<?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 error using mkl as default blas implementation on linux in Intel® oneAPI Math Kernel Library</title>
    <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/error-using-mkl-as-default-blas-implementation-on-linux/m-p/961050#M15903</link>
    <description>&lt;P&gt;Hi,&lt;/P&gt;
&lt;P&gt;I don't know if this makes much sense, but I tried to use update-alternatives on linux to make all the symlinks for libblas.so, libblas.so.3gf, liblapack.so and liblapack.so.3gf point to libmkl_rt.so. I was hoping that this way numpy will use the mkl implementation of blas operations without needing to recompile it. But, as it turns out, the loader is having some trouble with the multi-threading library, which I expected to be libmkl_gnu_thread.so instead of&amp;nbsp;libmkl_intel_thread.so.&lt;/P&gt;
&lt;P&gt;python -c "from numpy import zeros; a = zeros((17,17)); b = a.dot(a)"&lt;/P&gt;
&lt;P&gt;python: symbol lookup error: /opt/intel/mkl/lib/intel64/libmkl_intel_thread.so: undefined symbol: omp_get_num_procs&lt;/P&gt;
&lt;P&gt;I am using mkl 11.0 update 5, which I upgraded from mkl 11.0 update 2, which didn't work either. Everything is on 64bits. I also noticed that for matrices smaller than (17, 17) ther error doesn't occur.&lt;/P&gt;
&lt;P&gt;Thanks,&lt;/P&gt;
&lt;P&gt;Bogdan&amp;nbsp;&lt;/P&gt;</description>
    <pubDate>Thu, 25 Jul 2013 21:58:39 GMT</pubDate>
    <dc:creator>Bogdan_B_2</dc:creator>
    <dc:date>2013-07-25T21:58:39Z</dc:date>
    <item>
      <title>error using mkl as default blas implementation on linux</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/error-using-mkl-as-default-blas-implementation-on-linux/m-p/961050#M15903</link>
      <description>&lt;P&gt;Hi,&lt;/P&gt;
&lt;P&gt;I don't know if this makes much sense, but I tried to use update-alternatives on linux to make all the symlinks for libblas.so, libblas.so.3gf, liblapack.so and liblapack.so.3gf point to libmkl_rt.so. I was hoping that this way numpy will use the mkl implementation of blas operations without needing to recompile it. But, as it turns out, the loader is having some trouble with the multi-threading library, which I expected to be libmkl_gnu_thread.so instead of&amp;nbsp;libmkl_intel_thread.so.&lt;/P&gt;
&lt;P&gt;python -c "from numpy import zeros; a = zeros((17,17)); b = a.dot(a)"&lt;/P&gt;
&lt;P&gt;python: symbol lookup error: /opt/intel/mkl/lib/intel64/libmkl_intel_thread.so: undefined symbol: omp_get_num_procs&lt;/P&gt;
&lt;P&gt;I am using mkl 11.0 update 5, which I upgraded from mkl 11.0 update 2, which didn't work either. Everything is on 64bits. I also noticed that for matrices smaller than (17, 17) ther error doesn't occur.&lt;/P&gt;
&lt;P&gt;Thanks,&lt;/P&gt;
&lt;P&gt;Bogdan&amp;nbsp;&lt;/P&gt;</description>
      <pubDate>Thu, 25 Jul 2013 21:58:39 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/error-using-mkl-as-default-blas-implementation-on-linux/m-p/961050#M15903</guid>
      <dc:creator>Bogdan_B_2</dc:creator>
      <dc:date>2013-07-25T21:58:39Z</dc:date>
    </item>
    <item>
      <title>&gt;&gt;.../intel/mkl/lib/intel64</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/error-using-mkl-as-default-blas-implementation-on-linux/m-p/961051#M15904</link>
      <description>&amp;gt;&amp;gt;.../intel/mkl/lib/intel64/libmkl_intel_thread.so: undefined symbol: &lt;STRONG&gt;omp_get_num_procs&lt;/STRONG&gt;

