<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<p>Dear Pablo<br>
</p>
<p>Final Coordinates and Job Done do not mean that relaxation
reached convergence you may have reached the max number of steps
and relaxation stopped. <br>
</p>
<p>Check few lines before "Begin Final coordinates" you should find
something like "bfgs converged in <n> scf cycles" <br>
</p>
<p>Did you try to use the same relaxation run for the 2 different
first steps ? Do final results coincide in this case ?<br>
</p>
<br>
<div class="moz-cite-prefix">On 28/12/2016 19:01, Pablo García
Risueño wrote:<br>
</div>
<blockquote
cite="mid:CAL1uXO1hOdOAm2C-r_6ugcZuV4qrzZkND1uH=GU8H0Qx9PNq5w@mail.gmail.com"
type="cite">
<div dir="ltr">
<div>Dear Stefano<br>
<br>
</div>
Thank you very much for your reply. I answer to your questions
below.<br>
<div class="gmail_extra"><br>
<div class="gmail_quote">2016-12-28 16:29 GMT+01:00 Stefano de
Gironcoli <span dir="ltr"><<a moz-do-not-send="true"
href="mailto:degironc@sissa.it" target="_blank">degironc@sissa.it</a>></span>:<br>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">Dear
Pablo Garcia Risueño<br>
<br>
I'm not sure I understand the problem.<br>
the two final positions for ecut=30 are identical to
my eyes.<br>
the two final positions for ecut=60 differ by about
10-5 A. !<br>
If the property you are interested in depends so
strongly on the<br>
atomic positions you better learn how to live with it.
Even assuming<br>
that you are using perfect xc functional, perfect
pseudopotentials and<br>
converged cutoff and k-points (which likely you don't),
your<br>
calculation is neglecting zero point energy, thermal
expansion,<br>
quantum nature of H motion, ...<br>
<i><br>
</i></blockquote>
<div><i>Yes, my concern is not that the results are not
perfect (of course DFT makes non-negligible errors). My
concern is that fully identical input files run in the
same machine and with the same executable provide
different results.</i><br>
</div>
<div> </div>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">
As for your specific question. In principle if you
tighten enough<br>
the convergence the minima should tend to converge toward
a common<br>
value. However the stopping criteria will make the two
calculations<br>
stop as soon they are satisfied. In your case as soon as
each force<br>
component is lower than 10-6 Ry/au and the energy does not
change more<br>
that 10-8. Did the calculation complete ? 10-8 for the
energy is<br>
easily reachable but 10-6 for the forces looks to me very
demanding.<br>
<i><br>
</i></blockquote>
<div><i>Yes, the calculations do finish. I read 'Begin final
coordinates' (...) and 'JOB DONE' at the end of the
output files. My system is very small (just 26 atoms and
56 electrons).</i><br>
</div>
<div> </div>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">
Anyhow... even if the relaxation did complete
successfully two<br>
equivalent calculations may still differ. If you run your
calculations<br>
with exactly the same input on exactly the same machine
with exactly<br>
the same mpi setup the results should be identical </blockquote>
<div><i>This is exactly what I do.</i><br>
<br>
</div>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">..but
why doing<br>
