<?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 ScaLAPACK for platform_mpi in Intel® oneAPI Math Kernel Library</title>
    <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/ScaLAPACK-for-platform-mpi/m-p/811906#M4017</link>
    <description>Sam,&lt;BR /&gt;I did not notice any significant overhead. In response to my question to HP-MPI developers in 2007 was:&lt;BR /&gt;&lt;PRE&gt;  As far as performance penalty, there IS a thin binary translation layer used, however it should not impact performance too much.&lt;BR /&gt;&lt;/PRE&gt;From simple ping-pong tests, I wasn't able to see any measurable effect of the wrappers either. &lt;BR /&gt;&lt;BR /&gt;Alexei&lt;BR /&gt;&lt;BR /&gt;P.S. For Windows, I did have to compile our own BLACS and ScaLAPACK. But then you can still link them with MKL.&lt;BR /&gt;&lt;BR /&gt;</description>
    <pubDate>Wed, 09 Feb 2011 14:18:34 GMT</pubDate>
    <dc:creator>Alexei_Yakovlev</dc:creator>
    <dc:date>2011-02-09T14:18:34Z</dc:date>
    <item>
      <title>ScaLAPACK for platform_mpi</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/ScaLAPACK-for-platform-mpi/m-p/811900#M4011</link>
      <description>I test a mpi code MUMPS which uses ScaLAPACK. MUMPS crashes as long as I use more than one processor. I suspect I didn't link the right version of ScaLAPACK and BLACS. I am using platform_mpi and couldn't find BLCAS lib for paltform_mpi. Instead i use intelmpi BLACS. I suspect intelmpi BLACS is not comptatible with platform_mpi. Here is my linking line for ScaLAPACK and BLACS:&lt;BR /&gt;-L/u/xeons03/people/xian/intel/composerxe-2011.1.107/mkl/lib/intel64/ -lmkl_scalapack_lp64 -lmkl_gf_lp64 -lmkl_blacs_intelmpi_lp64 -lmkl_intel_lp64 -lmkl_intel_thread -lmkl_core -L/u/xeons03/people/xian/intel/composerxe-2011.1.107/compiler/lib/intel64/ -liomp5 -lpthread -lm&lt;BR /&gt;&lt;BR /&gt;Any idea what's going wrong?&lt;BR /&gt;&lt;BR /&gt;Thanks</description>
      <pubDate>Mon, 31 Jan 2011 18:06:10 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/ScaLAPACK-for-platform-mpi/m-p/811900#M4011</guid>
      <dc:creator>xian-zhong_guous_cd-</dc:creator>
      <dc:date>2011-01-31T18:06:10Z</dc:date>
    </item>
    <item>
      <title>ScaLAPACK for platform_mpi</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/ScaLAPACK-for-platform-mpi/m-p/811901#M4012</link>
      <description>If you're using Intel Fortran, you can't use a gfortran library; in any case, you can't use the gfortran and Intel versions together. I don't expect compatibility between Platform MPI and any of the MPI versions supported by MKL. Of course, the MKL libraries which don't use MPI should be compatible.</description>
      <pubDate>Mon, 31 Jan 2011 20:23:34 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/ScaLAPACK-for-platform-mpi/m-p/811901#M4012</guid>
      <dc:creator>TimP</dc:creator>
      <dc:date>2011-01-31T20:23:34Z</dc:date>
    </item>
    <item>
      <title>ScaLAPACK for platform_mpi</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/ScaLAPACK-for-platform-mpi/m-p/811902#M4013</link>
      <description>Does this mean if I use platform_mpi, I have to recompile BLACS and ScaLAPACL using platform_mpi myself?</description>
      <pubDate>Tue, 01 Feb 2011 21:35:54 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/ScaLAPACK-for-platform-mpi/m-p/811902#M4013</guid>
      <dc:creator>xian-zhong_guous_cd-</dc:creator>
      <dc:date>2011-02-01T21:35:54Z</dc:date>
    </item>
    <item>
      <title>ScaLAPACK for platform_mpi</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/ScaLAPACK-for-platform-mpi/m-p/811903#M4014</link>
      <description>It seems you're right. The only way is to download ScaLAPACK &amp;amp; BLACS from NETLIB, build them with platform_mpi and change MKL scalapack and blacs libraries with the built ones, but leave other MKL libraries as is in order to use MKL BLAS.&lt;DIV&gt;&lt;/DIV&gt;&lt;DIV&gt;Regards,&lt;/DIV&gt;&lt;DIV&gt;Konstantin&lt;/DIV&gt;</description>
      <pubDate>Wed, 02 Feb 2011 03:10:56 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/ScaLAPACK-for-platform-mpi/m-p/811903#M4014</guid>
      <dc:creator>Konstantin_A_Intel</dc:creator>
      <dc:date>2011-02-02T03:10:56Z</dc:date>
    </item>
    <item>
      <title>ScaLAPACK for platform_mpi</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/ScaLAPACK-for-platform-mpi/m-p/811904#M4015</link>
      <description>Actually, you can use the MKL's BLACS/ScaLAPACK if you use the Platform's include files and libraries from the MPICH1.2 subdirectory. That's assuming you use Platform-MPI 7.1 a.k.a HP-MPI. We've been using MKL with HP-MPI this way for several years now and had no trouble with this combo.&lt;BR /&gt;&lt;BR /&gt;Alexei</description>
      <pubDate>Fri, 04 Feb 2011 17:26:16 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/ScaLAPACK-for-platform-mpi/m-p/811904#M4015</guid>
      <dc:creator>Alexei_Yakovlev</dc:creator>
      <dc:date>2011-02-04T17:26:16Z</dc:date>
    </item>
    <item>
      <title>ScaLAPACK for platform_mpi</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/ScaLAPACK-for-platform-mpi/m-p/811905#M4016</link>
      <description>Alexei,&lt;BR /&gt; I was able to combine MKL ScaLAPACK with platform_mpi using mpich2 wrapper. However, Platform suggests recompiling netlib BLACS/ScaLAPACK using platform_mpi due to overhead of mpich2 wrapper. Do you notice any ovehead?&lt;BR /&gt;&lt;BR /&gt;Sam</description>
      <pubDate>Tue, 08 Feb 2011 18:05:46 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/ScaLAPACK-for-platform-mpi/m-p/811905#M4016</guid>
      <dc:creator>xian-zhong_guous_cd-</dc:creator>
      <dc:date>2011-02-08T18:05:46Z</dc:date>
    </item>
    <item>
      <title>ScaLAPACK for platform_mpi</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/ScaLAPACK-for-platform-mpi/m-p/811906#M4017</link>
      <description>Sam,&lt;BR /&gt;I did not notice any significant overhead. In response to my question to HP-MPI developers in 2007 was:&lt;BR /&gt;&lt;PRE&gt;  As far as performance penalty, there IS a thin binary translation layer used, however it should not impact performance too much.&lt;BR /&gt;&lt;/PRE&gt;From simple ping-pong tests, I wasn't able to see any measurable effect of the wrappers either. &lt;BR /&gt;&lt;BR /&gt;Alexei&lt;BR /&gt;&lt;BR /&gt;P.S. For Windows, I did have to compile our own BLACS and ScaLAPACK. But then you can still link them with MKL.&lt;BR /&gt;&lt;BR /&gt;</description>
      <pubDate>Wed, 09 Feb 2011 14:18:34 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/ScaLAPACK-for-platform-mpi/m-p/811906#M4017</guid>
      <dc:creator>Alexei_Yakovlev</dc:creator>
      <dc:date>2011-02-09T14:18:34Z</dc:date>
    </item>
  </channel>
</rss>