It is defined in these libraries:

libiomp5md
libiomp5mt
libiompprof5md
libiompprof5mt
libiompstubs5md
libiompstubs5mt</description>
      <pubDate>Fri, 26 Jul 2013 01:13:36 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/error-using-mkl-as-default-blas-implementation-on-linux/m-p/961051#M15904</guid>
      <dc:creator>SergeyKostrov</dc:creator>
      <dc:date>2013-07-26T01:13:36Z</dc:date>
    </item>
    <item>
      <title>Shouldn't one of the</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/error-using-mkl-as-default-blas-implementation-on-linux/m-p/961052#M15905</link>
      <description>&lt;P&gt;Shouldn't one of the appropriate libraries be loaded by libmkl_rt.so?&lt;/P&gt;</description>
      <pubDate>Fri, 26 Jul 2013 05:36:17 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/error-using-mkl-as-default-blas-implementation-on-linux/m-p/961052#M15905</guid>
      <dc:creator>Bogdan_B_2</dc:creator>
      <dc:date>2013-07-26T05:36:17Z</dc:date>
    </item>
    <item>
      <title>yes, it will load libiomp via</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/error-using-mkl-as-default-blas-implementation-on-linux/m-p/961053#M15906</link>
      <description>&lt;P&gt;yes, it will load libiomp via dlopen == &amp;gt; you need to add &amp;lt;composer_xe_install_path&amp;gt;\compiler/lib/intel64\ to your systems paths.&lt;/P&gt;
&lt;P&gt;&lt;/P&gt;</description>
      <pubDate>Fri, 26 Jul 2013 07:05:23 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/error-using-mkl-as-default-blas-implementation-on-linux/m-p/961053#M15906</guid>
      <dc:creator>Gennady_F_Intel</dc:creator>
      <dc:date>2013-07-26T07:05:23Z</dc:date>
    </item>
    <item>
      <title>Thanks a lot. I got it fixed</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/error-using-mkl-as-default-blas-implementation-on-linux/m-p/961054#M15907</link>
      <description>&lt;P&gt;Thanks a lot. I got it fixed by adding the path to loader's search paths in /etc/ld.so.conf.d&lt;/P&gt;</description>
      <pubDate>Sat, 27 Jul 2013 13:02:42 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/error-using-mkl-as-default-blas-implementation-on-linux/m-p/961054#M15907</guid>
      <dc:creator>Bogdan_B_2</dc:creator>
      <dc:date>2013-07-27T13:02:42Z</dc:date>
    </item>
    <item>
      <title>&gt;&gt;...Shouldn't one of the</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/error-using-mkl-as-default-blas-implementation-on-linux/m-p/961055#M15908</link>
      <description>&amp;gt;&amp;gt;...Shouldn't one of the appropriate libraries be loaded by libmkl_rt.so?

I checked dependencies and the answer is Yes. However, if a search path is Not defined ( this is what you had and fixed already ) then the loader will fail to load that module. Thanks for the update.</description>
      <pubDate>Sat, 27 Jul 2013 23:45:01 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/error-using-mkl-as-default-blas-implementation-on-linux/m-p/961055#M15908</guid>
      <dc:creator>SergeyKostrov</dc:creator>
      <dc:date>2013-07-27T23:45:01Z</dc:date>
    </item>
    <item>
      <title>&gt;&gt;&gt;&gt;...Shouldn't one of the</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/error-using-mkl-as-default-blas-implementation-on-linux/m-p/961056#M15909</link>
      <description>&amp;gt;&amp;gt;&amp;gt;&amp;gt;...Shouldn't one of the appropriate libraries be loaded by libmkl_rt.so?
&amp;gt;&amp;gt;
&amp;gt;&amp;gt;I checked dependencies and the answer is Yes.

I was talking about &lt;STRONG&gt;libiomp5m?&lt;/STRONG&gt; modules.</description>
      <pubDate>Sat, 27 Jul 2013 23:47:22 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/error-using-mkl-as-default-blas-implementation-on-linux/m-p/961056#M15909</guid>
      <dc:creator>SergeyKostrov</dc:creator>
      <dc:date>2013-07-27T23:47:22Z</dc:date>
    </item>
  </channel>
</rss>

