<p dir="ltr">Order your unknowns appropriately and set the local sizes to match your distribution.</p>
<div class="gmail_quote">On Apr 28, 2013 4:16 PM, "Panruo Wu" <<a href="mailto:armiuswu@gmail.com">armiuswu@gmail.com</a>> wrote:<br type="attribution"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div dir="ltr">Thanks Jed!<div><br></div><div>I understand that MatCreateMPIAdj takes adjacency information</div><div>and partition the matrix for me; what if I don't want it to partition</div><div>
for me, that I already have a particular partitioning?</div><div><br></div><div>I guess my question is, how do I tell Petsc to use my own partitioning?</div><div><br></div><div>Thanks,</div><div>
Panruo</div></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Sun, Apr 28, 2013 at 5:26 AM, Jed Brown <span dir="ltr"><<a href="mailto:jedbrown@mcs.anl.gov" target="_blank">jedbrown@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"><div><div>Panruo Wu <<a href="mailto:armiuswu@gmail.com" target="_blank">armiuswu@gmail.com</a>> writes:<br>
<br>
> Hello,<br>
><br>
> I have a question about the matrix distribution in Petsc.<br>
> Can I define the distribution pattern as the output<br>
> of graph partition software like METIS? Pointers<br>
> to documentation/code about matrix distribution in Petsc<br>
> would be very helpful.<br>
<br>
</div></div><a href="http://www.mcs.anl.gov/petsc/petsc-current/docs/manualpages/Mat/MatCreateMPIAdj.html" target="_blank">http://www.mcs.anl.gov/petsc/petsc-current/docs/manualpages/Mat/MatCreateMPIAdj.html</a><br>
<a href="http://www.mcs.anl.gov/petsc/petsc-current/docs/manualpages/MatOrderings/MatPartitioningCreate.html" target="_blank">http://www.mcs.anl.gov/petsc/petsc-current/docs/manualpages/MatOrderings/MatPartitioningCreate.html</a><br>
<br>
After applying the partitioning, relabel your mesh (or whatever is the<br>
source of your problem), migrate that data, and create a new matrix<br>
using the new distribution.<br>
</blockquote></div><br></div>
</blockquote></div>