<div dir="ltr">Antonio,<div><br></div><div>Thank you for your response. Will setting of this variable result in using the "sock" channel instead of the "nemesis" channel?</div><div><br></div><div>--Jiri</div>
</div><div class="gmail_extra"><br><br><div class="gmail_quote">On Fri, Sep 13, 2013 at 8:56 AM, Antonio J. Peña <span dir="ltr"><<a href="mailto:apenya@mcs.anl.gov" target="_blank">apenya@mcs.anl.gov</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><u></u>
<div style="font-family:'Sans Serif';font-size:9pt;font-weight:400;font-style:normal">
<p style="margin-top:0px;margin-bottom:0px;margin-left:0px;margin-right:0px;text-indent:0px"> </p>
<p style="margin-top:0px;margin-bottom:0px;margin-left:0px;margin-right:0px;text-indent:0px">You can set the MPIR_PARAM_CH3_NO_LOCAL environment variable to 1. Refer to the README.envvar file.</p><span class="HOEnZb"><font color="#888888">
<p style="margin-top:0px;margin-bottom:0px;margin-left:0px;margin-right:0px;text-indent:0px"> </p>
<p style="margin-top:0px;margin-bottom:0px;margin-left:0px;margin-right:0px;text-indent:0px"> Antonio</p></font></span><div><div class="h5">
<p style="margin-top:0px;margin-bottom:0px;margin-left:0px;margin-right:0px;text-indent:0px"> </p>
<p style="margin-top:0px;margin-bottom:0px;margin-left:0px;margin-right:0px;text-indent:0px"> </p>
<p style="margin-top:0px;margin-bottom:0px;margin-left:0px;margin-right:0px;text-indent:0px">On Friday, September 13, 2013 08:53:36 AM Jiri Simsa wrote:<br></p>
<p style="margin-top:12px;margin-bottom:0px;margin-left:40px;margin-right:40px;text-indent:0px">Pavan,</p>
<p style="margin-top:0px;margin-bottom:0px;margin-left:40px;margin-right:40px;text-indent:0px"><br></p>
<p style="margin-top:0px;margin-bottom:0px;margin-left:40px;margin-right:40px;text-indent:0px">Thank you for your answer. That's precisely what I was looking for. Any chance there is a way to force the intranode communication to use tcp?</p>
<p style="margin-top:0px;margin-bottom:0px;margin-left:40px;margin-right:40px;text-indent:0px"><br></p>
<p style="margin-top:0px;margin-bottom:0px;margin-left:40px;margin-right:40px;text-indent:0px">--Jiri</p>
<p style="margin-top:0px;margin-bottom:0px;margin-left:40px;margin-right:40px;text-indent:0px"><br></p>
<p style="margin-top:0px;margin-bottom:0px;margin-left:46px;margin-right:40px;text-indent:0px"><span style="font-family:'arial,sans-serif'">Within the node, it uses shared memory. Outside the node, it depends on the netmod you configured with. tcp is the default netmod.</span><span style="font-family:'arial, sans-serif'"><br>
</span><span style="font-family:'arial,sans-serif'"> -- Pavan</span><span style="font-family:'arial, sans-serif'"><br></span><span style="font-family:'arial,sans-serif'">On Sep 12, 2013, at 2:24 PM, Jiri Simsa wrote:</span><span style="font-family:'arial, sans-serif'"><br>
</span><span style="font-family:'arial,sans-serif'">> The high-order bit of my question is: What OS interface(s) does MPICH use to transfer data from one MPI process to another?<br>><br>><br>> On Thu, Sep 12, 2013 at 1:36 PM, Jiri Simsa <</span><a href="mailto:jsimsa@cs.cmu.edu" target="_blank"><span style="font-family:'arial,sans-serif';text-decoration:underline;color:#0057ae">jsimsa@cs.cmu.edu</span></a><span style="font-family:'arial,sans-serif'">> wrote:<br>
> Hello,<br>><br>> I have been trying to understand how MPICH implements collective operations. To do so, I have been reading the MPICH source code and stepping through mpiexec executions.<br>><br>> For the sake of this discussion, let's assume that all MPI processes are executed on the same computer using: mpiexec -n <n> <mpi_program><br>
><br>> This is my current abstract understanding of MPICH:<br>><br>> - mpiexec spawns a hydra_pmi_proxy process, which in turn spawns <n> instances of <mpi_program><br>> - hydra_pmi_proxy process uses socket pairs to communicate with the instances of <mpi_program><br>
><br>> I am not quite sure though what happens under the hoods when a collective operation, such as MPI_Allreduce, is executed. I have noticed that instances of <mpi_program> create and listen on a socket in the course of executing MPI_Allreduce but I am not sure who connects to these sockets. Any chance someone could describe the data flow inside of MPICH when a collective operation, such as MPI_Allreduce, is executed? Thanks!<br>
><br>> Best,<br>><br>> --Jiri Simsa<br>><br>> _______________________________________________<br>> discuss mailing list </span><a href="mailto:discuss@mpich.org" target="_blank"><span style="font-family:'arial,sans-serif';text-decoration:underline;color:#0057ae">discuss@mpich.org<br>
</span></a><span style="font-family:'arial,sans-serif'">> To manage subscription options or unsubscribe:<br>> </span><a href="https://lists.mpich.org/mailman/listinfo/discuss" target="_blank"><span style="font-family:'arial,sans-serif';text-decoration:underline;color:#0057ae">https://lists.mpich.org/mailman/listinfo/discuss</span></a><span style="font-family:'arial, sans-serif'"><br>
</span><span style="font-family:'arial,sans-serif'">--<br>Pavan Balaji<br></span><a href="http://www.mcs.anl.gov/~balaji" target="_blank"><span style="font-family:'arial,sans-serif';text-decoration:underline;color:#0057ae">http://www.mcs.anl.gov/~balaji</span></a></p>
<p style="margin-top:0px;margin-bottom:0px;margin-left:0px;margin-right:0px;text-indent:0px"><br><br></p></div></div></div></blockquote></div><br></div>