twice exactly the same calculation ?. </blockquote>
<div><i>The 'relax' calculation is the first step of a
complicated process. I run two different calculations
because the 6th step of the process is different. I
obtained very different results at the end of the full
process; then I tried to trace back the differences. And
then I saw that despite the fact that the 1st step is
identical for two different runs, the results of the 1st
step are different. This made me think that there might
be a bug, or a wrong parameter in my input files.</i><br>
</div>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">If
however you change anything<br>
in the input (Davidson vs CG, starting wavefunction
options, mixing<br>
mode, BFGS-related parameters) or parallel computational
setup the<br>
relaxation history will be slightly different and the
final<br>
configurations will differ by a certain amount allowed by
the<br>
tolerance defined by your stopping criteria.<br>
</blockquote>
<div><i>I changed nothing, this is why I am surprised with
the results...</i> <br>
</div>
<div><br>
<br>
</div>
<div>Thank you very much for your attention. Best regards.<br>
</div>
<div><br>
<br>
</div>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">
HTH<br>
<span class="HOEnZb"><font color="#888888"><br>
stefano<br>
</font></span>
<div class="HOEnZb">
<div class="h5"><br>
Quoting Pablo García Risueño <<a
moz-do-not-send="true"
href="mailto:garcia.risueno@gmail.com"><a class="moz-txt-link-abbreviated" href="mailto:garcia.risueno@gmail.com">garcia.risueno@gmail.com</a></a>>:<br>
<br>
> Dear professor<br>
><br>
> Thank you very much for your reply. The
differences are important in this<br>
> case, my final calculated quantities are very
sensitive to these optimized<br>
> positions. Should I change any of the ***conv_thr
variables, or other<br>
> variable, to have the same result for same
inputs?<br>
><br>
> Is there any random number in the algorithm of
relax calculations, can we<br>
> be sure that different outputs with the same
input are not due to any bug?<br>
><br>
> Thank you very much. Best regards.<br>
><br>
> 2016-12-28 14:10 GMT+01:00 Paolo Giannozzi <<a
moz-do-not-send="true"
href="mailto:p.giannozzi@gmail.com"><a class="moz-txt-link-abbreviated" href="mailto:p.giannozzi@gmail.com">p.giannozzi@gmail.com</a></a>>:<br>
><br>
>> The differences you find are very small.
Nothing to worry about in my<br>
>> opinion.<br>
>><br>
>> By the way: Davidson diagonalization is
typically faster than CG; do not<br>
>> specify incompatible options in K_POINTS
(gamma or automatic, not both;<br>
>> gamma should be used unless you have a good
reason not to)<br>
>><br>
>> Paolo<br>
>><br>
>> Il 28/dic/2016 01:46 PM, "Pablo García
Risueño" <<a moz-do-not-send="true"
href="mailto:garcia.risueno@gmail.com">garcia.risueno@gmail.com</a>><br>
>> ha scritto:<br>
>> ><br>
>> > Dear Espresso community<br>
>> ><br>
>> > I have one problem that is important for
me; it is somewhat surprising.<br>
>> I run geometry optimization (relax)
calculations with pw.x with identical<br>
>> input files, and I obtain rather different
final coordinates. The problem<br>
>> does not happen if ecutwfc is 30, but it does
appear for cutoffs of 60, 80<br>
>> or 90. Below one can see the exact input
