<?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 Thanks a lot for your help, in Intel® oneAPI Math Kernel Library</title>
    <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/3D-hybrid-MPI-OpenMP-Fourier-transform/m-p/982416#M17501</link>
    <description>Thanks a lot for your help, the manual is clear on how to handle higher dimensional transforms. So if I understood correctly, if I compile with the appropriate options, I will be able to run a 3D transform using hybrid MPI/Openmp?

Giorgos</description>
    <pubDate>Thu, 03 Jan 2013 17:53:36 GMT</pubDate>
    <dc:creator>giorgos17</dc:creator>
    <dc:date>2013-01-03T17:53:36Z</dc:date>
    <item>
      <title>3D hybrid MPI/OpenMP Fourier transform</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/3D-hybrid-MPI-OpenMP-Fourier-transform/m-p/982414#M17499</link>
      <description>&lt;P&gt;Hello,&lt;/P&gt;
&lt;P&gt;I would like to use a hybrid MPI/OpenMP Fourier transform routine on a Linux cluster with parallel composer 2011 update 4 and IMPI 4.0.1.007. I found out that hybrid mode is supported for 1D transforms, but could not find any documentation. Is it possible to do a 3D FFT in hybrid mode using MKL on this particular system?&lt;/P&gt;
&lt;P&gt;Thanks,&lt;/P&gt;
&lt;P&gt;Giorgos&lt;/P&gt;</description>
      <pubDate>Wed, 02 Jan 2013 12:27:01 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/3D-hybrid-MPI-OpenMP-Fourier-transform/m-p/982414#M17499</guid>
      <dc:creator>giorgos17</dc:creator>
      <dc:date>2013-01-02T12:27:01Z</dc:date>
    </item>
    <item>
      <title>I have great news for you.</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/3D-hybrid-MPI-OpenMP-Fourier-transform/m-p/982415#M17500</link>
      <description>I have great news for you. Check out page 2754 of the MKL Reference Manual. 

&lt;A href="http://software.intel.com/en-us/articles/intel-math-kernel-library-documentation" target="_blank"&gt;http://software.intel.com/en-us/articles/intel-math-kernel-library-documentation&lt;/A&gt;

There is an entire section devoted to explaining how MKL's Cluster FFT works and how it handles higher dimensionality. Please take a look through that and let me know if you have any more questions.</description>
      <pubDate>Wed, 02 Jan 2013 19:23:31 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/3D-hybrid-MPI-OpenMP-Fourier-transform/m-p/982415#M17500</guid>
      <dc:creator>Noah_C_Intel</dc:creator>
      <dc:date>2013-01-02T19:23:31Z</dc:date>
    </item>
    <item>
      <title>Thanks a lot for your help,</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/3D-hybrid-MPI-OpenMP-Fourier-transform/m-p/982416#M17501</link>
      <description>Thanks a lot for your help, the manual is clear on how to handle higher dimensional transforms. So if I understood correctly, if I compile with the appropriate options, I will be able to run a 3D transform using hybrid MPI/Openmp?

Giorgos</description>
      <pubDate>Thu, 03 Jan 2013 17:53:36 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/3D-hybrid-MPI-OpenMP-Fourier-transform/m-p/982416#M17501</guid>
      <dc:creator>giorgos17</dc:creator>
      <dc:date>2013-01-03T17:53:36Z</dc:date>
    </item>
    <item>
      <title>Yes, what is most interesting</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/3D-hybrid-MPI-OpenMP-Fourier-transform/m-p/982417#M17502</link>
      <description>Yes, what is most interesting here is that with a simple function call the management of MPI and OpenMP are handled for you. I believe in your initial inquiry, you were trying to direct all of the communcation between nodes and between cores yourself, but in the case of MKL Cluster FFT, everything is "up-leveled" for you. So you are passing in that function FFT function call a wide variety of information about your problem and how your MPI procs are setup and then the MKL library will utilize MPI/OpenMP to solve the problem on the nodes you have specified.</description>
      <pubDate>Thu, 03 Jan 2013 19:22:27 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/3D-hybrid-MPI-OpenMP-Fourier-transform/m-p/982417#M17502</guid>
      <dc:creator>Noah_C_Intel</dc:creator>
      <dc:date>2013-01-03T19:22:27Z</dc:date>
    </item>
  </channel>
</rss>

