If you do &quot;man orte_hosts&quot;, you&#39;ll see a full explanation of how the various machinefile options work.<div><br></div><div>The default mapper doesn&#39;t do any type of sorting - it is a round-robin mapper that just works its way through the provided nodes. We don&#39;t reorder them in any way.</div>
<div><br></div><div>However, it does depend on the number of slots we are told each node has, so that might be what you are encountering. If you do a --display-map and send it along, I might be able to spot the issue.</div>
<div><br></div><div>Thanks</div><div>Ralph</div><div><br><br><div class="gmail_quote">On Fri, Jun 19, 2009 at 1:35 PM, Rajesh Sudarsan <span dir="ltr">&lt;<a href="mailto:rsudarsan@gmail.com">rsudarsan@gmail.com</a>&gt;</span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">Hi,<br>
<br>
I tested a simple hello world program on 5 nodes each with dual<br>
quad-core processors. I noticed that openmpi does not always follow<br>
the order of the processors indicated in the machinefile. Depending<br>
upon the number of processors requested, openmpi does some type of<br>
sorting to find the best node fit for a particular job and runs on<br>
them.  Is there a way to make openmpi to turn off this sorting and<br>
strictly follow the order indicated in the machinefile?<br>
<br>
mpiexec supports three options to specify the machinefile -<br>
default-machinefile, hostfile, and machinefile. Can anyone tell what<br>
is the difference between these three options?<br>
<br>
Any help would be greatly appreciated.<br>
<br>
Thanks,<br>
Rajesh<br>
_______________________________________________<br>
users mailing list<br>
<a href="mailto:users@open-mpi.org">users@open-mpi.org</a><br>
<a href="http://www.open-mpi.org/mailman/listinfo.cgi/users" target="_blank">http://www.open-mpi.org/mailman/listinfo.cgi/users</a><br>
</blockquote></div><br></div>