<?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 Hi in Intel® oneAPI Math Kernel Library</title>
    <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/xxmr2d-out-of-memory-even-with-64-bit-libraries/m-p/779918#M1377</link>
    <description>&lt;P&gt;Hi&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;

&lt;P&gt;still have problems distributing a 43496x43496 matrix to the slaves using pdgemr2d it looks to be related to&lt;/P&gt;

&lt;P&gt;&lt;A href="http://icl.cs.utk.edu/lapack-forum/viewtopic.php?t=491" target="_blank"&gt;http://icl.cs.utk.edu/lapack-forum/viewtopic.php?t=491&lt;/A&gt;&lt;/P&gt;

&lt;P&gt;and&lt;/P&gt;

&lt;P&gt;&lt;A href="https://icl.cs.utk.edu/lapack-forum/viewtopic.php?t=465" target="_blank"&gt;https://icl.cs.utk.edu/lapack-forum/viewtopic.php?t=465&lt;/A&gt;&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;

&lt;P&gt;will this be fixed soon?&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;

&lt;P&gt;TK&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;</description>
    <pubDate>Thu, 29 Jan 2015 08:57:14 GMT</pubDate>
    <dc:creator>Thomas_K_6</dc:creator>
    <dc:date>2015-01-29T08:57:14Z</dc:date>
    <item>
      <title>xxmr2d:out of memory even with 64-bit libraries</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/xxmr2d-out-of-memory-even-with-64-bit-libraries/m-p/779913#M1372</link>
      <description>Hello,&lt;BR /&gt;&lt;BR /&gt;I am writing a small fortran90 code to do (amongst othet things)&lt;BR /&gt;memory-distributed matrix matrix multiplication A*B on a computational cluster.&lt;BR /&gt;&lt;BR /&gt;For this, I use ScaLAPACK + BLACS .. in the MKL libraries.&lt;BR /&gt;The code works fast and fine for matrices &amp;lt;2GB (&amp;lt;16384x16384)&lt;BR /&gt;When I try to use the 64-bit libraries something goes wrong. &lt;BR /&gt;&lt;BR /&gt;The code compiles just fine with&lt;BR /&gt;&lt;BR /&gt;MKLPATH = /site/VERSIONS/intel-11.1u7/mkl/lib/em64t&lt;BR /&gt;&lt;BR /&gt;$(MKLPATH)/libmkl_scalapack_ilp64.a -Wl,--start-group $(MKLPATH)/libmkl_inte&lt;BR /&gt;l_ilp64.a $(MKLPATH)/libmkl_sequential.a $(MKLPATH)/libmkl_core.a $(MKLPATH)/libmkl_blacs_op&lt;BR /&gt;enmpi_ilp64.a -Wl,--end-group &lt;BR /&gt;&lt;BR /&gt;and the ifort mpi compile wrapper &lt;BR /&gt;&lt;BR /&gt;mpif90 &lt;BR /&gt;&lt;BR /&gt;with the extra flags &lt;BR /&gt;&lt;BR /&gt;-mcmodel=medium -i-dynamic -i8&lt;BR /&gt;&lt;BR /&gt;During runtime I get an error from the pdgemr2d routine:&lt;BR /&gt;&lt;BR /&gt;&amp;gt;&amp;gt;&amp;gt; xxmr2d:out of memory&lt;BR /&gt;&lt;BR /&gt;Naturally, the same happens if I use the pdgeadd routine for&lt;BR /&gt;the task of block-cyclic distribution.&lt;BR /&gt;&lt;BR /&gt;When browsing other forums I found this solution:&lt;BR /&gt;********************************&lt;BR /&gt;You need to modify the file REDIST/SRC/pgemraux.c and change&lt;BR /&gt;void *&lt;BR /&gt;mr2d_malloc(n)&lt;BR /&gt;unsigned int n;&lt;BR /&gt;{&lt;BR /&gt;&lt;BR /&gt;to&lt;BR /&gt;void *&lt;BR /&gt;mr2d_malloc(n)&lt;BR /&gt;unsigned long int n;&lt;BR /&gt;{&lt;BR /&gt;********************************&lt;BR /&gt;&lt;BR /&gt;I.e, a large enough workspace isn't allocated...&lt;BR /&gt;However, this is the exact meaning of using ilp64 libraries i guess?!&lt;BR /&gt;&lt;BR /&gt;Could someone point me in the right direction or perhaps let me know what I am&lt;BR /&gt;doing wrong? I would really appreciate the help :) Oh, the multiplication routine&lt;BR /&gt;pdgemm works fine for large (&amp;gt;2gb) matrices, so the program really uses the &lt;BR /&gt;64-bit libraries.