<html><head><meta http-equiv="Content-Type" content="text/html; charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">So I tried:<div class=""><div class=""><br class=""></div><div class="">export KMP_STACKSIZE=512M</div><div class="">rcohen@tomcat3:~/CP2K/cp2k$ ./tools/regtesting/do_regtest -arch Linux-x86-64-intel -version psmp -restrictdir QS/regtest-dft-vdw-corr-1/ -restrictdir QS/regtest-dft-vdw-corr-2/ -restrictdir QS/regtest-dft-vdw-corr-3/ -restrictdir QS/regtest-dft-vdw-corr-3/ -nobuild -mpiranks 4 -ompthreads 4 -maxtasks 16 |& tee testwith512MKMP_STACKSIZE.out &</div><div class="">and I still get:</div><div class=""><br class=""></div><div class=""><div class="">< /home/rcohen/CP2K/cp2k/TEST-Linux-x86-64-intel-psmp-2019-01-07_18-24-16/tests/QS/regtest-dft-vdw-corr-3 (1 of 3) done in 775.00 sec</div><div class="">>>> /home/rcohen/CP2K/cp2k/TEST-Linux-x86-64-intel-psmp-2019-01-07_18-24-16/tests/QS/regtest-dft-vdw-corr-3</div><div class="">    argon05.inp                                               -85.02462435591488  WRONG RESULT TEST 1 </div><div class="">    argon06.inp                                               -85.18989253445228  WRONG RESULT TEST 1 </div><div class="">    argon07.inp                                               -85.05087192159809         RUNTIME FAIL </div><div class="">    argon08.inp                                               -85.05201740647929         RUNTIME FAIL </div><div class="">    argon09.inp                                               -85.05086520280044         RUNTIME FAIL </div><div class="">    argon10.inp                                               -85.05070440200512         RUNTIME FAIL </div><div class="">    argon11.inp                                               -84.69892988333885         RUNTIME FAIL </div><div class="">    argon12.inp                                               -84.69900817368848         RUNTIME FAIL </div><div class="">    argon13.inp                                               -84.81306482759408  WRONG RESULT TEST 1 </div><div class="">    argon14.inp                                               -84.69889654472566  WRONG RESULT TEST 1 </div><div class="">    argon-beef.inp                                            -42.46311172518392  WRONG RESULT TEST 1 </div><div class="">    dftd3bj_t1.inp                                             -0.00355123783846     OK (   1.19 sec) </div><div class="">    dftd3bj_t2.inp                                             -0.05897356220363     OK (   2.20 sec) </div><div class="">    dftd3bj_t3.inp                                             -0.00112424003807     OK (   3.75 sec) </div><div class="">    dftd3bj_t4.inp                                                -84.2983390350     OK (   3.86 sec) </div><div class=""><<< /home/rcohen/CP2K/cp2k/TEST-Linux-x86-64-intel-psmp-2019-01-07_18-24-16/tests/QS/regtest-dft-vdw-corr-3 (1 of 3) done in 775.00 sec</div><div class="">Starting regression tests in /home/rcohen/CP2K/cp2k/TEST-Linux-x86-64-intel-psmp-2019-01-07_18-24-16/tests/QS/regtest-dft-vdw-corr-2 (2 of 3)</div><div class="">Starting regression tests in /home/rcohen/CP2K/cp2k/TEST-Linux-x86-64-intel-psmp-2019-01-07_18-24-16/tests/QS/regtest-dft-vdw-corr-2 (2 of 3)</div><div class=""><br class=""></div></div><div class=""><br class=""></div><div class="">Almost all of the non vdw routines pass.</div><div class=""><br class=""></div><div class="">Sincerely,</div><div class=""><br class=""></div><div class="">Ron</div><div class=""><br class=""></div><div class="">
