<?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 Maybe I explained not so well in Intel® MPI Library</title>
    <link>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944744#M2829</link>
    <description>Maybe I explained not so well my problem: when I use mpirun -n 8 .HEART_MR, I expect to launch a parallel job on eight cores, and single process knows how many processes are running and its rank. Indeed, I obtain eight processes, that are inpidendent from each other.

In other words, my executable apply a domain decomposition over the number of the cores that I choose (if I use mpirun from openmpi). In the wrong case (with intel mpi), each process is associated to the whole computational domain.

Giacomo Rossi
PhD Student, Space Engineer
Mechanical and Aerospace Engineering Department
University of Rome "Sapienza"</description>
    <pubDate>Fri, 09 Nov 2012 08:18:09 GMT</pubDate>
    <dc:creator>Giacomo_R_</dc:creator>
    <dc:date>2012-11-09T08:18:09Z</dc:date>
    <item>
      <title>Bad mpirun behaviour</title>
      <link>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944742#M2827</link>
      <description>&lt;P&gt;I've installed intel mpi, use mpi-selector to choose intel mpi (installed in impi folder in /opt/intel/), but when I launch a job with&lt;/P&gt;
&lt;BLOCKQUOTE&gt;
&lt;P&gt;mpirun -n 8 ./HEART_MR&lt;/P&gt;
&lt;P&gt;&lt;/P&gt;
&lt;/BLOCKQUOTE&gt;
&lt;P&gt;where HEART_MR is the executable, I obtain 8 processes running each one on a core, when I want &amp;nbsp;a single process running on 8 cores. If I use openmpi, things happen how I expect to be.&lt;/P&gt;
&lt;P&gt;Can you help me?&lt;/P&gt;
&lt;P&gt;My machine has an eight core intel Xeon processor.&lt;/P&gt;
&lt;P&gt;&lt;/P&gt;
&lt;P&gt;&lt;SPAN&gt;Giacomo Rossi&lt;/SPAN&gt;&lt;BR /&gt;&lt;SPAN&gt;PhD Student, Space Engineer&lt;/SPAN&gt;&lt;BR /&gt;&lt;SPAN&gt;Mechanical and Aerospace Engineering Department&lt;/SPAN&gt;&lt;BR /&gt;&lt;SPAN&gt;University of Rome "Sapienza"&lt;/SPAN&gt;&lt;/P&gt;</description>
      <pubDate>Thu, 08 Nov 2012 13:42:12 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944742#M2827</guid>
      <dc:creator>Giacomo_R_</dc:creator>
      <dc:date>2012-11-08T13:42:12Z</dc:date>
    </item>
    <item>
      <title>Hi Giacomo,</title>
      <link>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944743#M2828</link>
      <description>Hi Giacomo,

Using -n specifies the number of processes to launch.  Is your application using threading?  You can use I_MPI_PIN_DOMAIN to select which cores are available to the process.

Sincerely,
James Tullos
Technical Consulting Engineer
Intel® Cluster Tools</description>
      <pubDate>Thu, 08 Nov 2012 14:37:45 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944743#M2828</guid>
      <dc:creator>James_T_Intel</dc:creator>
      <dc:date>2012-11-08T14:37:45Z</dc:date>
    </item>
    <item>
      <title>Maybe I explained not so well</title>
      <link>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944744#M2829</link>
      <description>Maybe I explained not so well my problem: when I use mpirun -n 8 .HEART_MR, I expect to launch a parallel job on eight cores, and single process knows how many processes are running and its rank. Indeed, I obtain eight processes, that are inpidendent from each other.

In other words, my executable apply a domain decomposition over the number of the cores that I choose (if I use mpirun from openmpi). In the wrong case (with intel mpi), each process is associated to the whole computational domain.

Giacomo Rossi
PhD Student, Space Engineer
Mechanical and Aerospace Engineering Department
University of Rome "Sapienza"</description>
      <pubDate>Fri, 09 Nov 2012 08:18:09 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944744#M2829</guid>
      <dc:creator>Giacomo_R_</dc:creator>
      <dc:date>2012-11-09T08:18:09Z</dc:date>
    </item>
    <item>
      <title>Hi Giacomo,</title>
      <link>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944745#M2830</link>
      <description>Hi Giacomo,

I see.  Please send me the output from the following commands:

env | grep I_MPI
mpirun -n 8 ./hello &lt;I&gt;hello can be any MPI hello world application&lt;/I&gt;

Sincerely,
James Tullos
Technical Consulting Engineer
Intel® Cluster Tools</description>
      <pubDate>Fri, 09 Nov 2012 11:41:01 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944745#M2830</guid>
      <dc:creator>James_T_Intel</dc:creator>
      <dc:date>2012-11-09T11:41:01Z</dc:date>
    </item>
    <item>
      <title>Here you have the results of</title>
      <link>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944746#M2831</link>
      <description>Here you have the results of the commands that you have requested to me:

nunzio@ALTAIR:~&amp;gt; mpirun -n 8 ./hello_world
 Hello World :-)
 Hello World :-)
 Hello World :-)
 Hello World :-)
 Hello World :-)
 Hello World :-)
 Hello World :-)
 Hello World :-)

nunzio@ALTAIR:~&amp;gt; env | grep I_MPI
I_MPI_ROOT=/opt/intel2/impi/4.1.0.024

Here some additional informations; this is a simple "hello world2 program with mpi

program hello
include 'mpif.h'
integer rank, size, ierror, tag, status(MPI_STATUS_SIZE)

   call MPI_INIT(ierror)
   call MPI_COMM_SIZE(MPI_COMM_WORLD, size, ierror)
   call MPI_COMM_RANK(MPI_COMM_WORLD, rank, ierror)
   print*, 'node', rank, ': Hello world'
   call MPI_FINALIZE(ierror)
end

and this is the result that I obtained when launch mpirun -n 8 prova_mpi:

nunzio@ALTAIR:~&amp;gt; mpirun -n 8 ./prova_mpi 
 node           0 : Hello world
 node           0 : Hello world
 node           0 : Hello world
 node           0 : Hello world
 node           0 : Hello world
 node           0 : Hello world
 node           0 : Hello world
 node           0 : Hello world


Thank you!

