly both in 5.1 and
                    5.1.2 versions except<br>
                    the disappointing obtained wannier90 band structure
                    by the later.<br>
                    Re 2) No difference.<br>
                    <br>
                    As more details, please find all in/output files for
                    both versions<br>
                    (except UNK ones which are too big) here:<br>
                    <br>
                    <a href="https://www.dropbox.com/sh/xhd3etb6zfz2e7c/AAClTpe-aGkwXVrzDcIq6L7Aa?dl=0" rel="noreferrer" target="_blank">https://www.dropbox.com/sh/xhd3etb6zfz2e7c/AAClTpe-aGkwXVrzDcIq6L7Aa?dl=0</a><br>
                    <br>
                    Thanks again,<br>
                    <br>
                    On Mon, Jun 15, 2015 at 12:23 PM, Nicola Marzari
                    <<a href="mailto:nicola.marzari@epfl.ch" target="_blank">nicola.marzari@epfl.ch</a><br>
                    <mailto:<a href="mailto:nicola.marzari@epfl.ch" target="_blank">nicola.marzari@epfl.ch</a>>>
                    wrote:<br>
                    <br>
                    <br>
                    <br>
                        Hi Sani,<br>
                    <br>
                    <br>
                        I had mentioned in my email that you should
                    check both<br>
                        your input and your output files - since the
                    problems seem due<br>
                        to the change in version of quantum-espresso (I
                    do appreciate you<br>
                        did check your input files, but I do wonder why
                    you didn't<br>
                        read the email in full, since it was asking
                    about the output).<br>
                    <br>
                        So - do you see any difference in the output
                    files of quantum<br>
                    espresso?<br>
                    <br>
                         From your plots my first guess would be that
                    you have some states<br>
                    that<br>
                        are not converged at some k-point.<br>
                    <br>
                        1) can you confirm that in an older version of
                    QE everything<br>
                        worked perfectly (*which one*?, and in 5.1.2 you
                    have the problems<br>
                        below?<br>
                    <br>
                        2) do the outputs in the two cases differ - for
                    the QE runs, and<br>
                    for the<br>
                        wannier runs, when run with *identical* inputs?
                    If so, in what do<br>
                        they differ.<br>
                    <br>
                        Let us know re 1) and 2) - thanks a lot,<br>
                    <br>
                                                 nicola<br>
                    <br>
                    <br>
                    <br>
                        On 15/06/2015 09:40, Seyed Mojtaba Rezaei Sani
                    wrote:<br>
                    <br>
                            Dear all<br>
                    <br>
                            I am looking forward to receive any kind
                    help.<br>
                    <br>
                            On Wed, Jun 10, 2015 at 2:59 PM, Seyed
                    Mojtaba Rezaei Sani<br>
                            <<a href="mailto:s.m.rezaeisani@gmail.com" target="_blank">s.m.rezaeisani@gmail.com</a>
                    <mailto:<a href="mailto:s.m.rezaeisani@gmail.com" target="_blank">s.m.rezaeisani@gmail.com</a>><br>
                            <mailto:<a href="mailto:s.m.rezaeisani@gmail.com" target="_blank">s.m.rezaeisani@gmail.com</a><br>
                            <mailto:<a href="mailto:s.m.rezaeisani@gmail.com" target="_blank">s.m.rezaeisani@gmail.com</a>>>>
                    wrote:<br>
                    <br>
                                 Dear Prof. Marzari<br>
                    <br>
                                 As you suggested I checked input files
                    and they were exactly<br>
                                 identical. but the difference in result
                    remains.<br>
                                 I tried to make a change in
                    calculations so I increased<br>
                            number of<br>
                                 kpoints in non scf and wannier
                    calculations from<br>
                                 a mesh of 12*12*1 to 16*16*1. The
                    result was an improvement<br>
                            in band<br>
                                 structure. you can find the plot
                    attached. But<br>
                                 there are still some inadequacies to
                    the pw band in the<br>
                            region gamma<br>
                                 to X.<br>
                                 Again I increased the grid to a 24*24*1
                    mesh which<br>
                            encountered this<br>
                                 error:<br>
                    <br>
                                 Unable to satisfy B1 with any of the
                    first  12 shells<br>
                                   Your cell might be very long, or you
                    may have an<br>
                            irregular MP grid<br>
                                   Try increasing the parameter
                    search_shells in the win<br>
                            file (default=12)<br>
                    <br>
                                 Setting the parameter search_shells =
                    60 solved the problem<br>
                            but now<br>
                                 inadequacies are in the region S to Y
                    (the plot is<br>
