<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">Paul:</div><div class="gmail_quote">It might be caused by our changes in default shift strategy.</div><div class="gmail_quote">We previously used '-pc_factor_shift_type NONZERO' for ilu, then changed to '-pc_factor_shift_type NONE'.</div><div class="gmail_quote">For your test, I get</div><div class="gmail_quote"><div class="gmail_quote">./ex10 -f0 test.mat -rhs 0 -pc_type asm -pc_asm_overlap 12 -sub_pc_type ilu -sub_pc_factor_mat_ordering_type 1wd -sub_pc_factor_levels 4 -ksp_converged_reason -sub_pc_factor_shift_type NONZERO</div><div class="gmail_quote"><br></div><div class="gmail_quote">Linear solve converged due to CONVERGED_RTOL iterations 2</div><div class="gmail_quote">Number of iterations =   2</div><div class="gmail_quote">Residual norm 0.0116896</div><div class="gmail_quote"><br></div><div class="gmail_quote">with </div><div class="gmail_quote"><div class="gmail_quote">'-sub_pc_factor_shift_type INBLOCKS':</div><div class="gmail_quote"><br></div><div class="gmail_quote">Linear solve converged due to CONVERGED_RTOL iterations 2</div><div class="gmail_quote">Number of iterations =   2</div><div class="gmail_quote">Residual norm 0.00603736</div></div></div><div class="gmail_quote"><br></div><div class="gmail_quote">I guess your previous run might use one of these options.</div><div class="gmail_quote"><br></div><div class="gmail_quote">Hong</div><div class="gmail_quote"><br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex"><div dir="ltr">Thanks Hong,<br><div class="gmail_extra"><br><div class="gmail_quote"><span class="">On Thu, Jan 21, 2016 at 12:16 PM, Hong <span dir="ltr"><<a href="mailto:hzhang@mcs.anl.gov" target="_blank">hzhang@mcs.anl.gov</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex"><div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">Paul :</div><div class="gmail_quote">Using petsc-dev (we recently added feature for better displaying convergence behavior),</div></div></div></blockquote><div><br></div></span><div>OK, good to know, thanks.</div><span class=""><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex"><div dir="ltr"><div class="gmail_extra"><div class="gmail_quote"> I found that <span style="font-size:12.8px">'-sub_pc_factor_mat_ordering_</span><span style="font-size:12.8px">type 1wd' causes zero pivot:</span></div></div></div></blockquote><div><br></div></span><div>I figured it was something along these lines. So, just so I'm clear, likely this zero pivot was always there with this mat ordering (i.e. no mat ordering bits actually changed between 3.5 and 3.6) and this is a reflection of increased consistency checking in the newer PETSc?</div><div><br></div><div>Thanks much, </div><span class=""><font color="#888888"><div><br></div><div>Paul</div></font></span></div></div></div>
</blockquote></div><br></div></div>