<body><div id="__MailbirdStyleContent" style="font-size: 10pt;font-family: Arial;color: #000000">
I am using a single node with 64 gb memory with 32 cpus per task <div><br></div><div>I tried running the 32 atom calc again with 1x1x1 k-points and it got stuck as well.</div><div>But 2 atoms with 8x8x1 works.</div><div><br></div><div>$ top, I dont see the pw.x written explicitly.</div><div><br></div><div>Should I try letting pw.x run longer than an hour?</div><div class="mb_sig"></div><blockquote class="history_container" type="cite" style="border-left-style:solid;border-width:1px; margin-top:20px; margin-left:0px;padding-left:10px;">
<p style="color: #AAAAAA; margin-top: 10px;">On 5/9/2019 2:59:56 PM, Giuseppe Mattioli <giuseppe.mattioli@ism.cnr.it> wrote:</p><div style="font-family:Arial,Helvetica,sans-serif"><br>Hi<br>I don't know exactly why, because nothing seems clearly wrong in your <br>input file, but I see<br><br>> K_POINTS automatic<br>> 4 2 1 0 0 0<br><br>> total cpu time spent up to now is 1474.4 secs<br><br>Your calculation seems to be very slow, more than 20 minutes for the <br>first scf step, and the second step is often slower... And in your <br>previous calculation I saw<br><br>> K_POINTS automatic<br>> 8 8 1 0 0 0<br><br>that is going to be even slower. Are you sure pw.x hangs? If you type <br>$ top in a console (I suppose you use linux) do you see one active <br>pw.x process?<br>HTH<br>Giuseppe<br><br>Quoting Hud Wahab <hudwahab@gmail.com>:<br><br>> Hi<br>><br>> ok, even if I try with ibrav == 0, specify cell_parameters, and use <br>> cartesian coordinates for all as such:<br>><br>> &control<br>> calculation = 'scf'<br>> prefix = 'graphene'<br>> pseudo_dir = '/gscratch/hwahab/DFT-code/psp/' <br>> outdir = './'<br>> restart_mode = 'from_scratch'<br>> etot_conv_thr = 1.d-6<br>> forc_conv_thr = 1.d-5<br>> /<br>><br>> &system<br>> ibrav = 0<br>> nat = 48<br>> ntyp = 1<br>> ecutwfc = 80<br>> occupations = 'smearing'<br>> smearing = 'gaussian'<br>> degauss = 0.1<br>> vdw_corr='grimme-d2'<br>> /<br>><br>> &electrons<br>> diagonalization = 'david'<br>> diago_thr_init = 1.d-4<br>> mixing_mode = 'local-TF'<br>> mixing_beta = 0.7<br>> conv_thr = 1.d-8<br>> /<br>><br>> &ions<br>> <br>> /<br>><br>> ATOMIC_SPECIES<br>> C 12.0107 C.pbe-n-kjpaw_psl.1.0.0.UPF <br>><br>> ATOMIC_POSITIONS angstrom<br>> C -0.005594637 0.710607728 17.606217811<br>> C -0.007113039 2.128152101 17.584272182<br>> C 1.224800757 0.002792887 17.613037723<br>> C 1.220781849 2.833821539 17.553802163<br>> C 2.453370565 0.710361709 17.588200291<br>> C 2.452466724 2.129238290 17.551425829<br>> C 3.684434744 0.001530538 17.594476170<br>> C 3.684438920 2.839550201 17.533796214<br>> C 4.915544062 0.711333118 17.590036280<br>> C 4.915853909 2.129726547 17.558981625<br>> C 6.143486600 0.002227196 17.614492659<br>> C 6.144397239 2.838931198 17.567937591<br>> C -0.009386173 4.969207674 17.562655171<br>> C -0.007017413 6.394548846 17.533586100<br>> C 1.221221723 4.253471766 17.548052411<br>> C 1.198574699 7.115165393 17.592468497<br>> C 2.446729939 4.945634289 17.560357556<br>> C 2.429014083 6.377364466 17.843730988<br>> C 3.679896901 4.258356417 17.551643715<br>> C 3.704632453 7.115561848 17.867361404<br>> C 4.916336515 4.974358812 17.564482071<br>> C 4.935632583 6.385210475 