<?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 Profiling MPI applications using Vtune, collecting data only on select nodes in Intel® Moderncode for Parallel Architectures</title>
    <link>https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Profiling-MPI-applications-using-Vtune-collecting-data-only-on/m-p/1120992#M7560</link>
    <description>&lt;P&gt;Hello all.&amp;nbsp;&lt;/P&gt;

&lt;P&gt;I am trying to profile an MPI application. 6 nodes, 120 total MPI ranks.&amp;nbsp;&lt;/P&gt;

&lt;P&gt;I am interested in either a) Metrics on a single node or b) average metrics across all nodes.&amp;nbsp;&lt;/P&gt;

&lt;P&gt;Is there an easy way to do an option b) run?&lt;/P&gt;

&lt;P&gt;Can anyone think of any ways that my data can be skewed if I'm only collecting metrics on a single rank?&lt;/P&gt;

&lt;P&gt;Thank you.&amp;nbsp;&lt;/P&gt;</description>
    <pubDate>Thu, 22 Dec 2016 05:48:09 GMT</pubDate>
    <dc:creator>Paulius_V_</dc:creator>
    <dc:date>2016-12-22T05:48:09Z</dc:date>
    <item>
      <title>Profiling MPI applications using Vtune, collecting data only on select nodes</title>
      <link>https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Profiling-MPI-applications-using-Vtune-collecting-data-only-on/m-p/1120992#M7560</link>
      <description>&lt;P&gt;Hello all.&amp;nbsp;&lt;/P&gt;

&lt;P&gt;I am trying to profile an MPI application. 6 nodes, 120 total MPI ranks.&amp;nbsp;&lt;/P&gt;

&lt;P&gt;I am interested in either a) Metrics on a single node or b) average metrics across all nodes.&amp;nbsp;&lt;/P&gt;

&lt;P&gt;Is there an easy way to do an option b) run?&lt;/P&gt;

&lt;P&gt;Can anyone think of any ways that my data can be skewed if I'm only collecting metrics on a single rank?&lt;/P&gt;

&lt;P&gt;Thank you.&amp;nbsp;&lt;/P&gt;</description>
      <pubDate>Thu, 22 Dec 2016 05:48:09 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Profiling-MPI-applications-using-Vtune-collecting-data-only-on/m-p/1120992#M7560</guid>
      <dc:creator>Paulius_V_</dc:creator>
      <dc:date>2016-12-22T05:48:09Z</dc:date>
    </item>
    <item>
      <title>This subject seems more</title>
      <link>https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Profiling-MPI-applications-using-Vtune-collecting-data-only-on/m-p/1120993#M7561</link>
      <description>&lt;P&gt;This subject seems more likely to get the attention of experts if posted on the HPC clusters companion forum.&lt;/P&gt;

&lt;P&gt;I don't consider it easy, but you may need to run VTune command line, collecting a separate .tb5 for each rank.&amp;nbsp; This would enable you to analyze performance issues which aren't replicated in all ranks.&amp;nbsp; Presumably easier would be to use Intel Trace Analyzer to determine whether there are such issues.&lt;/P&gt;

&lt;P&gt;Useful results usually may be obtained by running a single copy of VTune when running a job on a single node.&lt;/P&gt;

&lt;P&gt;Trace Analyzer is a descendant of the original vampir which is available as open source; another open source alternative is jumpshot.&lt;/P&gt;</description>
      <pubDate>Thu, 22 Dec 2016 12:24:00 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Profiling-MPI-applications-using-Vtune-collecting-data-only-on/m-p/1120993#M7561</guid>
      <dc:creator>TimP</dc:creator>
      <dc:date>2016-12-22T12:24:00Z</dc:date>
    </item>
  </channel>
</rss>