<div style="color: rgb(0, 0, 0); font-family: Helvetica; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-position: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: 2; text-align: -webkit-auto; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">---<br class="">Ron Cohen<br class=""><a href="mailto:reco...@gmail.com" class="">reco...@gmail.com</a><br class="">skypename: ronaldcohen</div><div style="color: rgb(0, 0, 0); font-family: Helvetica; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-variant-east-asian: normal; font-variant-position: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: 2; text-align: -webkit-auto; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">twitter: @recohen3<br class=""><div class=""><br class=""></div></div><br class="Apple-interchange-newline"><br class="Apple-interchange-newline">
</div>
<div><br class=""><blockquote type="cite" class=""><div class="">On Jan 7, 2019, at 6:12 PM, Robert Schade <<a href="mailto:robert...@uni-paderborn.de" class="">robert...@uni-paderborn.de</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><div class=""><div class="protected-part"><div class="protected-title">Signed PGP part</div><div class="protected-content">Could you try setting KMP_STACKSIZE to something large in the terminal<br class="">session with "export KMP_STACKSIZE=512m" before you rerun the regtests<br class="">with your intel-psmp-binary that failed before?<br class="">Please also make sure that the general stack size is not the problem<br class="">by running "ulimt -s unlimited" in the same terminal where you want to<br class="">execute the regtests.<br class="">Best Wishes<br class="">Robert<br class=""><br class="">On 07.01.19 18:00, Ronald Cohen wrote:<br class="">> BTW, in case it was not clear. My Intel builds of POPT and PSMP<br class="">> versions were error free. The problems were all run time.<br class="">><br class="">> Ron<br class="">><br class="">> --- Ron Cohen <a href="mailto:reco...@gmail.com" class="">reco...@gmail.com</a> <<a href="mailto:reco...@gmail.com" class="">mailto:reco...@gmail.com</a>><br class="">> skypename: ronaldcohen twitter: @recohen3<br class="">><br class="">><br class="">><br class="">><br class="">>> On Jan 7, 2019, at 5:39 PM, Robert Schade<br class="">>> <<a href="mailto:robert...@uni-paderborn.de" class="">robert...@uni-paderborn.de</a><br class="">>> <<a href="mailto:robert...@uni-paderborn.de" class="">mailto:robert...@uni-paderborn.de</a>>> wrote:<br class="">>><br class="">>> Signed PGP part r is automatically private because it is the<br class="">>> first iteration variable. Every drho(s, i) is only read and<br class="">>> written in exactly one loop iteration. The statement<br class="">>> "COLLAPSE(3)" collapses the three perfectly nested loops into one<br class="">>> loop. So, IMHO, this code looks ok. Best Wishes Robert<br class="">>><br class="">>><br class="">>> On 07.01.19 14:52, Ronald Cohen wrote:<br class="">>>> Yes, I agree. I have tried the 2018.05 and the 2019.1 intel<br class="">>>> compilers. The POPT version runs fine, but the PSMP version<br class="">>>> fails in the vDW routines. I find things like: in<br class="">>>> qs_dispersion_nonloc.F<br class="">>>><br class="">>>> !$OMP PARALLEL DO DEFAULT(NONE)<span class="Apple-converted-space">                     </span> & !$OMP<br class="">>>> SHARED(ispin,i,n,lo,drho,drho_r)<span class="Apple-converted-space">  </span> & !$OMP<br class="">>>> PRIVATE(s) & !$OMP<span class="Apple-converted-space">            </span> COLLAPSE(3) DO r = 0, n(3)-1 DO<br class="">>>> q = 0, n(2)-1 DO p = 0, n(1)-1 s = r*n(2)*n(1)+q*n(1)+p+1<br class="">>>> drho(s, i) = drho(s, i)+drho_r(i, ispin)%pw%cr3d(p+lo(1), q<br class="">>>> +lo(2), r+lo(3)) END DO END DO END DO !