<?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 MKL Threads- BLAS level 2 routines in Intel® oneAPI Math Kernel Library</title>
    <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/MKL-Threads-BLAS-level-2-routines/m-p/816688#M4327</link>
    <description>&lt;P&gt;1)Yes, this routine is
threaded internally, butthe main question
what the scalability numbers are you expecting to see?&lt;/P&gt;

&lt;P&gt;In the mostly cases, for
the sparse matrixes, these are the cache and memory bandwidth problems.&lt;/P&gt;

&lt;P&gt;2)Please see here how to &lt;A href="http://software.intel.com/en-us/articles/intel-mkl-link-line-advisor/"&gt;link
MKL&lt;/A&gt; more properly&lt;/P&gt;</description>
    <pubDate>Fri, 07 Oct 2011 12:36:37 GMT</pubDate>
    <dc:creator>Gennady_F_Intel</dc:creator>
    <dc:date>2011-10-07T12:36:37Z</dc:date>
    <item>
      <title>MKL Threads- BLAS level 2 routines</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/MKL-Threads-BLAS-level-2-routines/m-p/816687#M4326</link>
      <description>Multithreading does not seem to work in my program where I am using mkl_dcsrmv subroutine to multiply large sparse matrices. I have tried using "mkl_set_num_threads(num_threads)" to set the number of threads to be used. The program gives correct output but the performance doesn't change as I change the number of threads.&lt;DIV&gt;&lt;SPAN style="font-family: Verdana, Arial, Helvetica, sans-serif;"&gt;According to mkl manual, mkl version &amp;gt;10.0 should maximum possible number of threads on processor, but that does not seem to be the case.&lt;BR /&gt;&lt;/SPAN&gt;&lt;DIV&gt;&lt;/DIV&gt;&lt;DIV&gt;Platform: Intel Xeon E5520 (4 cores/8 threads).&lt;/DIV&gt;&lt;DIV&gt;&lt;/DIV&gt;&lt;DIV&gt;#include "omp.h"&lt;/DIV&gt;&lt;DIV&gt;...&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV&gt;...&lt;/DIV&gt;&lt;DIV&gt;mkl_dcsrmv("N", &amp;amp;M, &amp;amp;N, α, "G**C", val, (int *)col, (int *)ptr, (int *)ptre, vec_aligned, α, y_vec);&lt;/DIV&gt;&lt;DIV&gt;...&lt;/DIV&gt;&lt;DIV&gt;&lt;/DIV&gt;&lt;DIV&gt;&lt;DIV&gt;Compile:&lt;/DIV&gt;&lt;DIV&gt;icc -mkl -I /opt/intel/Compiler/11.1/069/mkl/include/-L$(MKLROOT)/lib/intel64 -lmkl_intel_lp64 -lmkl_intel_thread -lmkl_core -openmp -lpthread&lt;SOURCE_FILES&gt; &lt;OBJECT_FILES&gt; -o run_mkl&lt;/OBJECT_FILES&gt;&lt;/SOURCE_FILES&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV&gt;&lt;/DIV&gt;&lt;DIV&gt;Is mkl_dcsrmv a threaded routine?&lt;/DIV&gt;</description>
      <pubDate>Fri, 07 Oct 2011 06:13:55 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/MKL-Threads-BLAS-level-2-routines/m-p/816687#M4326</guid>
      <dc:creator>kris_nagar</dc:creator>
      <dc:date>2011-10-07T06:13:55Z</dc:date>
    </item>
    <item>
      <title>MKL Threads- BLAS level 2 routines</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/MKL-Threads-BLAS-level-2-routines/m-p/816688#M4327</link>
      <description>&lt;P&gt;1)Yes, this routine is
threaded internally, butthe main question
what the scalability numbers are you expecting to see?&lt;/P&gt;

&lt;P&gt;In the mostly cases, for
the sparse matrixes, these are the cache and memory bandwidth problems.&lt;/P&gt;

&lt;P&gt;2)Please see here how to &lt;A href="http://software.intel.com/en-us/articles/intel-mkl-link-line-advisor/"&gt;link
MKL&lt;/A&gt; more properly&lt;/P&gt;</description>
      <pubDate>Fri, 07 Oct 2011 12:36:37 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/MKL-Threads-BLAS-level-2-routines/m-p/816688#M4327</guid>
      <dc:creator>Gennady_F_Intel</dc:creator>
      <dc:date>2011-10-07T12:36:37Z</dc:date>
    </item>
    <item>
      <title>MKL Threads- BLAS level 2 routines</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/MKL-Threads-BLAS-level-2-routines/m-p/816689#M4328</link>
      <description>In addition to what Gennady said, you might find it interesting (if using dynamic libiomp) to set&lt;BR /&gt;LD_PRELOAD=&lt;MKL_INSTALLATION_PATH&gt;/libiompprof5.so&lt;BR /&gt;and look at the guide.gvs file generated.&lt;/MKL_INSTALLATION_PATH&gt;</description>
      <pubDate>Fri, 07 Oct 2011 13:35:33 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/MKL-Threads-BLAS-level-2-routines/m-p/816689#M4328</guid>
      <dc:creator>TimP</dc:creator>
      <dc:date>2011-10-07T13:35:33Z</dc:date>
    </item>
    <item>
      <title>MKL Threads- BLAS level 2 routines</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/MKL-Threads-BLAS-level-2-routines/m-p/816690#M4329</link>
      <description>Thanks both of you for the reply.&lt;DIV&gt;I expect a speedup of 3-4x when going from serial to multithreaded code. And I am using matrices of size 8mx8m with 118 million entries.&lt;/DIV&gt;&lt;DIV&gt;From the guide.gvs file, I found that my program is not using 8 threads even when I try to set the threads manually.&lt;DIV&gt;I have another program where I use sgemm routine to multiply dense matrices. And that code uses multithreading. I am using the same settings and platform for both the programs.&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV&gt;&lt;/DIV&gt;&lt;DIV&gt;Thanks again!&lt;/DIV&gt;</description>
      <pubDate>Sun, 09 Oct 2011 18:00:54 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/MKL-Threads-BLAS-level-2-routines/m-p/816690#M4329</guid>
      <dc:creator>kris_nagar</dc:creator>
      <dc:date>2011-10-09T18:00:54Z</dc:date>
    </item>
    <item>
      <title>MKL Threads- BLAS level 2 routines</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/MKL-Threads-BLAS-level-2-routines/m-p/816691#M4330</link>
      <description>Finally got it working.. I just updated the icc version and now its invoking all the threads available.</description>
      <pubDate>Tue, 11 Oct 2011 20:20:11 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/MKL-Threads-BLAS-level-2-routines/m-p/816691#M4330</guid>
      <dc:creator>kris_nagar</dc:creator>
      <dc:date>2011-10-11T20:20:11Z</dc:date>
    </item>
  </channel>
</rss>

