<div dir="ltr"><div><div><div><div><div><div><div><div>Dear Molpro<br><br></div>I am having some problems with SAPT(DFT) calculations in large systems using the new DFSAPT module in Molpro15.<br><br></div>I have been able to run a job with 2328 basis functions and 168 electrons without problem, just asking for MEMORY,600,M.<br><br></div>However,
 when I increase the size of the system to 2372 basis functions and 198 
electrons the job always stops  with a segmentation fault message, just 
before ending the calculation of exch-dispersion contribution. I have 
increased the memory to MEMORY,1200,M and even to MEMORY,7000,M with no 
result.<br></div>In all cases the program stops before ending the Exch-disp calculation:<br><br> E(2)exch-disp<br> =============<br> (ia|jb).Sib/Sja...           <wbr>                CPU time:   107.51<br> (P|ia).SSi transformed integrals...          CPU time:   490.12<br> (P|jb).SSj transformed integrals...          CPU time:   323.79<br> (P|ia).Sij transformed integrals...          CPU time:   436.02<br> (P|jb).Sji transformed integrals...          CPU time:   452.26<br> done:     2.74%    CPU time:  1692.69<br> done:     5.48%    CPU time:  1619.34<br> done:     8.22%    CPU time:  1626.85<br> done:    10.96%    CPU time:  1574.67<br> done:    13.70%    CPU time:  1584.20<br> done:    16.44%    CPU time:  1433.22<br> done:    19.18%    CPU time:  1337.00<br> done:    21.92%    CPU time:  1284.13<br> done:    24.66%    CPU time:  1530.77<br> done:    27.40%    CPU time:  1619.63<br> done:    30.14%    CPU time:  1577.61<br> done:    32.88%    CPU time:  1620.97<br> done:    35.62%    CPU time:  1580.66<br> done:    38.36%    CPU time:  1615.58<br> done:    41.10%    CPU time:  1563.07<br> done:    43.84%    CPU time:  1582.54<br> done:    46.58%    CPU time:  1400.51<br> done:    49.32%    CPU time:  1306.42<br> done:    52.05%    CPU time:  1561.26<br> done:    54.79%    CPU time:  1538.78<br> done:    57.53%    CPU time:  1567.66<br> done:    60.27%    CPU time:  1296.91<br> done:    63.01%    CPU time:  1271.76<br> done:    65.75%    CPU time:  1273.03<br> done:    68.49%    CPU time:  1272.98<br> done:    71.23%    CPU time:  1276.71<br> done:    73.97%    CPU time:  1277.00<br> done:    76.71%    CPU time:  1277.53<br> done:    79.45%    CPU time:  1278.63<br> done:    82.19%    CPU time:  1270.43<br> done:    84.93%    CPU time:  1214.99<br> done:    87.67%    CPU time:  1215.03<br> done:    90.41%    CPU time:  1217.58<br> done:    93.15%    CPU time:  1214.26<br> done:    95.89%    CPU time:  1216.07<br> done:    98.63%    CPU time:  1215.97<br><br><br></div>I
 have seen in JCP 141, 094107 (2014), that much larger calculations can 
be carried out, so maybe I am misinterpreting something. <br><br></div>This is my input file:<br><br><br>MEMORY,1200,M<br>gdirect; gthresh,THROVL=1.e-9<br>angstrom<br>geomtyp=xyz<br>symmetry,nosym orient,noorient<br>geometry={<br>include, <a href="http://ghexa-hexa.xyz" target="_blank">ghexa-hexa.xyz</a><br>}<br><br><br>basis={<br>set,orbital; default,vtz<br>set,jkfit;   default,vtz/jkfit<br>set,mp2fit;  default,vtz/mp2fit<br>set,dflhf;   default,vtz/jkfit<br>}<br><br><br>!=============================<wbr>==============================<wbr>=====<br>ca=2101.2; cb=2102.2<br>oldnorm=1<br>ipghexa=0.30481511<br>iphexa=0.21937674<br>hoghexa=-0.263564<br>hohexa=-0.177575<br>deltaA=ipghexa+hoghexa<br>deltaB=iphexa+hohexa<br>!=============================<wbr>==============================<wbr>=====<br><br>dummy,69,70,71,72,73,74,75,76,<wbr>77,78,79,80,81,82,83,84,85,86,<wbr>87,88,89,90,91,92,93,94,95,96,<wbr>97,98,99,100,101,102,103,104,<wbr>105,106,107,108,109,110<br>{df-ks,lhf,pbex,pbec,basis=<wbr>jkfit; dftfac,0.25,0.75,1.0; asymp,deltaA; save,$ca; start,atdens}<br>monomerA,orb=$ca,core=26<br><br>dummy,1,2,3,4,5,6,7,8,9,10,11,<wbr>12,13,14,15,16,17,18,19,20,21,<wbr>22,23,24,25,26,27,28,29,30,31,<wbr>32,33,34,35,36,37,38,39,40,41,<wbr>42,43,44,45,46,47,48,49,50,51,<wbr>52,53,54,55,56,57,58,59,60,61,<wbr>62,63,64,65,66,67,68<br>{df-ks,lhf,pbex,pbec,basis=<wbr>jkfit; dftfac,0.25,0.75,1.0; asymp,deltaB; save,$cb; start,atdens}<br>monomerB,orb=$cb,core=26<br><br>{dfsapt,all=1,auxbas='mp2fit',<wbr>fcoul=1.0,fxc=1.0,fxcfit=1,<wbr>fxcsing=1,t678=3}<br><br><br></div>Thanks,<br></div>Quique<br clear="all"><br>-- <br><div class="gmail_signature" data-smartmail="gmail_signature">Enrique M. Cabaleiro Lago<br>Departamento de Química Física<br>Facultad de Ciencias-Lugo<br>Universidad de Santiago de Compostela<br>Tel: 982295800 Ext. 24129<br>Fax: 982285872<br>e-mail: <a href="mailto:caba.lago@usc.es" target="_blank">caba.lago@usc.es</a><br></div></div>