                    attached).<br>
                                 I don't know if there is any parameter
                    that I could change<br>
                            to solve<br>
                                 this problem. Please give me any
                    probable hint.<br>
                                 Thanks,<br>
                    <br>
                                 On Wed, Jun 10, 2015 at 2:53 PM, Seyed
                    Mojtaba Rezaei Sani<br>
                                 <<a href="mailto:s.m.rezaeisani@gmail.com" target="_blank">s.m.rezaeisani@gmail.com</a>
                    <mailto:<a href="mailto:s.m.rezaeisani@gmail.com" target="_blank">s.m.rezaeisani@gmail.com</a>><br>
                            <mailto:<a href="mailto:s.m.rezaeisani@gmail.com" target="_blank">s.m.rezaeisani@gmail.com</a><br>
                            <mailto:<a href="mailto:s.m.rezaeisani@gmail.com" target="_blank">s.m.rezaeisani@gmail.com</a>>>>
                    wrote:<br>
                    <br>
                                     Dear Prof. Marzari<br>
                    <br>
                                     As you suggested I checked input
                    files and they were<br>
                            exactly<br>
                                     identical. but the difference in
                    result remains.<br>
                                     I tried to make a change in
                    calculations so I increased<br>
                            number<br>
                                     of kpoints in non scf and wannier
                    calculations from<br>
                                     a mesh of 12*12*1 to 16*16*1. The
                    result was an<br>
                            improvement in<br>
                                     band structure. you can find the
                    plot attached. But<br>
                                     there are still some inadequacies
                    to the pw band in the<br>
                            region<br>
                                     gamma to X.<br>
                                     Again I increased the grid to a
                    24*24*1 mesh which<br>
                            encountered<br>
                                     this error:<br>
                    <br>
                                     Unable to satisfy B1 with any of
                    the first  12 shells<br>
                                       Your cell might be very long, or
                    you may have an<br>
                            irregular MP grid<br>
                                       Try increasing the parameter
                    search_shells in the win<br>
                            file (default=12)<br>
                    <br>
                                     Setting the parameter search_shells
                    = 60 solved the<br>
                            problem but<br>
                                     now inadequacies are in the region
                    S to Y (the plot is<br>
                            attached).<br>
                                     I don't know if there is any
                    parameter that I could<br>
                            change to<br>
                                     solve this problem. Please give me
                    any probable hint.<br>
                                     Thanks,<br>
                    <br>
                                     On Thu, May 28, 2015 at 12:39 PM,
                    Nicola Marzari<br>
                                     <<a href="mailto:nicola.marzari@epfl.ch" target="_blank">nicola.marzari@epfl.ch</a>
                    <mailto:<a href="mailto:nicola.marzari@epfl.ch" target="_blank">nicola.marzari@epfl.ch</a>><br>
                            <mailto:<a href="mailto:nicola.marzari@epfl.ch" target="_blank">nicola.marzari@epfl.ch</a>
                    <mailto:<a href="mailto:nicola.marzari@epfl.ch" target="_blank">nicola.marzari@epfl.ch</a>>>><br>
                            wrote:<br>
                    <br>
                    <br>
                    <br>
                                         Well, the obvious thing to do
                    is making sure your<br>
                                         inputs are exactly identical,
                    you re-run the<br>
                            calculations<br>
                                         to make sure the error shows up
                    again, and then you<br>
                            start<br>
                                         doing a diff in the output
                    files of<br>
                            quantum-espresso, and<br>
                                         see where there is a
                    difference.<br>
                    <br>
                                                                       
                      nicola<br>
                    <br>
                    <br>
                    <br>
                    <br>
                                         On 28/05/2015 09:07, Seyed
                    Mojtaba Rezaei Sani<br>
                    wrote:<br>
                    <br>
                                             Dear Wannier90 users,<br>
                    <br>
                                             I am trying to calculate
                    wannier functions for<br>
                                             phosphorene using<br>
                                             wannier90. first I tried to
                    plot band structure<br>
                            using both<br>
                                             quantum-espresso and
                    wannier90. Results were in<br>
                            good<br>
                                             agreement. Now I<br>
                                             have upgraded espresso to
