<div dir="ltr"><br><div class="gmail_extra">Dear Daniel<br></div><div class="gmail_extra"><div class="gmail_quote"><br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><p>I said "surprisingly" huge because I compared to the VASP code with the same parameters (Ecutoff, NKpoints, Nq) and with the HSE functional. Indeed, with the VASP code during one night on my computer I got a few tens of iterations compared to one with QE :(.</p></div></blockquote><div><br></div><div>With exactly the same parameters the computational time (on one processor) should be comparable. There are two important points you should consider to make a comparison:<br></div><div>-Different type of methodology: Using EXX with Ultrasoft pseudopotentials a very large number of augmentation charges are computed. As already suggested setting tqr=.true. should solve this problem.<br></div><div>-If you changed the default values of some variables such as Precfock or Nkred in Vasp the computational parameters are not anymore the same in Vasp and QE. <br></div><div>Best,<br></div><div>Dario<br></div><div><br><br> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div>
<p>From my point of view, I compare 2 pseudopotential-based methods for the same parameters which I can control via the input file : my question would then more preferably be "which default parameter should I modify to make the QE calculation time similar to the VASP one ?"</p>
<p>Best wishes,</p>
<p>Daniel</p><div><div class="h5">
<p> </p>
<p> </p>
<div> </div>
<p>Le 2016-09-09 09:47, Stefano de Gironcoli a écrit :</p>
<blockquote type="cite" style="padding-left:5px;border-left:#1010ff 2px solid;margin-left:5px;width:100%">
<div>Please read the literature about hybrid functionals in plane waves; understand why the computational time is, not unexpectedly, huge and start think about it.</div>
<div> If you come up with some smart idea on how to reduce the computational time without compromising accuracy (as for the ACE transformation recently proposed by Lin Lin and implemented in the latest QE version) let us know<br>Best,<br>stefano
<div>(sent from my phone)</div>
</div>
<div><br>On 09 Sep 2016, at 09:30, Daniel Stoeffler <<a href="mailto:daniel.stoeffler@ipcms.unistra.fr" target="_blank">daniel.stoeffler@ipcms.<wbr>unistra.fr</a>> wrote:<br><br></div>
<blockquote type="cite" style="padding-left:5px;border-left:#1010ff 2px solid;margin-left:5px;width:100%">
<div>
<p> </p>
<p>Hello,</p>
<p>As a new QE user, I do some tests with the HSE Functional for BCC ferromagnetic Fe with a cell containing 2 atoms (because a want to go to CoFe after that).</p>
<p>My input file is :</p>
<p> &control<br> title='Fe CC HSE',<br> calculation = 'scf',<br> prefix = 'Fe_CC_HSE',<br> pseudo_dir='./',<br> verbosity = 'high'<br> /<br> &system<br> ibrav = 6, <br> celldm(1) = 5.45, <br> celldm(3) = 1.0,<br> nat = 2, <br> ntyp = 1,<br> nspin = 2,<br> occupations='smearing', smearing='gauss', degauss=0.01,<br> starting_magnetization(1) = 1.0,<br> ecutwfc = 33,<br> input_dft='HSE', <br> report = 5<br> /<br> &electrons<br> mixing_beta = 0.7<br> /<br><br>ATOMIC_SPECIES<br> Fe 55.845 Fe.pbe-sp-van.UPF<br> <br>ATOMIC_POSITIONS (crystal)<br>Fe 0.000 0.000 0.000<br>Fe 0.500 0.500 0.500<br><br>K_POINTS (automatic)<br> 7 7 7 0 0 0<br><br></p>
<p>Surprisingly, the computation time is huge as writen into the output file :</p>
<p>==============================<wbr>=</p>
<p> iteration # 9 ecut= 33.00 Ry beta=0.70<br> Davidson diagonalization with overlap<br> ethr = 5.49E-08, avg # of iterations = 2.3<br> <br> Magnetic moment per site:<br> atom: 1 charge: 14.4350 magn: 2.6346 constr: 0.0000<br> atom: 2 charge: 14.4351 magn: 2.6343 constr: 0.0000<br><br> total cpu time spent up to now is 7.2 secs<br><br> End of self-consistent calculation<br><br></p>
<p> convergence has been achieved in 9 iterations<br><br> EXX: now go back to refine exchange calculation<br><br> total cpu time spent up to now is 4239.4 secs<br><br> Self-consistent Calculation<br><br> iteration # 1 ecut= 33.00 Ry beta=0.70<br> Davidson diagonalization with overlap<br> ethr = 5.49E-08, avg # of iterations = 4.9<br><br> total cpu time spent up to now is 52490.2 secs<br><br> iteration # 2 ecut= 33.00 Ry beta=0.70<br> Davidson diagonalization with overlap<br><br></p>
<p>==============================<wbr>=</p>
<p>So, the calculation takes 7 seconds for converging with PBE, 1h10min for calculating the EXX term and 14h35min for the first HSE iteration. Is this correct ? And if so, please, could you try to explain me how to reduce the computation time. Thanks in advance.</p>
<p>Best wishes,</p>
<p>Daniel</p>
<p> </p>
<div> </div>
</div>
</blockquote>
<blockquote type="cite" style="padding-left:5px;border-left:#1010ff 2px solid;margin-left:5px;width:100%">
<div><span>______________________________<wbr>_________________</span><br><span>Pw_forum mailing list</span><br><span><a href="mailto:Pw_forum@pwscf.org" target="_blank">Pw_forum@pwscf.org</a></span><br><span><a href="http://pwscf.org/mailman/listinfo/pw_forum" target="_blank">http://pwscf.org/mailman/<wbr>listinfo/pw_forum</a></span></div>
</blockquote>
<br>
<pre>______________________________<wbr>_________________
Pw_forum mailing list
<a href="mailto:Pw_forum@pwscf.org" target="_blank">Pw_forum@pwscf.org</a>
<a href="http://pwscf.org/mailman/listinfo/pw_forum" target="_blank">http://pwscf.org/mailman/<wbr>listinfo/pw_forum</a>
</pre>
</blockquote>
</div></div></div>
<br>______________________________<wbr>_________________<br>
Pw_forum mailing list<br>
<a href="mailto:Pw_forum@pwscf.org">Pw_forum@pwscf.org</a><br>
<a href="http://pwscf.org/mailman/listinfo/pw_forum" rel="noreferrer" target="_blank">http://pwscf.org/mailman/<wbr>listinfo/pw_forum</a><br></blockquote></div><br></div></div>