$OMP END PARALLEL DO END<br class="">>>> DO END DO<br class="">>>><br class="">>>> Doesn’t this have to be marked as a reduction? And shouldn’t r,<br class="">>>> q, p be labeled private? Perhaps this is automatic, but I do<br class="">>>> not see that said anywhere. Does gnu treat such differently<br class="">>>> than intel? Just ideas.<br class="">>>><br class="">>>> I am currently trying the toolchain, but it is building<br class="">>>> everything from scratch, including blas, lapack, scalapack etc<br class="">>>> etc, so will take days.<br class="">>>><br class="">>>> Thank you for your help,<br class="">>>><br class="">>>> Sincerely,<br class="">>>><br class="">>>> Ron<br class="">>>><br class="">>>> --- Ron Cohen <a href="mailto:reco...@gmail.com" class="">reco...@gmail.com</a> <<a href="mailto:reco...@gmail.com" class="">mailto:reco...@gmail.com</a>><br class="">>> <<a href="mailto:reco...@gmail.com" class="">mailto:reco...@gmail.com</a>><br class="">>>> skypename: ronaldcohen twitter: @recohen3<br class="">>>><br class="">>>><br class="">>>><br class="">>>><br class="">>>>> On Jan 7, 2019, at 2:16 PM, Robert Schade<br class="">>>>> <<a href="mailto:robert...@uni-paderborn.de" class="">robert...@uni-paderborn.de</a><br class="">>>>> <<a href="mailto:robert...@uni-paderborn.de" class="">mailto:robert...@uni-paderborn.de</a>><br class="">>>>> <<a href="mailto:robert...@uni-paderborn.de" class="">mailto:robert...@uni-paderborn.de</a>>> wrote:<br class="">>>>><br class="">>>>> Building cp2k on Intel Xeon Phi Knights Landing (KNL, not to<br class="">>>>> be confused with KNC!) is not different from building it on<br class="">>>>> any other Intel CPU. Hence, I think that the failing regtests<br class="">>>>> point to an underlying issue. Which exact version of the<br class="">>>>> Intel Compiler and MKL have you tried? Best Wishes Robert<br class="">>>>><br class="">>>>> On 06.01.19 01:59, Ronald Cohen wrote:<br class="">>>>>> OK—sorry for all the noise. I am trying:<br class="">>>>>> ./install_cp2k_toolchain.sh --with-elpa=install<br class="">>>>>> --with-libint=install --with-gcc=install I hate not being<br class="">>>>>> able to use my intel tools which work for me for everything<br class="">>>>>> else just fine.<br class="">>>>>><br class="">>>>>> Ron<br class="">>>>>><br class="">>>>><br class="">>>>> -- Robert Schade Paderborn Center for Parallel Computing<br class="">>>>> (PC2) University of Paderborn Warburger Str. 100 D-33098<br class="">>>>> Paderborn Germany <a href="mailto:robert...@uni-paderborn.de" class="">robert...@uni-paderborn.de</a><br class="">>> <<a href="mailto:robert...@uni-paderborn.de" class="">mailto:robert...@uni-paderborn.de</a>><br class="">>>>> <<a href="mailto:robert...@uni-paderborn.de" class="">mailto:robert...@uni-paderborn.de</a>> +49/(0)5251/60-5393<br class="">>>>><br class="">>>>> -- You received this message because you are subscribed to a<br class="">>>>> topic in the Google Groups "cp2k" group. To unsubscribe from<br class="">>>>> this topic, visit<br class="">>>>> <a href="https://groups.google.com/d/topic/cp2k/gzmRqKNt62U/unsubscribe" class="">https://groups.google.com/d/topic/cp2k/gzmRqKNt62U/unsubscribe</a>.<br class="">>><br class="">>>>><br class="">>> To unsubscribe from this group and all its topics, send an email<br class="">>>>> to <a href="mailto:cp2k+uns...@googlegroups.com" class="">cp2k+uns...@googlegroups.com</a><br class="">>> <<a href="mailto:cp2k+uns...@googlegroups.com" class="">mailto:cp2k+uns...@googlegroups.com</a>>. To post to this<br class="">>> group, send<br class="">>>>> email to <a href="mailto:cp...