<?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 Can you try using these in Intel® MPI Library</title>
    <link>https://community.intel.com/t5/Intel-MPI-Library/MPI-Rank-Binding/m-p/1026599#M4113</link>
    <description>&lt;P&gt;Can you try using these options:&amp;nbsp;&lt;/P&gt;

&lt;P&gt;(a) &amp;nbsp;&lt;/P&gt;

&lt;PRE class="brush:bash;"&gt;"-genv I_MPI_PIN_DOMAIN=core -genv I_MPI_PIN_ORDER=compact"&lt;/PRE&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;

&lt;P&gt;(b)&lt;/P&gt;

&lt;PRE class="brush:bash;"&gt;"-genv I_MPI_PIN_DOMAIN=core -genv I_MPI_PIN_ORDER=scatter"&lt;/PRE&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;</description>
    <pubDate>Tue, 19 Aug 2014 22:38:02 GMT</pubDate>
    <dc:creator>drMikeT</dc:creator>
    <dc:date>2014-08-19T22:38:02Z</dc:date>
    <item>
      <title>MPI Rank Binding</title>
      <link>https://community.intel.com/t5/Intel-MPI-Library/MPI-Rank-Binding/m-p/1026598#M4112</link>
      <description>&lt;P&gt;Hello all,&lt;/P&gt;

&lt;P&gt;Intel MPI 4.1.3 on RHEL6.4: trying to bind ranks in two simple fashions:(a) 2 ranks to the same processor socket and (b) 2 ranks to different processor sockets.&lt;/P&gt;

&lt;P&gt;&lt;SPAN style="font-size: 1em; line-height: 1.5;"&gt;Looking at the Intel MPI Reference Manual (&lt;/SPAN&gt;3.2. Process Pinning&amp;nbsp;pp.98+)&lt;SPAN style="font-size: 1em; line-height: 1.5;"&gt;, we should be able to use options in mpiexec.hydra when the hostfile points to the same host&lt;/SPAN&gt;&lt;/P&gt;

&lt;PRE class="brush:bash;"&gt;-genv I_MPI_PIN 1  -genv I_MPI_PIN_PROCESSOR_LIST all:bunch&lt;/PRE&gt;

&lt;PRE class="brush:bash;"&gt;-genv I_MPI_PIN 1  -genv I_MPI_PIN_PROCESSOR_LIST all:scatter&lt;/PRE&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;

&lt;P&gt;Unfortunately, the "scatter" option STILL binds the MPI ranks to the same socket.&amp;nbsp;&lt;/P&gt;

&lt;P&gt;Should I be using the&amp;nbsp;&amp;nbsp;"I_MPI_PIN_DOMAIN" instead?&lt;/P&gt;

&lt;P&gt;Any sugestions?&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;

&lt;P&gt;Thanks, Michael&lt;/P&gt;</description>
      <pubDate>Tue, 19 Aug 2014 21:43:55 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-MPI-Library/MPI-Rank-Binding/m-p/1026598#M4112</guid>
      <dc:creator>HPCS</dc:creator>
      <dc:date>2014-08-19T21:43:55Z</dc:date>
    </item>
    <item>
      <title>Can you try using these</title>
      <link>https://community.intel.com/t5/Intel-MPI-Library/MPI-Rank-Binding/m-p/1026599#M4113</link>
      <description>&lt;P&gt;Can you try using these options:&amp;nbsp;&lt;/P&gt;

&lt;P&gt;(a) &amp;nbsp;&lt;/P&gt;

&lt;PRE class="brush:bash;"&gt;"-genv I_MPI_PIN_DOMAIN=core -genv I_MPI_PIN_ORDER=compact"&lt;/PRE&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;

&lt;P&gt;(b)&lt;/P&gt;

&lt;PRE class="brush:bash;"&gt;"-genv I_MPI_PIN_DOMAIN=core -genv I_MPI_PIN_ORDER=scatter"&lt;/PRE&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;</description>
      <pubDate>Tue, 19 Aug 2014 22:38:02 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-MPI-Library/MPI-Rank-Binding/m-p/1026599#M4113</guid>
      <dc:creator>drMikeT</dc:creator>
      <dc:date>2014-08-19T22:38:02Z</dc:date>
    </item>
  </channel>
</rss>