17.597930230<br>> C 6.145763199 4.260558789 17.557531939<br>> C 6.139436097 7.104085900 17.546822753<br>> C -0.008057730 9.237466939 17.548881513<br>> C -0.005309676 10.658828114 17.540319093<br>> C 1.220546640 8.524488204 17.583337938<br>> C 1.225791434 11.368226350 17.566890173<br>> C 2.458810059 9.240431880 17.612862514<br>> C 2.454941023 10.659162890 17.592221246<br>> C 3.688725076 8.555854012 17.643635937<br>> C 3.683918063 11.368771580 17.603925777<br>> C 4.910904632 9.244786635 17.581661741<br>> C 4.913395819 10.664333371 17.573847419<br>> C 6.143923926 8.528771033 17.556446236<br>> C 6.142987294 11.370161820 17.555480374<br>> C -0.004309434 13.496944454 17.583281339<br>> C -0.004505934 14.915030012 17.613620426<br>> C 1.226582522 12.786684076 17.587667221<br>> C 1.224762481 15.623384029 17.618836753<br>> C 2.455170254 13.495545076 17.620810601<br>> C 2.455394258 14.913895553 17.626574694<br>> C 3.684071873 12.787455574 17.618041427<br>> C 3.684499812 15.622198165 17.627916140<br>> C 4.913552069 13.495405863 17.614905357<br>> C 4.913417489 14.913710743 17.621452935<br>> C 6.142676777 12.788081224 17.582175333<br>> C 6.143988489 15.623015254 17.616347215<br>><br>><br>> K_POINTS automatic<br>> 4 2 1 0 0 0<br>><br>> CELL_PARAMETERS angstrom<br>> 7.378073983 0.000000000 0.0000000000<br>> 0.000000000 17.038932 0.0000000000<br>> 0.000000000 0.000000000 25.00<br>><br>> I still get it stuck here in output: <br>><br>> Estimated max dynamical RAM per process > 9360.31MB<br>><br>> Check: negative/imaginary core charge= -0.000003 0.000000<br>><br>> Initial potential from superposition of free atoms<br>> Check: negative starting charge= -0.001323<br>><br>> starting charge 191.99800, renormalised to 192.00000<br>><br>> negative rho (up, down): 1.323E-03 0.000E+00<br>> Starting wfc are 192 randomized atomic wfcs<br>><br>> total cpu time spent up to now is 1474.4 secs<br>><br>> per-process dynamical memory: 1195.0 Mb<br>><br>> Self-consistent Calculation<br>><br>> iteration # 1 ecut= 80.00 Ry beta=0.70<br>> Davidson diagonalization with overlap<br>><br>><br>> This code has worked previously with qe/5.4.0, but gets stuck with <br>> 6.1 serial.<br>><br>> -Hud<br>> On 5/9/2019 1:23:39 PM, Giuseppe Mattioli <br>> <giuseppe.mattioli@ism.cnr.it> wrote:<br>><br>> You are mixing two different ways to indicate cell parameters<br>><br>> ///---<br>> EITHER:<br>><br>> +--------------------------------------------------------------------<br>> Variable: celldm(i), i=1,6<br>><br>> Type: REAL<br>> See: ibrav<br>> Description: Crystallographic constants - see the "ibrav" variable.<br>> Specify either these OR<br>> "A","B","C","cosAB","cosBC","cosAC" NOT both.<br>> Only needed values (depending on "ibrav") must<br>> be specified<br>> alat = "celldm"(1) is the lattice parameter "a"<br>> (in BOHR)<br>> If "ibrav"==0, only "celldm"(1) is used if present;<br>> cell vectors are read from card "CELL_PARAMETERS"<br>> +--------------------------------------------------------------------<br>><br>> OR:<br>><br>> +--------------------------------------------------------------------<br>> Variables: A, B, C, cosAB, cosAC, cosBC<br>><br>> Type: REAL<br>> See: ibrav<br>> Description: Traditional crystallographic constants:<br>><br>> a,b,c in ANGSTROM<br>> cosAB = cosine of the angle between axis a<br>> and b (gamma)<br>> cosAC = cosine of the angle between axis a<br>> and c (beta)<br>> cosBC = cosine of the angle between axis b<br>> and c (alpha)<br>><br>> The axis are chosen according to the value of<br>> @ref ibrav.