<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>topic Having Problems using  PARDISO  OOC in Intel® oneAPI Math Kernel Library</title>
    <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Having-Problems-using-PARDISO-OOC/m-p/894017#M10681</link>
    <description>&lt;BR /&gt;Hi
&lt;BR /&gt;
&lt;BR /&gt; I am having problems in PARDISO  out of core solver. In my code, I set these variables (using C convention):
&lt;BR /&gt;iparam[2] = 1;
&lt;BR /&gt;iparam[59] = 2;
&lt;BR /&gt;
&lt;BR /&gt;I use VS2005 (Windows XP, 32-bit, P4, 2GB)  and I call PARDISO from inside a COM DLL. The size of the matrix I try to solve = ~340,000 with 7.5 million nonzeros. 
&lt;BR /&gt;
&lt;BR /&gt;First I  define "MKL_PARDISO_OOC_MAX_CORE_SIZE  = 2000"    in "pardiso_ooc.cfg" and then, I am getting  "-2" error code (insufficient memory). And If it is "MKL_PARDISO_OOC_MAX_CORE_SIZE  = 250", I am getting -9 (insufficient memory for OOC). I tried a few other numbers, but i still get either -2 or -9.
&lt;BR /&gt;
&lt;BR /&gt;Interestingly, if I create a simple Windows-console application and try to solve my matrix,  It works successfully with both in-core and OOC PARDISO.
&lt;BR /&gt;
&lt;BR /&gt;Any help greatly appreciated
&lt;BR /&gt;regards
&lt;BR /&gt;bulent
&lt;BR /&gt;
&lt;BR /&gt;
&lt;BR /&gt;
&lt;BR /&gt;</description>
    <pubDate>Mon, 01 Sep 2008 16:31:13 GMT</pubDate>
    <dc:creator>Alemdar__Bulent</dc:creator>
    <dc:date>2008-09-01T16:31:13Z</dc:date>
    <item>
      <title>Having Problems using  PARDISO  OOC</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Having-Problems-using-PARDISO-OOC/m-p/894017#M10681</link>
      <description>&lt;BR /&gt;Hi
&lt;BR /&gt;
&lt;BR /&gt; I am having problems in PARDISO  out of core solver. In my code, I set these variables (using C convention):
&lt;BR /&gt;iparam[2] = 1;
&lt;BR /&gt;iparam[59] = 2;
&lt;BR /&gt;
&lt;BR /&gt;I use VS2005 (Windows XP, 32-bit, P4, 2GB)  and I call PARDISO from inside a COM DLL. The size of the matrix I try to solve = ~340,000 with 7.5 million nonzeros. 
&lt;BR /&gt;
&lt;BR /&gt;First I  define "MKL_PARDISO_OOC_MAX_CORE_SIZE  = 2000"    in "pardiso_ooc.cfg" and then, I am getting  "-2" error code (insufficient memory). And If it is "MKL_PARDISO_OOC_MAX_CORE_SIZE  = 250", I am getting -9 (insufficient memory for OOC). I tried a few other numbers, but i still get either -2 or -9.
&lt;BR /&gt;
&lt;BR /&gt;Interestingly, if I create a simple Windows-console application and try to solve my matrix,  It works successfully with both in-core and OOC PARDISO.
&lt;BR /&gt;
&lt;BR /&gt;Any help greatly appreciated
&lt;BR /&gt;regards
&lt;BR /&gt;bulent
&lt;BR /&gt;
&lt;BR /&gt;
&lt;BR /&gt;
&lt;BR /&gt;</description>
      <pubDate>Mon, 01 Sep 2008 16:31:13 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Having-Problems-using-PARDISO-OOC/m-p/894017#M10681</guid>
      <dc:creator>Alemdar__Bulent</dc:creator>
      <dc:date>2008-09-01T16:31:13Z</dc:date>
    </item>
    <item>
      <title>Re: Having Problems using  PARDISO  OOC</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Having-Problems-using-PARDISO-OOC/m-p/894018#M10682</link>
      <description>Make sure you are using the latest Update 4 for MKL 10 . I found a  problem with OOC and PARDISO in Update 3 - it was targeted to be fixed in Update 4.</description>
      <pubDate>Tue, 02 Sep 2008 17:02:06 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Having-Problems-using-PARDISO-OOC/m-p/894018#M10682</guid>
      <dc:creator>AndrewC</dc:creator>
      <dc:date>2008-09-02T17:02:06Z</dc:date>
    </item>
    <item>
      <title>Re: Having Problems using  PARDISO  OOC</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Having-Problems-using-PARDISO-OOC/m-p/894019#M10683</link>
      <description>&lt;P&gt;I just checked my version: 10.0.4.023&lt;/P&gt;
&lt;P&gt;b.&lt;/P&gt;</description>
      <pubDate>Tue, 02 Sep 2008 17:44:47 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Having-Problems-using-PARDISO-OOC/m-p/894019#M10683</guid>
      <dc:creator>Alemdar__Bulent</dc:creator>
      <dc:date>2008-09-02T17:44:47Z</dc:date>
    </item>
    <item>
      <title>Re: Having Problems using  PARDISO  OOC</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Having-Problems-using-PARDISO-OOC/m-p/894020#M10684</link>