file, and examples of the<br>
>> difference between the final coordinates for
both runs (both run with<br>
>> identical input) for given cutoffs.<br>
>> ><br>
>> > Could anybody give me a clue on the
origin of the problem, and how to<br>
>> solve it?<br>
>> ><br>
>> > Thank you very much. Best regards.<br>
>> ><br>
>> ><br>
>> ><br>
>> > Input file:<br>
>> ><br>
>> ><br>
>> ><br>
>> > &CONTROL<br>
>> ><br>
>> > calculation = 'relax',<br>
>> ><br>
>> > restart_mode = 'from_scratch',<br>
>> ><br>
>> > prefix='',<br>
>> ><br>
>> > outdir = './',<br>
>> ><br>
>> > pseudo_dir = '/path_xxx/PP/',<br>
>> ><br>
>> > forc_conv_thr = 1.0D-6 ,<br>
>> ><br>
>> > etot_conv_thr = 1.0D-8 ,<br>
>> ><br>
>> > /<br>
>> ><br>
>> > &system<br>
>> ><br>
>> > ibrav = 0, a=18.0,<br>
>> ><br>
>> > nat= 26, ntyp= 2,<br>
>> ><br>
>> > ecutwfc = 30d0,<br>
>> ><br>
>> > nbnd = 100,<br>
>> ><br>
>> > /<br>
>> ><br>
>> > &electrons<br>
>> ><br>
>> > conv_thr = 1.0e-9,<br>
>> ><br>
>> > mixing_beta = 0.7,<br>
>> ><br>
>> > mixing_mode = 'plain',<br>
>> ><br>
>> > diagonalization = 'cg'<br>
>> ><br>
>> > /<br>
>> ><br>
>> > &IONS<br>
>> ><br>
>> > /<br>
>> ><br>
>> ><br>
>> ><br>
>> > ATOMIC_SPECIES<br>
>> ><br>
>> > C 12.0107 C.pz-vbc.UPF<br>
>> ><br>
>> > H 1.007825035 H.pz-vbc.UPF<br>
>> ><br>
>> ><br>
>> ><br>
>> > ATOMIC_POSITIONS { angstrom }<br>
>> ><br>
>> > C 8.891700e+00 8.891700e+00
8.891700e+00<br>
>> ><br>
>> > C 9.783400e+00 9.783400e+00
8.000000e+00<br>
>> ><br>
>> > C 9.783400e+00 8.000000e+00
9.783400e+00<br>
>> ><br>
>> > C 8.000000e+00 9.783400e+00
9.783400e+00<br>
>> ><br>
>> > C 8.891700e+00 1.067510e+01
1.067510e+01<br>
>> ><br>
>> > C 1.067510e+01 1.067510e+01
8.891700e+00<br>
>> ><br>
>> > C 1.067510e+01 8.891700e+00
1.067510e+01<br>
>> ><br>
>> > C 9.783400e+00 1.156680e+01
9.783400e+00<br>
>> ><br>
>> > C 9.783400e+00 9.783400e+00
1.156680e+01<br>
>> ><br>
>> > C 1.156680e+01 9.783400e+00
9.783400e+00<br>
>> ><br>
>> > H 8.391700e+00 8.391700e+00
8.391700e+00<br>
>> ><br>
>> > H 7.500000e+00 1.028340e+01
9.283400e+00<br>
>> ><br>
>> > H 7.500000e+00 9.283400e+00
1.028340e+01<br>
>> ><br>
>> > H 9.283400e+00 1.028340e+01
7.500000e+00<br>
>> ><br>
>> > H 1.028340e+01 9.283400e+00
7.500000e+00<br>
>> ><br>
>> > H 9.283400e+00 7.500000e+00
1.028340e+01<br>
>> ><br>
>> > H 1.028340e+01 7.500000e+00
9.283400e+00<br>
>> ><br>
>> > H 9.283400e+00 1.206680e+01
9.283400e+00<br>
>> ><br>
>> > H 1.206680e+01 9.283400e+00
9.283400e+00<br>
>> ><br>
>> > H 9.283400e+00 9.283400e+00
1.206680e+01<br>
>> ><br>
>> > H 8.391700e+00 1.117510e+01
1.117510e+01<br>
>> ><br>
>> > H 1.117510e+01 1.117510e+01
8.391700e+00<br>
>> ><br>
>> > H 1.117510e+01 8.391700e+00
1.117510e+01<br>
>> ><br>
>> > H 1.028340e+01 1.206680e+01
1.028340e+01<br>
>> ><br>
>> > H 1.028340e+01 1.028340e+01
1.206680e+01<br>
>> ><br>
>> > H 1.206680e+01 1.028340e+01
1.028340e+01<br>
>> ><br>
>> ><br>
>> ><br>
>> ><br>
>> ><br>
>> > CELL_PARAMETERS {cubic}<br>
>> ><br>
>> > 1.00 0.00 0.00<br>
>> ><br>
>> > 0.00 1.00 0.00<br>
>> ><br>
>> > 0.00 0.00 1.00<br>
>> ><br>
>> ><br>
>> ><br>
>> > K_POINTS {gamma} {automatic}<br>
>> ><br>
>> > 1 1 1 0 0 0<br>
>> ><br>
>> ><br>
>> ><br>
>> ><br>
>> ><br>
>> > The program is run with<br>
>> ><br>
>> ><br>
