<html><head><meta http-equiv="Content-Type" content="text/html charset=iso-8859-1"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><div apple-content-edited="true"><div style="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: -webkit-auto; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-text-size-adjust: auto; -webkit-text-stroke-width: 0px; word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; ">that is quite interesting. In fact, in the very beginning I used the serial SuperLU and</div><div style="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: -webkit-auto; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-text-size-adjust: auto; -webkit-text-stroke-width: 0px; word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; ">then transitioned to SuperLU Dist. There, I realised that even for the serial operation</div><div style="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: -webkit-auto; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-text-size-adjust: auto; -webkit-text-stroke-width: 0px; word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; ">of SuperLU Dist it was considerably better and more efficient than SuperLU,</div><div style="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: -webkit-auto; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-text-size-adjust: auto; -webkit-text-stroke-width: 0px; word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><br></div><div style="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: -webkit-auto; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-text-size-adjust: auto; -webkit-text-stroke-width: 0px; word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; ">Thus, I think even in the serial case it pays off to use SuperLU Dist.</div><div style="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: -webkit-auto; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-text-size-adjust: auto; -webkit-text-stroke-width: 0px; word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><br></div><div style="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: -webkit-auto; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-text-size-adjust: auto; -webkit-text-stroke-width: 0px; word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; ">Greetings, Benedikt</div><div style="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: -webkit-auto; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-text-size-adjust: auto; -webkit-text-stroke-width: 0px; word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><br></div><div style="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: -webkit-auto; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-text-size-adjust: auto; -webkit-text-stroke-width: 0px; word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><br></div>
</div>
<br><div><div>On Oct 31, 2013, at 2:53 PM, Dominic Kempf <<a href="mailto:dominic.r.kempf@gmail.com">dominic.r.kempf@gmail.com</a>> wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite">Hey Benedikt,<br>I was referring to the serial version of SuperLU, as there is no<br>solver in ISTL using SuperLUDist.<br>Best,<br>Dominic<br><br>On Thu, Oct 31, 2013 at 2:42 PM, Benedikt Oswald<br><<a href="mailto:benedikt.oswald@lspr.ch">benedikt.oswald@lspr.ch</a>> wrote:<br><blockquote type="cite">Hello Dominic, thanks for these numbers!<br>In fact, I am using SuperLU Dist (MPI, parallel, complex distributed<br>matrix).<br><br>When you say SuperLU, do you mean SuperLU Dist or the serial version ?<br><br>Greetings, Benedikt<br><br><br><br>--------------------------------------------------------------------------------------------------------------------------------------------------------------<br>Dr. sc. techn. Benedikt Oswald - first engineer - LSPR AG - phone - +41 43<br>366 90 74<br>Technoparkstrasse 1, CH-8005 Zürich, <a href="mailto:benedikt.oswald@lspr.ch">benedikt.oswald@lspr.ch</a><br>"Passion is required for any great work, and for the Revolution passion and<br>audacity are required in big doses."<br>Ernesto 'Che' Guevara, Letter to his parents.<br>--------------------------------------------------------------------------------------------------------------------------------------------------------------<br><br>On Oct 31, 2013, at 1:45 PM, Dominic Kempf <<a href="mailto:dominic.r.kempf@gmail.com">dominic.r.kempf@gmail.com</a>><br>wrote:<br><br>Hey Christoph,<br>Hey devel-list,<br><br>while doing a writeup of my work for Markus, I reran some of the<br>performance tests and realized that infact SuperLU and UMFPack<br>performance results are NOT within the same order of magnitude.<br>UMFPack shows significantly better performance, see this table (I hope<br>its somehwat readable)<br><br>dofs 1e4 4e4 1.6e5 6.4e5<br>T_{superlu} 0.076 0.395 2.707 19.839<br>T_{umfpack} 0.064 0.245 1.367 9.176<br>time savings 15.79% 37.98% 49.51% 53.75%<br>$M_{superlu} 30.78 MB 123.23 MB 493.2 MB<br>$M_{umfpack} 25.72 MB 88.66 MB 418.9 MB<br>mem savings 16.44% 28.06% 15.07%<br>$P_{superlu} 13.4 MB 55.64 MB 253.18 MB 1.16 GB<br>$P_{umfpack} 11.8 MB 41.33 MB 177.14 MB 771.16 MB<br>peakmem sav. 11.95% 25.72% 30.04% 33.53%<br><br>T is the wallclock time here, M the total allocation size as given by<br>valgrind and P the peak memory usage as given by /usr/bin/time -v.<br>As you can see, there is something to be gained here!<br><br>Best,<br>Dominic<br><br>On Thu, Oct 17, 2013 at 10:33 AM, Dominic Kempf<br><<a href="mailto:dominic.r.kempf@gmail.com">dominic.r.kempf@gmail.com</a>> wrote:<br><br>Hey Christoph,<br>thanks for testing the branch. My system gave me memory with zeroes,<br>yours didnt... thats exactly where I need you.<br><br>I did a bit of measurements to compare superlu and umfpack, though not<br>on hard problems. In those tests umfpack was a bit faster and needed<br>less memory, within the same order of magnitude though. The original<br>reasons to implement umfpack support were conceptional difficulties<br>with superlu in multithreaded environment. Markus can elaborate more<br>on that (and whether umfpack actually solved the issue).<br><br><br>Best,<br>Dominic<br><br>2013/10/17 Christoph Grüninger <<a href="mailto:christoph.grueninger@iws.uni-stuttgart.de">christoph.grueninger@iws.uni-stuttgart.de</a>>:<br><br>Hi Dominic,<br>thanks for the quick patch. Using commit ...da3510863 it does no longer<br>segfault.<br><br>Bye<br>Christoph<br><br>--<br>Une science n'était vraiment développée que quand elle<br>pouvait utiliser les mathématiques. (Paul Lafargue)<br>*********************************************<br>CMWR 2014: 10th - 13th June 2014 in Stuttgart<br> Please visit <a href="http://www.cmwr14.de">www.cmwr14.de</a><br>*********************************************<br><br>_______________________________________________<br>Dune-devel mailing list<br><a href="mailto:Dune-devel@dune-project.org">Dune-devel@dune-project.org</a><br>http://lists.dune-project.org/mailman/listinfo/dune-devel<br><br><br>_______________________________________________<br>Dune-devel mailing list<br>Dune-devel@dune-project.org<br>http://lists.dune-project.org/mailman/listinfo/dune-devel<br><br><br><br>_______________________________________________<br>Dune-devel mailing list<br>Dune-devel@dune-project.org<br>http://lists.dune-project.org/mailman/listinfo/dune-devel<br><br></blockquote></blockquote></div><br></body></html>