                    5.1.2 version but<br>
                            with the<br>
                                             same inputs the<br>
                                             obtained results are not
                    similar.<br>
                                             All files are attached and
                    here in .win input:<br>
                    <br>
                                             num_bands        = 10<br>
                                             num_wann         = 10<br>
                                             num_iter         = 500<br>
                    <br>
                                             Begin Unit_Cell_Cart<br>
                                             Bohr<br>
                                                   6.282237122   
                     0.000000000<br>
                    0.0000000000<br>
                                                   0.000000000   
                     8.651074073<br>
                    0.0000000000<br>
                                                   0.000000000   
                     0.000000000<br>
                    26.456165734<br>
                                             End Unit_Cell_Cart<br>
                    <br>
                                             Begin Atoms_Frac<br>
                                             P1        0.010000000 
                     0.086656382<br>
                    0.152592466<br>
                                             P1        0.010000000 
                     0.913343530<br>
                    0.001000000<br>
                                             P2        0.510000000 
                     0.413343618<br>
                    0.152592466<br>
                                             P2        0.510000000 
                     0.586656470<br>
                    0.001000000<br>
                                             End Atoms_Frac<br>
                    <br>
                                             Begin Projections<br>
                                             P1:sp3<br>
                                             P2:py<br>
                                             End Projections<br>
                    <br>
                                             bands_plot = true<br>
                    <br>
                                             begin kpoint_path<br>
                                             G 0.0000 0.0000 0.0000 X
                    0.5000 0.0000 0.0000<br>
                                             x 0.5000 0.0000 0.0000 S
                    0.5000 0.5000 0.0000<br>
                                             S 0.5000 0.5000 0.0000 Y
                    0.0000 0.5000 0.0000<br>
                                             Y 0.0000 0.5000 0.0000 G
                    0.0000 0.0000 0.0000<br>
                                             end kpoint_path<br>
                    <br>
                                             mp_grid = 12 12 1<br>
                    <br>
                                             begin kpoints<br>
                                                 0.00000000  0.00000000 
                    0.00000000<br>
                                                .<br>
                                                .<br>
                                                .<br>
                                                 0.91666667  0.91666667 
                    0.00000000<br>
                                             End Kpoints<br>
                    <br>
                                             I would greatly appreciate
                    it if you kindly<br>
                            give me some<br>
                                             hints.<br>
                    <br>
                                             Best regards,<br>
                    <br>
                                             --<br>
                                             Seyed Mojtaba Rezaei Sani<br>
                    <br>
                                             Institute for Research in
                    Fundamental Sciences<br>
                            (IPM)<br>
                                             School of Nano-Science<br>
                                             Shahid Farbin Alley<br>
                                             Shahid Lavasani st<br>
                                             P.O. Box 19395-5531<br>
                                             Tehran, Iran<br>
                                             Tel: +98 21 2310  (3069)<br>
                    <br>
                    <br>
                                           
                     _______________________________________________<br>
                                             Wannier mailing list<br>
                            <a href="mailto:Wannier@quantum-espresso.org" target="_blank">Wannier@quantum-espresso.org</a><br>
                    <mailto:<a href="mailto:Wannier@quantum-espresso.org" target="_blank">Wannier@quantum-espresso.org</a>><br>
                                             <mailto:<a href="mailto:Wannier@quantum-espresso.org" target="_blank">Wannier@quantum-espresso.org</a><br>
                            <mailto:<a href="mailto:Wannier@quantum-espresso.org" target="_blank">Wannier@quantum-espresso.org</a>>><br>
                    <br>
                            <a href="http://mailman.qe-forge.org/cgi-bin/mailman/listinfo/wannier" rel="noreferrer" target="_blank">http://mailman.qe-forge.org/cgi-bin/mailman/listinfo/wannier</a><br>
                    <br>
                    <br>
                                         --<br>
                    <br>
                    <br>
                    <br>
----------------------------------------------------------------------<br>
                                         Prof Nicola Marzari, Chair of
                    Theory and<br>
