<?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 Multi-threaded performance on MKL on highly loaded CPUs.... in Intel® oneAPI Math Kernel Library</title>
    <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Multi-threaded-performance-on-MKL-on-highly-loaded-CPUs/m-p/834936#M5987</link>
    <description>No, I don't expect MKL to perform any checks on which cores are busy. When you don't set KMP_AFFINITY, you leave scheduling to the OS. As you hinted, you can cut back to 1 thread by setting MKL_NUM_THREADS or OMP_NUM_THREADS.</description>
    <pubDate>Thu, 22 Apr 2010 18:13:48 GMT</pubDate>
    <dc:creator>TimP</dc:creator>
    <dc:date>2010-04-22T18:13:48Z</dc:date>
    <item>
      <title>Multi-threaded performance on MKL on highly loaded CPUs....</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Multi-threaded-performance-on-MKL-on-highly-loaded-CPUs/m-p/834935#M5986</link>
      <description>&lt;BR /&gt;Suppose I have dual core machine and I am using MKL in my application. I have not set MKL_NUM_THREADS, so my MKL routines will try to use both cores. Suppose though that another application is hammering both my cores; will MKL release that both cores are fully utilised and will back off and use only 1? &lt;BR /&gt;&lt;BR /&gt;This is not stated anywhere in the documentation (unless I missed it), so some clarification would be helpful please.&lt;BR /&gt;&lt;BR /&gt;thanks!&lt;BR /&gt;Tony</description>
      <pubDate>Thu, 22 Apr 2010 17:05:03 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Multi-threaded-performance-on-MKL-on-highly-loaded-CPUs/m-p/834935#M5986</guid>
      <dc:creator>Tony_Garratt</dc:creator>
      <dc:date>2010-04-22T17:05:03Z</dc:date>
    </item>
    <item>
      <title>Multi-threaded performance on MKL on highly loaded CPUs....</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Multi-threaded-performance-on-MKL-on-highly-loaded-CPUs/m-p/834936#M5987</link>
      <description>No, I don't expect MKL to perform any checks on which cores are busy. When you don't set KMP_AFFINITY, you leave scheduling to the OS. As you hinted, you can cut back to 1 thread by setting MKL_NUM_THREADS or OMP_NUM_THREADS.</description>
      <pubDate>Thu, 22 Apr 2010 18:13:48 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Multi-threaded-performance-on-MKL-on-highly-loaded-CPUs/m-p/834936#M5987</guid>
      <dc:creator>TimP</dc:creator>
      <dc:date>2010-04-22T18:13:48Z</dc:date>
    </item>
    <item>
      <title>Multi-threaded performance on MKL on highly loaded CPUs....</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Multi-threaded-performance-on-MKL-on-highly-loaded-CPUs/m-p/834937#M5988</link>
      <description>That's right. MKL doesn't check&lt;SPAN style="font-family: verdana, sans-serif;"&gt;whether the processors are busy or in any application or not.&lt;/SPAN&gt;&lt;DIV&gt;&lt;SPAN style="font-family: verdana, sans-serif;"&gt;&lt;/SPAN&gt;--Gennady&lt;BR /&gt;&lt;DIV&gt;&lt;SPAN style="font-family: verdana, sans-serif;"&gt;&lt;BR /&gt;&lt;/SPAN&gt;&lt;/DIV&gt;&lt;/DIV&gt;</description>
      <pubDate>Fri, 23 Apr 2010 04:30:02 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Multi-threaded-performance-on-MKL-on-highly-loaded-CPUs/m-p/834937#M5988</guid>
      <dc:creator>Gennady_F_Intel</dc:creator>
      <dc:date>2010-04-23T04:30:02Z</dc:date>
    </item>
  </channel>
</rss>

