<html><head></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><div><br></div>Dear Christopher,<div><br><div><div>Il giorno 05/mar/2011, alle ore 04.59, Christopher O'Brien ha scritto:</div><br class="Apple-interchange-newline"><blockquote type="cite"><div>Thank you for your suggestions regarding my problem. In my case, a further decrease in the electronic convergence threshold, or increase in ecutrho, and ecutwfc had little effect. I usually work with supercells or NEB calculations with little or no symmetry.<br>It turns out that the parameter that affected the most significant reduction in force was the "upscale" parameter. Your recommended value of 1.D3 was usually sufficient, but really difficult cases can be handled with 1.D4. This parameter seemed to be the 'magic bullet' for me when nothing else worked. Why is the default value so small (10.0)?<br></div></blockquote><div><br></div><div>There's really no magic bullet, it is just a straightforward method to save computer time. You could have chosen conv_thr=1.d-9 and upscale=10, or conv_thr=1.d-6 and upscale=1.d4, you would end up in both cases with an electronic convergence threshold of 1.d-10 when the optimization procedure requires that (i.e. close to the convergence, when forces are small and energy changes very little, because you need very accurate forces to locate the minimum with a good precision). The point is that at the beginning of the relaxation your system is usually far from the optimum ionic configuration, hence it is not worth to spend a lot of cpu time in computing very accurate forces. </div><div><br></div><div>Defaults are often good, but cannot encompass all possible needs of the users. </div><div><br></div><div><br></div><div>Regards</div><div><br></div><div>GS</div><div><br></div><br><blockquote type="cite"><div>===================================================================<br>Christopher J. O'Brien<br><a href="mailto:cjobrien@ncsu.edu">cjobrien@ncsu.edu</a><br>https://sites.google.com/a/ncsu.edu/cjobrien/<br><br>Ph.D. Candidate<br>Computational Materials Group<br>Department of Materials Science & Engineering<br>North Carolina State University<br>__________________________________________________________________<br>Please send all documents in PDF. <br>For Word documents: Please use the 'Save as PDF' option before sending.<br>===================================================================<br><br>On Feb 24, 2011, at 4:08 AM, Giuseppe Mattioli wrote:<br><br><blockquote type="cite">Dear Christopher<br></blockquote><blockquote type="cite">You should try to reduce conv_thr in &ELECTRONS, in order to have a more accurate calculation of forces and stress. Try, for instance, with conv_thr = 1.0d-9 (and with a lower mixing_beta, 0.2~0.3 or so, if the electronic calculation does not converge).<br></blockquote><blockquote type="cite">HTH<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">Giuseppe<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">Quoting Christopher O'Brien <cjobrien@ncsu.edu>:<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"><blockquote type="cite">Below this is a super-cell of an interface of Cu2O and Ti. The  problem is simply that the forc_conv_thr is not met within 100  iterations. etot_conv_thr is almost always met first (even  decreasing etot_conv_thr to 1.0d-6). Typically, the force in this  system decreases to ~0.01 Ry/B with a delta F on the same order of  magnitude but fluctuates about this value. This means that I  continually get the warning telling me to reduce the convergence  threshold. Ideally I would like obtain a maximum force of ~1D-4.<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Does anyone have any suggestions? I am at a loss.<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Thanks in advance,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Chris O'Brien<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">title = 'Cu2O-Ti-case2' ,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">calculation = 'vc-relax' ,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">restart_mode = 'from_scratch' ,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">wf_collect = .false. ,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">outdir = './',<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">wfcdir = './',<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">pseudo_dir = '/share2/cjobrien/pseudo/',<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">prefix = 'Cu2OTi',<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">disk_io = 'low',<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">verbosity = 'minimal',<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">tstress = .true.,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">tprnfor = .true.,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">forc_conv_thr = 1.0d-4,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">etot_conv_thr = 1.0d-5,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">/<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">&SYSTEM<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">ibrav = 0,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">celldm(1) = 8.1878,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">nat = 14,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">ntyp = 3,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">ecutwfc = 30.0 ,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">ecutrho = 300.0 ,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">occupations = 'smearing' ,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">smearing = 'mv' ,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">degauss = 0.01 ,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">/<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">&ELECTRONS<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">conv_thr = 1.0d-6 ,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">mixing_beta = 0.7 ,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">/<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">&IONS<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">ion_dynamics = 'bfgs',<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">/<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">&CELL<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">cell_dynamics = 'bfgs',<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">press = 0.0,<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">cell_dofree = 'xyz',<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">/<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">CELL_PARAMETERS<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">0.929429709 0.0 0.0<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">0.0 0.929429709 0.0<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">0.0 0.0 2.924945006<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">ATOMIC_SPECIES<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Cu   58.70000  Cu.pw91-n-van_ak.UPF<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Ti   47.88000  Ti.pw91-nsp-van.UPF<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">O    16.00000  O.pw91-van_ak.UPF<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">ATOMIC_POSITIONS alat<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Cu      0       0       0<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Cu      0.504691013     0.504691013     0<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Cu      0.484702934     0.019988079     0.553031754<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Cu      0.019988079     0.484702934     0.553031754<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Cu      0.039976158     0.039976158     1.106063508<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Cu      0.464714854     0.464714854     1.106063508<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">O       0.252345506     0.252345506     0.285738983<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">O       0.024221288     0.024221288     1.544049139<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Ti      0.480469725     0.024221288     1.544049139<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Ti      0.024221288     0.480469725     1.544049139<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Ti      0.484702934     0.484702934     2.015504257<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Ti      0.019988079     0.019988079     2.015504257<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Ti      0.01575487      0.488936142     2.486959376<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Ti      0.488936142     0.01575487      2.486959376<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">K_POINTS automatic<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">8 8 4 0 0 0<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">===================================================================<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Christopher J. O'Brien<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">cjobrien@ncsu.edu<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">https://sites.google.com/a/ncsu.edu/cjobrien/<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Ph.D. Candidate<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Computational Materials Group<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Department of Materials Science & Engineering<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">North Carolina State University<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">__________________________________________________________________<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Please send all documents in PDF.<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">For Word documents: Please use the 'Save as PDF' option before sending.<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">===================================================================<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">_______________________________________________<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Pw_forum mailing list<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Pw_forum@pwscf.org<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">http://www.democritos.it/mailman/listinfo/pw_forum<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">Giuseppe Mattioli<br></blockquote><blockquote type="cite">ISM-CNR, Italy<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"><br></blockquote><br>_______________________________________________<br>Pw_forum mailing list<br>Pw_forum@pwscf.org<br>http://www.democritos.it/mailman/listinfo/pw_forum<br></div></blockquote></div><br><div>
<span class="Apple-style-span" style="border-collapse: separate; color: rgb(0, 0, 0); font-family: Helvetica; font-size: medium; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: 2; text-align: auto; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-border-horizontal-spacing: 0px; -webkit-border-vertical-spacing: 0px; -webkit-text-decorations-in-effect: none; -webkit-text-size-adjust: auto; -webkit-text-stroke-width: 0px; "><div><span class="Apple-style-span" style="color: rgb(126, 126, 126); font-size: 16px; font-style: italic; "><br class="Apple-interchange-newline">§ Gabriele Sclauzero, EPFL SB ITP CSEA</span></div><div><font class="Apple-style-span" color="#7E7E7E"><i>   PH H2 462, Station 3, CH-1015 Lausanne</i></font></div></span>
</div>
<br></div></body></html>