<br>> Specify either these OR @ref celldm but NOT both.<br>> Only needed values (depending on @ref ibrav)<br>> must be specified.<br>><br>> The lattice parameter alat = A (in ANGSTROM ).<br>><br>> If @ref ibrav == 0, only A is used if present, and<br>> cell vectors are read from card @ref CELL_PARAMETERS.<br>> +--------------------------------------------------------------------<br>><br>> \\\---<br>><br>> This might be the cause of the strange behavior, supposing that your<br>> machine has the ~4GB of free RAM to perform the calculation indicated<br>> in the output. However, in the case of a regular hcp supercell you<br>> should not need at all to indicate the cosAB and cosAC values.<br>> HTH<br>> Giuseppe<br>><br>><br>> Quoting "H1@GMAIL" :<br>><br>>> Hi Giuseppe<br>>><br>>> apologies. My input file:<br>>><br>>> &control<br>>> calculation = 'scf'<br>>> prefix = 'graphene'<br>>> pseudo_dir = '/gscratch/hwahab/DFT-code/psp/'<br>>> outdir = './'<br>>> restart_mode = 'from_scratch'<br>>> etot_conv_thr = 1.d-6<br>>> forc_conv_thr = 1.d-5<br>>> /<br>>> &system<br>>> ibrav = 4<br>>> celldm(1) = 9.84<br>>> celldm(3) = 10<br>>> cosAB = -0.5<br>>> cosAC = 1<br>>> nat = 32<br>>> ntyp = 1<br>>> ecutwfc = 80<br>>> occupations = 'smearing'<br>>> smearing = 'gaussian'<br>>> degauss = 0.1<br>>> vdw_corr='grimme-d2'<br>>> /<br>>><br>>> &electrons<br>>> diagonalization = 'david'<br>>> diago_thr_init = 1.d-4<br>>> mixing_mode = 'local-TF'<br>>> mixing_beta = 0.7<br>>> conv_thr = 1.d-10<br>>> /<br>>><br>>> &ions<br>>> /<br>>><br>>> ATOMIC_SPECIES<br>>> C 12.0107 C.pbe-mt_gipaw.UPF<br>>><br>>> ATOMIC_POSITIONS crystal<br>>> C 0.16667 0.08333 0.00000<br>>> C 0.41667 0.08333 0.00000<br>>> C 0.66667 0.08333 0.00000<br>>> C 0.91667 0.08333 0.00000<br>>> C 0.08333 0.16667 0.00000<br>>> C 0.33333 0.16667 0.00000<br>>> C 0.58333 0.16667 0.00000<br>>> C 0.83333 0.16667 0.00000<br>>> C 0.16667 0.33333 0.00000<br>>> C 0.41667 0.33333 0.00000<br>>> C 0.66667 0.33333 0.00000<br>>> C 0.91667 0.33333 0.00000<br>>> C 0.08333 0.41667 0.00000<br>>> C 0.33333 0.41667 0.00000<br>>> C 0.58333 0.41667 0.00000<br>>> C 0.83333 0.41667 0.00000<br>>> C 0.16667 0.58333 0.00000<br>>> C 0.41667 0.58333 0.00000<br>>> C 0.66667 0.58333 0.00000<br>>> C 0.91667 0.58333 0.00000<br>>> C 0.08333 0.66667 0.00000<br>>> C 0.33333 0.66667 0.00000<br>>> C 0.58333 0.66667 0.00000<br>>> C 0.83333 0.66667 0.00000<br>>> C 0.16667 0.83333 0.00000<br>>> C 0.41667 0.83333 0.00000<br>>> C 0.66667 0.83333 0.00000<br>>> C 0.91667 0.83333 0.00000<br>>> C 0.08333 0.91667 0.00000<br>>> C 0.33333 0.91667 0.00000<br>>> C 0.58333 0.91667 0.00000<br>>> C 0.83333 0.91667 0.00000<br>>><br>>><br>>> K_POINTS automatic<br>>> 8 8 1 0 0 0<br>>><br>>> And the end snippet of the output:<br>>><br>>> Estimated max dynamical RAM per process > 3558.76MB<br>>><br>>> Initial potential from superposition of free atoms<br>>><br>>> starting charge 111.99996, renormalised to 128.00000<br>>><br>>> negative rho (up, down): 5.479E-05 0.000E+00<br>>> Starting wfc are 256 randomized atomic wfcs<br>>><br>>> There is no error outputs, it just gets stuck there..