                    Simulation of<br>
                                         Materials, EPFL<br>
                                         Director, National Centre for
                    Competence in<br>
                            Research NCCR<br>
                                         MARVEL, EPFL<br>
                    <br>
                    <br>
                    <br>
                    <br>
                                     --<br>
                                     Seyed Mojtaba Rezaei Sani<br>
                    <br>
                                     Institute for Research in
                    Fundamental Sciences (IPM)<br>
                                     School of Nano-Science<br>
                                     Shahid Farbin Alley<br>
                                     Shahid Lavasani st<br>
                                     P.O. Box 19395-5531<br>
                                     Tehran, Iran<br>
                                     Tel: +98 21 2310  (3069)<br>
                    <br>
                    <br>
                    <br>
                    <br>
                                 --<br>
                                 Seyed Mojtaba Rezaei Sani<br>
                    <br>
                                 Institute for Research in Fundamental
                    Sciences (IPM)<br>
                                 School of Nano-Science<br>
                                 Shahid Farbin Alley<br>
                                 Shahid Lavasani st<br>
                                 P.O. Box 19395-5531<br>
                                 Tehran, Iran<br>
                                 Tel: +98 21 2310  (3069)<br>
                    <br>
                    <br>
                    <br>
                    <br>
                            --<br>
                            Seyed Mojtaba Rezaei Sani<br>
                    <br>
                            Institute for Research in Fundamental
                    Sciences (IPM)<br>
                            School of Nano-Science<br>
                            Shahid Farbin Alley<br>
                            Shahid Lavasani st<br>
                            P.O. Box 19395-5531<br>
                            Tehran, Iran<br>
                            Tel: +98 21 2310  (3069)<br>
                    <br>
                    <br>
                        --<br>
                    <br>
                    <br>
----------------------------------------------------------------------<br>
                        Prof Nicola Marzari, Chair of Theory and
                    Simulation of Materials,<br>
                    EPFL<br>
                        Director, National Centre for Competence in
                    Research NCCR MARVEL,<br>
                    EPFL<br>
                    <br>
                    <br>
                    <br>
                    <br>
                    --<br>
                    Seyed Mojtaba Rezaei Sani<br>
                    <br>
                    Institute for Research in Fundamental Sciences (IPM)<br>
                    School of Nano-Science<br>
                    Shahid Farbin Alley<br>
                    Shahid Lavasani st<br>
                    P.O. Box 19395-5531<br>
                    Tehran, Iran<br>
                    Tel: +98 21 2310  (3069)<br>
                  </blockquote>
                  <br>
                </blockquote>
                <br>
                -- <br>
                <br>
----------------------------------------------------------------------<br>
                Prof Nicola Marzari, Chair of Theory and Simulation of
                Materials, EPFL<br>
                Director, National Centre for Competence in Research
                NCCR MARVEL, EPFL<br>
              </div>
            </div>
          </blockquote>
        </div>
        <br>
        <br clear="all">
        <br>
        -- <br>
        <div>
          <div dir="ltr"><span><font color="#888888"><span style="font-family:arial,helvetica,sans-serif">Seyed
                  Mojtaba Rezaei Sani<br>
                  <br>
                  Institute for Research in Fundamental Sciences (IPM)<br>
                  School of Nano-Science<br>
                  <font><span>Shahid Farbin Alley<br>
                      Shahid Lavasani st</span></font><br>
                  <font><span><font><span>P.O. Box 19395-5531<br>
                        </span></font>Tehran, Iran<br>
                    </span></font>Tel: +98 21 2310  (3069)<br>
                </span></font></span></div>
        </div>
      </div>
      <br>
      <fieldset></fieldset>
      <br>
      </div></div><pre><div><div class="h5">_______________________________________________
Wannier mailing list
<a href="mailto:Wannier@quantum-espresso.org" target="_blank">Wannier@quantum-espresso.org</a>
</div></div><a href="http://mailman.qe-forge.org/cgi-bin/mailman/listinfo/wannier" target="_blank">http://mailman.qe-forge.org/cgi-bin/mailman/listinfo/wannier</a>
</pre>
    </blockquote>
    <br>
  </div>

<br>_______________________________________________<br>
Wannier mailing list<br>
<a href="mailto:Wannier@quantum-espresso.org">Wannier@quantum-espresso.org</a><br>
<a href="http://mailman.qe-forge.org/cgi-bin/mailman/listinfo/wannier" rel="noreferrer" target="_blank">http://mailman.qe-forge.org/cgi-bin/mailman/listinfo/wannier</a><br>
<br></blockquote></div><br><br clear="all"><br>-- <br><div class="gmail_signature"><div dir="ltr"><span><font color="#888888"><span style="font-family:arial,helvetica,sans-serif">Seyed Mojtaba Rezaei Sani<br><br>Institute for Research in Fundamental Sciences (IPM)<br>
School of Nano-Science<br><font><span>Shahid Farbin Alley<br>
                Shahid Lavasani st</span></font><br><font><span><font><span>P.O. Box 19395-5531<br></span></font>Tehran, Iran<br></span></font>Tel: +98 21 2310  (3069)<br></span></font></span></div></div>
</div>