      <description>&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;Hi Bulent,&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;I think that the cause of the problem you met is limitation for allocation memory from DLL ( I guess that this limit something around 2 GB ). Some people compains about 1 Gb only.&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;/SPAN&gt;&lt;SPAN&gt;&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;Look at your case: &lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;The Input data: Size of the matrix ~ 340.000, nonzero(A) ~ 7.500.000 &lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;Therefore after reordering we can expect the number of nonzeros LU factors may be for example &amp;gt; 500.000.000. So, for allocation this LU factors requires more that 2 Gb.&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;++++++++++++++++++++++++++++++++++++++++&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;As an example: I have tried to reproduce the task&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;Input matrix size = 350 000&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;Mtype == 2 (real and symmetric positive definite matrix)&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;and please look at the snip of statistics report:&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&amp;lt; Linear system Ax = b&amp;gt; &lt;TRANSPOSE&gt;&lt;P&gt;&lt;/P&gt;&lt;/TRANSPOSE&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;SPAN&gt; &lt;/SPAN&gt;#equations:&lt;SPAN&gt; &lt;/SPAN&gt;350000&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;SPAN&gt; &lt;/SPAN&gt;&lt;/SPAN&gt;&lt;SPAN&gt;#non-zeros in A:&lt;SPAN&gt; &lt;/SPAN&gt;1384500&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;SPAN&gt; &lt;/SPAN&gt;non-zeros in A ():&lt;SPAN&gt; &lt;/SPAN&gt;0.001130&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;SPAN&gt; &lt;/SPAN&gt;&lt;/SPAN&gt;&lt;SPAN&gt;#right-hand sides:&lt;SPAN&gt; &lt;/SPAN&gt;1&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&amp;lt; Factors L and U &amp;gt;&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;SPAN&gt; &lt;/SPAN&gt;#columns for each panel:&lt;SPAN&gt; &lt;/SPAN&gt;80&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;SPAN&gt;&amp;amp;
nbsp; &lt;/SPAN&gt;#independent subgraphs:&lt;SPAN&gt; &lt;/SPAN&gt;0&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&amp;lt; Preprocessing with state of the art partitioning metis&amp;gt;&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;SPAN&gt; &lt;/SPAN&gt;#supernodes:&lt;SPAN&gt; &lt;/SPAN&gt;230149&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;SPAN&gt; &lt;/SPAN&gt;size of largest supernode:&lt;SPAN&gt; &lt;/SPAN&gt;5062&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;SPAN&gt; &lt;/SPAN&gt;number of nonzeros in L&lt;SPAN&gt; &lt;/SPAN&gt;152892710&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;SPAN&gt; &lt;/SPAN&gt;number of nonzeros in U&lt;SPAN&gt; &lt;/SPAN&gt;&lt;SPAN&gt;&lt;/SPAN&gt;1&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;SPAN&gt; &lt;/SPAN&gt;number of nonzeros in L+U&lt;SPAN&gt; &lt;/SPAN&gt;152892711&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;++++++++++++++++++++++++++++++++++++++++++&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;Please pay attention for this matrix (only 1.4 million non-zeroes) we have nonzeros for LU ~ 150 million, therefore for allocation this array require ~ 8*150 million &amp;gt; 1.2 Gb&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;This is just suggestion. You can easy to check this suggestion to reduce the task size. &lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;Please let me know if any further question.&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;Regards, &lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;
&lt;P class="MsoNormal"&gt;&lt;SPAN&gt;--Gennady&lt;P&gt;&lt;/P&gt;&lt;/SPAN&gt;&lt;/P&gt;</description>
      <pubDate>Wed, 03 Sep 2008 12:40:36 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Having-Problems-using-PARDISO-OOC/m-p/894020#M10684</guid>
      <dc:creator>Gennady_F_Intel</dc:creator>
      <dc:date>2008-09-03T12:40:36Z</dc:date>
    </item>
    <item>
      <title>Re: Having Problems using  PARDISO  OOC</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Having-Problems-using-PARDISO-OOC/m-p/894021#M10685</link>
      <description>&lt;P&gt;Gennady&lt;/P&gt;