Giacomo Rossi</description>
      <pubDate>Fri, 09 Nov 2012 12:53:07 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944746#M2831</guid>
      <dc:creator>Giacomo_R_</dc:creator>
      <dc:date>2012-11-09T12:53:07Z</dc:date>
    </item>
    <item>
      <title>Are you running in a job</title>
      <link>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944747#M2832</link>
      <description>Are you running in a job scheduler?  Please try the following.  There is a test program included with Intel® MPI, in /opt/intel2/mpi/4.1.0.024/test.  Compile any of the files in that folder, and run it with

mpirun -n 8 -verbose -genv I_MPI_DEBUG 5 ./a.out &amp;gt; output.txt

Please attach the output.txt file.

James.</description>
      <pubDate>Fri, 09 Nov 2012 12:58:12 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944747#M2832</guid>
      <dc:creator>James_T_Intel</dc:creator>
      <dc:date>2012-11-09T12:58:12Z</dc:date>
    </item>
    <item>
      <title>Unfortunately I haven't this</title>
      <link>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944748#M2833</link>
      <description>Unfortunately I haven't this directory...

nunzio@ALTAIR:/opt/intel2/impi/4.1.0.024&amp;gt; lt
totale 152
-rw-r--r-- 1 root root  9398 31 ago 16.48 README.txt
-rw-r--r-- 1 root root 28556 31 ago 16.48 Doc_Index.html
-rw-r--r-- 1 root root  2770 31 ago 16.48 redist-rt.txt
-rw-r--r-- 1 root root 28728 31 ago 16.48 mpi-rtEULA.txt
-rw-r--r-- 1 root root   491  7 set 13.12 mpi-rtsupport.txt
-rwxr-xr-x 1 root root 41314  7 set 13.12 uninstall.sh
-rw-r--r-- 1 root root  3036  8 nov 11.51 uninstall.log
lrwxrwxrwx 1 root root     8  8 nov 11.55 etc -&amp;gt; ia32/etc
lrwxrwxrwx 1 root root     8  8 nov 11.55 bin -&amp;gt; ia32/bin
drwxr-xr-x 5 root root  4096  8 nov 11.55 ia32
lrwxrwxrwx 1 root root     8  8 nov 11.55 lib -&amp;gt; ia32/lib
drwxr-xr-x 3 root root  4096  8 nov 11.55 data
drwxr-xr-x 4 root root  4096  8 nov 11.55 doc
lrwxrwxrwx 1 root root    11  8 nov 11.55 etc64 -&amp;gt; intel64/etc
lrwxrwxrwx 1 root root    11  8 nov 11.55 bin64 -&amp;gt; intel64/bin
drwxr-xr-x 5 root root  4096  8 nov 11.55 intel64
lrwxrwxrwx 1 root root    11  8 nov 11.55 lib64 -&amp;gt; intel64/lib
drwxr-xr-x 5 root root  4096  8 nov 11.55 mic
-rw-r--r-- 1 root root   340  8 nov 11.55 impi.uninstall.config</description>
      <pubDate>Fri, 09 Nov 2012 13:09:42 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944748#M2833</guid>
      <dc:creator>Giacomo_R_</dc:creator>
      <dc:date>2012-11-09T13:09:42Z</dc:date>
    </item>
    <item>
      <title>Do you have the full SDK, or</title>
      <link>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944749#M2834</link>
      <description>Do you have the full SDK, or only the runtime?</description>
      <pubDate>Fri, 09 Nov 2012 19:37:19 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944749#M2834</guid>
      <dc:creator>James_T_Intel</dc:creator>
      <dc:date>2012-11-09T19:37:19Z</dc:date>
    </item>
    <item>
      <title>I've only the runtime.</title>
      <link>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944750#M2835</link>
      <description>I've only the runtime.</description>
      <pubDate>Mon, 12 Nov 2012 08:09:05 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944750#M2835</guid>
      <dc:creator>Giacomo_R_</dc:creator>
      <dc:date>2012-11-12T08:09:05Z</dc:date>
    </item>
    <item>
      <title>Ok, the program you run will</title>
      <link>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944751#M2836</link>
      <description>Ok, the program you run will need to be compiled with the Intel® MPI Library to run with the Intel® MPI Library.  We have some binary compatibility with MPICH2, so you might be able to use our MPI to run an MPICH2 compiled program as well.</description>
      <pubDate>Tue, 13 Nov 2012 13:15:44 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-MPI-Library/Bad-mpirun-behaviour/m-p/944751#M2836</guid>
      <dc:creator>James_T_Intel</dc:creator>
      <dc:date>2012-11-13T13:15:44Z</dc:date>
    </item>
  </channel>
</rss>