&lt;BR /&gt;&lt;BR /&gt;MORE DETAILS AND CODE excerpt:&lt;BR /&gt;&lt;BR /&gt;! create the root-node context where the entire A and B matrices&lt;BR /&gt;! reside in memory, called gloA and gloB&lt;BR /&gt;call sl_init (rootNodeContext, 1, 1) &lt;BR /&gt;! prep the descriptors for A B and C&lt;BR /&gt;! the C descriptor is used later for moving &lt;BR /&gt;! the resulting C sub arrays back to the root node&lt;BR /&gt;&lt;BR /&gt;if (Iam==0) then &lt;BR /&gt; nr_gloA_row = numroc( m, m, myrow, 0, nprow ) &lt;BR /&gt; nr_gloB_row = numroc( k, k, myrow, 0, nprow ) &lt;BR /&gt; nr_gloC_row = numroc( m, m, myrow, 0, nprow ) &lt;BR /&gt; call descinit( desc_gloA, m, k, m, k, 0, 0, &amp;amp; &lt;BR /&gt; rootNodeContext, max(1, nr_gloA_row), info) &lt;BR /&gt; call descinit( desc_gloB, k, n, k, n, 0, 0, &amp;amp; &lt;BR /&gt; rootNodeContext, max(1, nr_gloB_row), info) &lt;BR /&gt; call descinit (desc_gloC, m, n, m, n, 0, 0, &amp;amp; &lt;BR /&gt; rootNodeContext, max(1, nr_gloC_row), info) &lt;BR /&gt; &lt;BR /&gt;else &lt;BR /&gt; desc_gloA(1:9) = 0 &lt;BR /&gt; desc_gloB(1:9) = 0 &lt;BR /&gt; desc_gloC(1:9) = 0 &lt;BR /&gt; desc_gloA(2) = -1 &lt;BR /&gt; desc_gloB(2) = -1 &lt;BR /&gt; desc_gloC(2) = -1 &lt;BR /&gt;end if &lt;BR /&gt; &lt;BR /&gt;call pdgemr2d( m, k, gloA, one, one, desc_gloA, locA, &amp;amp; &lt;BR /&gt; one, one, desc_locA, desc_locA( 2 )) &lt;BR /&gt;call pdgemr2d( k, n, gloB, one, one, desc_gloB, locB, &amp;amp; &lt;BR /&gt; one, one, desc_locB, desc_locB( 2 )) &lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;All the best,&lt;BR /&gt;&lt;BR /&gt; Andreas&lt;BR /&gt;&lt;BR /&gt;</description>
      <pubDate>Mon, 22 Nov 2010 16:35:48 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/xxmr2d-out-of-memory-even-with-64-bit-libraries/m-p/779913#M1372</guid>
      <dc:creator>jaeks</dc:creator>
      <dc:date>2010-11-22T16:35:48Z</dc:date>
    </item>
    <item>
      <title>xxmr2d:out of memory even with 64-bit libraries</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/xxmr2d-out-of-memory-even-with-64-bit-libraries/m-p/779914#M1373</link>
      <description>Andreas,&lt;DIV&gt;it looks like the defect in iLP64 implementation. We will check and let you know.&lt;/DIV&gt;&lt;DIV&gt;--Gennady&lt;/DIV&gt;</description>
      <pubDate>Tue, 23 Nov 2010 12:12:52 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/xxmr2d-out-of-memory-even-with-64-bit-libraries/m-p/779914#M1373</guid>
      <dc:creator>Gennady_F_Intel</dc:creator>
      <dc:date>2010-11-23T12:12:52Z</dc:date>
    </item>
    <item>
      <title>xxmr2d:out of memory even with 64-bit libraries</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/xxmr2d-out-of-memory-even-with-64-bit-libraries/m-p/779915#M1374</link>
      <description>Hi Gennady,&lt;BR /&gt;&lt;BR /&gt;I was just wondering if you managed to reproduce my problem, and if so, did you find a fix to it?!&lt;BR /&gt;&lt;BR /&gt;All the best,&lt;BR /&gt;&lt;BR /&gt; Andreas</description>
      <pubDate>Thu, 09 Dec 2010 14:39:17 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/xxmr2d-out-of-memory-even-with-64-bit-libraries/m-p/779915#M1374</guid>
      <dc:creator>jaeks</dc:creator>
      <dc:date>2010-12-09T14:39:17Z</dc:date>
    </item>
    <item>
      <title>xxmr2d:out of memory even with 64-bit libraries</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/xxmr2d-out-of-memory-even-with-64-bit-libraries/m-p/779916#M1375</link>
      <description>Andreas,&lt;BR /&gt;&lt;BR /&gt;I could not reproduce the described problem. All works fine with 17000x17000. Could you please provide test case?&lt;BR /&gt;&lt;BR /&gt;Best,&lt;BR /&gt;andrew</description>
      <pubDate>Tue, 08 Feb 2011 13:56:43 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/xxmr2d-out-of-memory-even-with-64-bit-libraries/m-p/779916#M1375</guid>
      <dc:creator>Andrei_Moskalev__Int</dc:creator>
      <dc:date>2011-02-08T13:56:43Z</dc:date>
    </item>
    <item>
      <title>which mpi library is this ?</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/xxmr2d-out-of-memory-even-with-64-bit-libraries/m-p/779917#M1376</link>
      <description>&lt;P&gt;which mpi library is this ?&lt;/P&gt;

