Mm, i&#39;m not sure. Suppose this:<div><br></div><div>$pragma omp parallel num_thread(1)</div><div>{</div><div><meta http-equiv="content-type" content="text/html; charset=utf-8">hwloc_set_cpubind(*topology, set,  HWLOC_CPUBIND_THREAD | HWLOC_CPUBIND_STRICT |   HWLOC_CPUBIND_NOMEMBIND);</div>
<div>}</div><div><br></div><div>is equivalent to?</div><div><br></div><div><meta http-equiv="content-type" content="text/html; charset=utf-8"><div>$pragma omp parallel num_thread(1)</div><div>{</div></div><div><meta http-equiv="content-type" content="text/html; charset=utf-8">hwloc_set_cpubind(*topology, set,  HWLOC_CPUBIND_THREAD);<br>
hwloc_set_cpubind(*topology, set, HWLOC_CPUBIND_STRICT);<br>hwloc_set_cpubind(*topology, set, HWLOC_CPUBIND_NOMEMBIND);</div><div><br></div><div><meta http-equiv="content-type" content="text/html; charset=utf-8">}</div><div>
<br></div><div><br></div><div>You said <meta http-equiv="content-type" content="text/html; charset=utf-8">HWLOC_CPUBIND_STRICT bind process and memory. Why also the memory? It should strictly design which CPU will assigned to a process/thread, or does more?<br>
<br><div class="gmail_quote">2011/8/2 Samuel Thibault <span dir="ltr">&lt;<a href="mailto:samuel.thibault@inria.fr">samuel.thibault@inria.fr</a>&gt;</span><br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
Gabriele Fatigati, le Tue 02 Aug 2011 16:23:12 +0200, a écrit :<br>
<div class="im">&gt; hwloc_set_cpubind(*topology, set,  HWLOC_CPUBIND_THREAD | HWLOC_CPUBIND_STRICT<br>
&gt; |   HWLOC_CPUBIND_NOMEMBIND);<br>
&gt;<br>
&gt; is it possible do multiple call to hwloc_set_cpubind passing each flag per<br>
&gt; time? <br>
&gt;<br>
&gt; hwloc_set_cpubind(*topology, set,  HWLOC_CPUBIND_THREAD);<br>
&gt; hwloc_set_cpubind(*topology, set, HWLOC_CPUBIND_STRICT);<br>
&gt; hwloc_set_cpubind(*topology, set, HWLOC_CPUBIND_NOMEMBIND);<br>
&gt;<br>
&gt; or only the last have effect?<br>
<br>
</div>Err, it will simply do the three operations, i.e. first bind the current<br>
thread and memory, then strictly bind the whole process and memory, and<br>
eventually bind the process but not memory (but it will still bound<br>
since it was by the second call).<br>
<br>
Samuel<br>
_______________________________________________<br>
hwloc-users mailing list<br>
<a href="mailto:hwloc-users@open-mpi.org">hwloc-users@open-mpi.org</a><br>
<a href="http://www.open-mpi.org/mailman/listinfo.cgi/hwloc-users" target="_blank">http://www.open-mpi.org/mailman/listinfo.cgi/hwloc-users</a><br>
</blockquote></div><br><br clear="all"><br>-- <br>Ing. Gabriele Fatigati<br><br>Parallel programmer<br><br>CINECA Systems &amp; Tecnologies Department<br><br>Supercomputing Group<br><br>Via Magnanelli 6/3, Casalecchio di Reno (BO) Italy<br>
<br><a href="http://www.cineca.it">www.cineca.it</a>                    Tel:   +39 051 6171722<br><br>g.fatigati [AT] <a href="http://cineca.it">cineca.it</a>           <br>
</div>