<br>>><br>>> Hope this makes sense.<br>>><br>>> Hud Wahab<br>>> University of Wyoming<br>>> 1000 E University Ave<br>>> Laramie WY, 82072<br>>> Email: hwahab@uwyo.edu<br>>> On 5/9/2019 12:50:50 PM, Giuseppe Mattioli<br>>> wrote:<br>>><br>>> Dear Hud (please sign always with full name and scientific affiliation<br>>> the posts to this forum, we appreciate it)<br>>> It is impossible to help you if you don't post the input of your<br>>> calculation and the relevant part of your output (where does the code<br>>> stop?). Is there any system error like a segfault printed, e.g., in a<br>>> nohup.out file? It is primarily important to look into such kind of<br>>> information, in order to see if the error is reproducible on different<br>>> machines/architectures or by using different compilers/libraries.<br>>> HTH<br>>> Giuseppe<br>>><br>>> Quoting "H1@GMAIL" :<br>>><br>>>> Hello<br>>>> I run on 6.1-serial. My pw.x scf runs ok with small size systems 2<br>>>> atoms, but nothing happens when scaled to 4x4 or 32 atoms. I let it<br>>>> run for more than an hour and don't expect that the calculation<br>>>> takes that long.<br>>>><br>>>> From the troubleshooting in User Guide I see it might be a<br>>>> floating-point error causing endless NaNs - how to handle for such<br>>>> exceptions?<br>>>><br>>>> As I can't provide the error output, I am not sure what details you<br>>>> need to troubleshoot, but let me know if something is missing<br>>>><br>>>> Cheers, Hud<br>>>> Dept. Chemical Engineering<br>>>> University of Wyoming<br>>><br>>><br>>><br>>> GIUSEPPE MATTIOLI<br>>> CNR - ISTITUTO DI STRUTTURA DELLA MATERIA<br>>> Via Salaria Km 29,300 - C.P. 10<br>>> I-00015 - Monterotondo Scalo (RM)<br>>> Mob (*preferred*) +39 373 7305625<br>>> Tel + 39 06 90672342 - Fax +39 06 90672316<br>>> E-mail:<br>>><br>>> _______________________________________________<br>>> Quantum Espresso is supported by MaX (www.max-centre.eu/quantum-espresso)<br>>> users mailing list users@lists.quantum-espresso.org<br>>> https://lists.quantum-espresso.org/mailman/listinfo/users<br>><br>><br>><br>> GIUSEPPE MATTIOLI<br>> CNR - ISTITUTO DI STRUTTURA DELLA MATERIA<br>> Via Salaria Km 29,300 - C.P. 10<br>> I-00015 - Monterotondo Scalo (RM)<br>> Mob (*preferred*) +39 373 7305625<br>> Tel + 39 06 90672342 - Fax +39 06 90672316<br>> E-mail:<br>><br>> _______________________________________________<br>> Quantum Espresso is supported by MaX (www.max-centre.eu/quantum-espresso)<br>> users mailing list users@lists.quantum-espresso.org<br>> https://lists.quantum-espresso.org/mailman/listinfo/users<br><br><br><br>GIUSEPPE MATTIOLI<br>CNR - ISTITUTO DI STRUTTURA DELLA MATERIA<br>Via Salaria Km 29,300 - C.P. 10<br>I-00015 - Monterotondo Scalo (RM)<br>Mob (*preferred*) +39 373 7305625<br>Tel + 39 06 90672342 - Fax +39 06 90672316<br>E-mail: <giuseppe.mattioli@ism.cnr.it><br><br>_______________________________________________<br>Quantum Espresso is supported by MaX (www.max-centre.eu/quantum-espresso)<br>users mailing list users@lists.quantum-espresso.org<br>https://lists.quantum-espresso.org/mailman/listinfo/users</giuseppe.mattioli@ism.cnr.it></giuseppe.mattioli@ism.cnr.it></hudwahab@gmail.com></div></blockquote>
</div></body>