>> ><br>
>> > mpirun -np 32 /path_xxx/pw.x<br>
>> ><br>
>> ><br>
>> ><br>
>> > Final coordinates: The first two rows
for ecutwfc=30 for two runs with<br>
>> identical input are:<br>
>> ><br>
>> > ATOMIC_POSITIONS (angstrom)<br>
>> > C 8.898432121 8.898432121 8.898432121<br>
>> > C 9.783306562 9.783306562 8.019670107<br>
>> ><br>
>> > ATOMIC_POSITIONS (angstrom)<br>
>> > C 8.898432121 8.898432121 8.898432121<br>
>> > C 9.783306562 9.783306562 8.019670109<br>
>> ><br>
>> ><br>
>> > Final coordinates: The first two rows
for ecutwfc=60 for two runs with<br>
>> identical input are:<br>
>> ><br>
>> > ATOMIC_POSITIONS (angstrom)<br>
>> > C 8.904988579 8.904988579 8.904988579<br>
>> > C 9.783426086 9.783426086 8.031809869<br>
>> ><br>
>> > ATOMIC_POSITIONS (angstrom)<br>
>> > C 8.904962246 8.904962246 8.904962246<br>
>> > C 9.783425401 9.783425401 8.031847251<br>
>> ><br>
>> > --<br>
>> > --<br>
>> ><br>
>> > Dr. Pablo García Risueño<br>
>> ><br>
>> > Institut für Physikalische Chemie,
Universität Hamburg, Grindelallee<br>
>> 117, 20146 Hamburg<br>
>> ><br>
>> > Tel. +49 040 42 83 84 82 7<br>
>> ><br>
>> > ______________________________<wbr>_________________<br>
>> > Pw_forum mailing list<br>
>> > <a moz-do-not-send="true"
href="mailto:Pw_forum@pwscf.org">Pw_forum@pwscf.org</a><br>
>> > <a moz-do-not-send="true"
href="http://pwscf.org/mailman/listinfo/pw_forum"
rel="noreferrer" target="_blank">http://pwscf.org/mailman/<wbr>listinfo/pw_forum</a><br>
>><br>
>><br>
>> ______________________________<wbr>_________________<br>
>> Pw_forum mailing list<br>
>> <a moz-do-not-send="true"
href="mailto:Pw_forum@pwscf.org">Pw_forum@pwscf.org</a><br>
>> <a moz-do-not-send="true"
href="http://pwscf.org/mailman/listinfo/pw_forum"
rel="noreferrer" target="_blank">http://pwscf.org/mailman/<wbr>listinfo/pw_forum</a><br>
>><br>
><br>
><br>
><br>
> --<br>
> --<br>
><br>
> Dr. Pablo García Risueño<br>
><br>
> Institut für Physikalische Chemie, Universität
Hamburg, Grindelallee 117,<br>
> 20146 Hamburg<br>
><br>
> Tel. <a moz-do-not-send="true"
href="tel:%2B49%20040%2042%2083%2084%2082%207"
value="+4940428384827">+49 040 42 83 84 82 7</a><br>
<br>
<br>
<br>
______________________________<wbr>_________________<br>
Pw_forum mailing list<br>
<a moz-do-not-send="true"
href="mailto:Pw_forum@pwscf.org">Pw_forum@pwscf.org</a><br>
<a moz-do-not-send="true"
href="http://pwscf.org/mailman/listinfo/pw_forum"
rel="noreferrer" target="_blank">http://pwscf.org/mailman/<wbr>listinfo/pw_forum</a></div>
</div>
</blockquote>
</div>
<br>
<br clear="all">
<br>
-- <br>
<div class="gmail_signature" data-smartmail="gmail_signature">
<div dir="ltr">
<div>
<div dir="ltr">
<div>
<div dir="ltr">--<br>
<br>
Dr. Pablo García Risueño<br>
<br>
Institut für Physikalische Chemie, Universität
Hamburg, Grindelallee 117, 20146 Hamburg<br>
<br>
Tel. +49 040 42 83 84 82 7</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset>
<br>
<pre wrap="">_______________________________________________
Pw_forum mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Pw_forum@pwscf.org">Pw_forum@pwscf.org</a>
<a class="moz-txt-link-freetext" href="http://pwscf.org/mailman/listinfo/pw_forum">http://pwscf.org/mailman/listinfo/pw_forum</a></pre>
</blockquote>
<br>
</body>
</html>