&lt;P&gt;Thanks for your comments. I run in-core version of PARDISO in my little test program (calling PARDISO routines from a windows-console application). Here is the results I have:&lt;/P&gt;
&lt;P&gt;&amp;lt; Linear system Ax = b&amp;gt; &lt;TRANSPOSE&gt;&lt;BR /&gt; #equations: 324885&lt;BR /&gt; #non-zeros in A: 7440848&lt;BR /&gt; non-zeros in A (%): 0.007050&lt;BR /&gt; #right-hand sides: 1&lt;/TRANSPOSE&gt;&lt;/P&gt;
&lt;P&gt;&amp;lt; Factors L and U &amp;gt;&lt;BR /&gt; #columns for each panel: 80&lt;BR /&gt; #independent subgraphs: 0&lt;BR /&gt;&amp;lt; Preprocessing with state of the art partitioning metis&amp;gt;&lt;BR /&gt; #supernodes: 55249&lt;BR /&gt; size of largest supernode: 2479&lt;BR /&gt; number of nonzeros in L 63468206&lt;BR /&gt; number of nonzeros in U 1&lt;BR /&gt; number of nonzeros in L+U 63468207&lt;/P&gt;
&lt;P&gt;For factorization part, apparently, the programs needs 63468207 * 8 /1E6 = 507.75 MB memory. &lt;/P&gt;
&lt;P&gt;I see your comments regarding high memory requirements and I totally agree with you.. And, I understand that 
when I call PARDISO from a DLL, it may be the reason that either OS or DLL does not allow PARDISO use ~500 MB from the heap memory. &lt;/P&gt;
&lt;P&gt;On the other had, what is confusing me is that I run into the same limitation when I use out-of-core version of PARDISO called from my DLL. (for OOC version, I believe this 500MB is stored in files , and then the solver partially reads some of these files into memory during factorization).&lt;/P&gt;
&lt;P&gt;Regards&lt;/P&gt;
&lt;P&gt;Bulent&lt;/P&gt;</description>
      <pubDate>Thu, 04 Sep 2008 15:22:06 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Having-Problems-using-PARDISO-OOC/m-p/894021#M10685</guid>
      <dc:creator>Alemdar__Bulent</dc:creator>
      <dc:date>2008-09-04T15:22:06Z</dc:date>
    </item>
    <item>
      <title>Re: Having Problems using  PARDISO  OOC</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Having-Problems-using-PARDISO-OOC/m-p/894022#M10686</link>
      <description>&lt;DIV style="margin:0px;"&gt;
&lt;DIV id="quote_reply" style="margin-top: 5px; width: 100%;"&gt;
&lt;DIV style="margin-left:2px;margin-right:2px;"&gt;Quoting - &lt;A href="https://community.intel.com/en-us/profile/240567"&gt;bulent.alemdar@bentley.com&lt;/A&gt;&lt;/DIV&gt;
&lt;DIV style="background-color:#E5E5E5; padding:5px;border: 1px; border-style: inset;margin-left:2px;margin-right:2px;"&gt;&lt;EM&gt;&lt;BR /&gt;Hi &lt;BR /&gt;&lt;BR /&gt;I am having problems in PARDISO out of core solver. In my code, I set these variables (using C convention): &lt;BR /&gt;iparam[2] = 1; &lt;BR /&gt;iparam[59] = 2; &lt;BR /&gt;&lt;BR /&gt;I use VS2005 (Windows XP, 32-bit, P4, 2GB) and I call PARDISO from inside a COM DLL. The size of the matrix I try to solve = ~340,000 with 7.5 million nonzeros. &lt;BR /&gt;&lt;BR /&gt;First I define "MKL_PARDISO_OOC_MAX_CORE_SIZE = 2000" in "pardiso_ooc.cfg" and then, I am getting "-2" error code (insufficient memory). And If it is "MKL_PARDISO_OOC_MAX_CORE_SIZE = 250", I am getting -9 (insufficient memory for OOC). I tried a few other numbers, but i still get either -2 or -9. &lt;BR /&gt;&lt;BR /&gt;Interestingly, if I create a simple Windows-console application and try to solve my matrix, It works successfully with both in-core and OOC PARDISO. &lt;BR /&gt;&lt;BR /&gt;Any help greatly appreciated &lt;BR /&gt;regards &lt;BR /&gt;bulent &lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;/EM&gt;&lt;/DIV&gt;
&lt;/DIV&gt;
&lt;/DIV&gt;
&lt;P&gt;Hi, Bulent,&lt;/P&gt;
&lt;P&gt;You are not alone. I reported the exact same problem to Intel premier support 4 months ago, with a complete sample project, a dataset (&lt;A href="ftp://ftp.orca3d.com/pub/Intel/PardisoAlone.zip"&gt;ftp://ftp.orca3d.com/pub/Intel/PardisoAlone.zip&lt;/A&gt;), and the computer description,yet still had no answer up to today. Accidentally, we run the same executable in a single processor machine last week, and it runthrough without any problem. Please join me to ask MKL developteam to fix this dual-core related PARDISO-OOC bug ASAP.&lt;/P&gt;
&lt;P&gt;Regards&lt;/P&gt;
&lt;P&gt;Ming&lt;/P&gt;
&lt;P&gt;&lt;/P&gt;
&lt;P&gt;&lt;/P&gt;</description>
      <pubDate>Mon, 13 Oct 2008 18:04:06 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Having-Problems-using-PARDISO-OOC/m-p/894022#M10686</guid>
      <dc:creator>mma1</dc:creator>
      <dc:date>2008-10-13T18:04:06Z</dc:date>
    </item>
  </channel>
</rss>