@googlegroups.com" class="">cp...@googlegroups.com</a><br class="">>>>> <<a href="mailto:cp...@googlegroups.com" class="">mailto:cp...@googlegroups.com</a>>.<br class="">>> Visit this group at<br class="">>>>> <a href="https://groups.google.com/group/cp2k" class="">https://groups.google.com/group/cp2k</a>. For more options,<br class="">>>>> visit <a href="https://groups.google.com/d/optout" class="">https://groups.google.com/d/optout</a>.<br class="">>>><br class="">>>> -- You received this message because you are subscribed to the<br class="">>>> Google Groups "cp2k" group. To unsubscribe from this group and<br class="">>>> stop receiving emails from it, send an email to<br class="">>>> <a href="mailto:cp2k+uns...@googlegroups.com" class="">cp2k+uns...@googlegroups.com</a><br class="">>> <<a href="mailto:cp2k+uns...@googlegroups.com" class="">mailto:cp2k+uns...@googlegroups.com</a>><br class="">>>> <<a href="mailto:cp2k+uns...@googlegroups.com" class="">mailto:cp2k+uns...@googlegroups.com</a>>. To post to this<br class="">>>> group, send email to <a href="mailto:cp...@googlegroups.com" class="">cp...@googlegroups.com</a><br class="">>>> <<a href="mailto:cp...@googlegroups.com" class="">mailto:cp...@googlegroups.com</a>> <<a href="mailto:cp...@googlegroups.com" class="">mailto:cp...@googlegroups.com</a>>.<br class="">>>> Visit this group at <a href="https://groups.google.com/group/cp2k" class="">https://groups.google.com/group/cp2k</a>. For<br class="">>>> more options, visit <a href="https://groups.google.com/d/optout" class="">https://groups.google.com/d/optout</a>.<br class="">>><br class="">>> -- Robert Schade Paderborn Center for Parallel Computing (PC2)<br class="">>> University of Paderborn Warburger Str. 100 D-33098 Paderborn<br class="">>> Germany <a href="mailto:robert...@uni-paderborn.de" class="">robert...@uni-paderborn.de</a><br class="">>> <<a href="mailto:robert...@uni-paderborn.de" class="">mailto:robert...@uni-paderborn.de</a>> +49/(0)5251/60-5393<br class="">>><br class="">><br class="">> -- You received this message because you are subscribed to the<br class="">> Google Groups "cp2k" group. To unsubscribe from this group and stop<br class="">> receiving emails from it, send an email to<br class="">> <a href="mailto:cp2k+uns...@googlegroups.com" class="">cp2k+uns...@googlegroups.com</a><br class="">> <<a href="mailto:cp2k+uns...@googlegroups.com" class="">mailto:cp2k+uns...@googlegroups.com</a>>. To post to this group,<br class="">> send email to <a href="mailto:cp...@googlegroups.com" class="">cp...@googlegroups.com</a><br class="">> <<a href="mailto:cp...@googlegroups.com" class="">mailto:cp...@googlegroups.com</a>>. Visit this group at<br class="">> <a href="https://groups.google.com/group/cp2k" class="">https://groups.google.com/group/cp2k</a>. For more options, visit<br class="">> <a href="https://groups.google.com/d/optout" class="">https://groups.google.com/d/optout</a>.<br class=""><br class="">--<br class="">Robert Schade<br class="">Paderborn Center for Parallel Computing (PC2)<br class="">University of Paderborn<br class="">Warburger Str. 100<br class="">D-33098 Paderborn<br class="">Germany<br class=""><a href="mailto:robert...@uni-paderborn.de" class="">robert...@uni-paderborn.de</a><br class="">+49/(0)5251/60-5393</div></div><br class=""><iframe class="untrusted-content-test" scrolling="auto" width="200" height="20" style="border:none;display:block;overflow:auto;" data-src="data:text/html;charset=UTF-8;base64,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" sandbox="allow-scripts"></iframe></div></div></blockquote></div><br class=""></div></body></html>