<?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 Windows MKL 10.2 vs 10.1 - defaulting to single core/single threaded in Intel® oneAPI Math Kernel Library</title>
    <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Windows-MKL-10-2-vs-10-1-defaulting-to-single-core-single/m-p/910993#M12156</link>
    <description>I am running various regression cases on MKL 10.1 vs 10.2 update 2 and 10.2 seems to resolutely insist on only using one core on my machine -( Intel Pentium D 3.Ghz running Windows XP64 ( 64 bit MKL). Of course, OMP_NUM_THREADS=2&lt;BR /&gt;I am linking against the OpenMP compatibility libraries&lt;BR /&gt;&lt;BR /&gt;To run 10.1 vs 10.2 I just drop the new DLLs into my app folder, and re-run the regression cases so repeating various tests s quite easy. It is very,very obvious that 10.2 only uses one core to detriment of solution times.&lt;BR /&gt;&lt;BR /&gt;Any ideas why this would happen?&lt;BR /&gt;&lt;BR /&gt;More...&lt;BR /&gt;If I reboot my machine, then run a test case then the first run will use both cores. Subsequent runs will only use one core - until I reboot , that is.....</description>
    <pubDate>Wed, 14 Oct 2009 19:01:46 GMT</pubDate>
    <dc:creator>AndrewC</dc:creator>
    <dc:date>2009-10-14T19:01:46Z</dc:date>
    <item>
      <title>Windows MKL 10.2 vs 10.1 - defaulting to single core/single threaded</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Windows-MKL-10-2-vs-10-1-defaulting-to-single-core-single/m-p/910993#M12156</link>
      <description>I am running various regression cases on MKL 10.1 vs 10.2 update 2 and 10.2 seems to resolutely insist on only using one core on my machine -( Intel Pentium D 3.Ghz running Windows XP64 ( 64 bit MKL). Of course, OMP_NUM_THREADS=2&lt;BR /&gt;I am linking against the OpenMP compatibility libraries&lt;BR /&gt;&lt;BR /&gt;To run 10.1 vs 10.2 I just drop the new DLLs into my app folder, and re-run the regression cases so repeating various tests s quite easy. It is very,very obvious that 10.2 only uses one core to detriment of solution times.&lt;BR /&gt;&lt;BR /&gt;Any ideas why this would happen?&lt;BR /&gt;&lt;BR /&gt;More...&lt;BR /&gt;If I reboot my machine, then run a test case then the first run will use both cores. Subsequent runs will only use one core - until I reboot , that is.....</description>
      <pubDate>Wed, 14 Oct 2009 19:01:46 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Windows-MKL-10-2-vs-10-1-defaulting-to-single-core-single/m-p/910993#M12156</guid>
      <dc:creator>AndrewC</dc:creator>
      <dc:date>2009-10-14T19:01:46Z</dc:date>
    </item>
    <item>
      <title>Re: Windows MKL 10.2 vs 10.1 - defaulting to single core/single</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Windows-MKL-10-2-vs-10-1-defaulting-to-single-core-single/m-p/910994#M12157</link>
      <description>&lt;DIV style="margin:0px;"&gt;
&lt;DIV id="quote_reply" style="width: 100%; margin-top: 5px;"&gt;
&lt;DIV style="margin-left:2px;margin-right:2px;"&gt;Quoting - &lt;A href="https://community.intel.com/en-us/profile/371334"&gt;vasci_intel&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;I am running various regression cases on MKL 10.1 vs 10.2 update 2 and 10.2 seems to resolutely insist on only using one core on my machine -( Intel Pentium D 3.Ghz running Windows XP64 ( 64 bit MKL). Of course, OMP_NUM_THREADS=2&lt;BR /&gt;I am linking against the OpenMP compatibility libraries&lt;BR /&gt;&lt;BR /&gt;To run 10.1 vs 10.2 I just drop the new DLLs into my app folder, and re-run the regression cases so repeating various tests s quite easy. It is very,very obvious that 10.2 only uses one core to detriment of solution times.&lt;BR /&gt;&lt;BR /&gt;Any ideas why this would happen?&lt;BR /&gt;&lt;BR /&gt;More...&lt;BR /&gt;If I reboot my machine, then run a test case then the first run will use both cores. Subsequent runs will only use one core - until I reboot , that is.....&lt;/EM&gt;&lt;/DIV&gt;
&lt;/DIV&gt;
&lt;/DIV&gt;
&lt;BR /&gt;I have answered my own question. I was not setting MKL_DYNAMIC so MKL, in its infinite wisdom, was setting the number of threads to 1. Setting MKL_DYNAMIC=FALSE forces MKL to use the number of threads set by , say, OMP_NUM_THREADS&lt;BR /&gt;</description>
      <pubDate>Wed, 14 Oct 2009 21:27:46 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Windows-MKL-10-2-vs-10-1-defaulting-to-single-core-single/m-p/910994#M12157</guid>
      <dc:creator>AndrewC</dc:creator>
      <dc:date>2009-10-14T21:27:46Z</dc:date>
    </item>
  </channel>
</rss>