&lt;P&gt;32 bit integer or 64 bit integer ? OpenMPI, IntelMPI, MPICH2?&lt;/P&gt;

&lt;P&gt;it sounds to be related to "Out of memory error with Cpzgemr2d" topic 509048&lt;/P&gt;

&lt;P&gt;Best Regards&lt;/P&gt;

&lt;P&gt;Thomas Kjaergaard&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;</description>
      <pubDate>Mon, 28 Apr 2014 12:51:05 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/xxmr2d-out-of-memory-even-with-64-bit-libraries/m-p/779917#M1376</guid>
      <dc:creator>Thomas_K_6</dc:creator>
      <dc:date>2014-04-28T12:51:05Z</dc:date>
    </item>
    <item>
      <title>Hi</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/xxmr2d-out-of-memory-even-with-64-bit-libraries/m-p/779918#M1377</link>
      <description>&lt;P&gt;Hi&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;

&lt;P&gt;still have problems distributing a 43496x43496 matrix to the slaves using pdgemr2d it looks to be related to&lt;/P&gt;

&lt;P&gt;&lt;A href="http://icl.cs.utk.edu/lapack-forum/viewtopic.php?t=491" target="_blank"&gt;http://icl.cs.utk.edu/lapack-forum/viewtopic.php?t=491&lt;/A&gt;&lt;/P&gt;

&lt;P&gt;and&lt;/P&gt;

&lt;P&gt;&lt;A href="https://icl.cs.utk.edu/lapack-forum/viewtopic.php?t=465" target="_blank"&gt;https://icl.cs.utk.edu/lapack-forum/viewtopic.php?t=465&lt;/A&gt;&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;

&lt;P&gt;will this be fixed soon?&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;

&lt;P&gt;TK&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;

&lt;P&gt;&amp;nbsp;&lt;/P&gt;</description>
      <pubDate>Thu, 29 Jan 2015 08:57:14 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/xxmr2d-out-of-memory-even-with-64-bit-libraries/m-p/779918#M1377</guid>
      <dc:creator>Thomas_K_6</dc:creator>
      <dc:date>2015-01-29T08:57:14Z</dc:date>
    </item>
  </channel>
</rss>

