Stefano :<br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hong, INTERLACED is the default since <a href="http://petsc.cs.iit.edu/petsc/petsc-dev/rev/67df4fdba5bd" target="_blank">http://petsc.cs.iit.edu/petsc/petsc-dev/rev/67df4fdba5bd</a><br>
<br>See also <a href="http://petsc.cs.iit.edu/petsc/petsc-dev/rev/c04222d17a3d" target="_blank">http://petsc.cs.iit.edu/petsc/petsc-dev/rev/c04222d17a3d</a></blockquote><div><br></div><div>Hmm, you are correct :-(</div><div>
 </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5"><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">

Hong, in petsc-dev, default is interlaced. I ran some tests, I got malloc errors when inserting mat values using subcomm of contiguous type. I think that PCRedundant with PETSC_SUBCOMM_CONTIGUOUS is buggy.</blockquote></div>
</div></div></blockquote><div> </div><div>PETSC_SUBCOMM_CONTIGUOUS was introduced for  multiprocessor bjacobi preconditioner,</div><div>which has not been used/tested in other situations, likely causes problems.</div><div>
Can you send us a stand-alone code that reproduces the malloc errors?</div><div><br></div><div>Hong</div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb">
<div class="h5"><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div>
<div><br><br><div class="gmail_quote">
2012/10/15 Hong Zhang <span dir="ltr"><<a href="mailto:hzhang@mcs.anl.gov" target="_blank">hzhang@mcs.anl.gov</a>></span><br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">


Stefano :<div class="gmail_quote"><div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><br>is it possible to use PCRedundant with PETSC_SUBCOMM_CONTIGUOUS as subcomm type?<br>



</blockquote></div><div>I think "PETSC_SUBCOMM_CONTIGUOUS" is the default for PCRedundant.</div><span><font color="#888888"><div><br></div><div>Hong</div></font></span></div>
</blockquote></div><br><br clear="all"><br></div></div><span><font color="#888888">-- <br>Stefano<br>
</font></span></blockquote></div><br><br clear="all"><br></div></div><span class="HOEnZb"><font color="#888888">-- <br>Stefano<br>
</font></span></blockquote></div><br>