<div dir="ltr">If you ignore entries for the rows and columns associated with BCs, but do not<br>eliminate them from the ordering, do you remember to put something on the<br>diagonal of the Jacobian?<br><br> Matt<br><br><div class="gmail_quote">
On Tue, Aug 12, 2008 at 11:08 AM, Cristian Tibirna <span dir="ltr"><<a href="mailto:ctibirna@giref.ulaval.ca">ctibirna@giref.ulaval.ca</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<br>
Hello<br>
<br>
We recently decided to use PETSc's negative indices features in a FEM code<br>
that originated long time ago, long before PETSc 2.0.24 (which, it seems,<br>
introduced negative indices for matrices and vectors).<br>
<br>
I attempted to write an algorithm that produces negative indices at assembly<br>
for matrix lines and columns and vector entries imposed by Dirichlet boundary<br>
conditions. All seemed straightforward up until the final testing.<br>
<br>
Our FEM solver is written "in terms of correction". Thus, we get the solution<br>
as a correction vector, which should have null entries corresponding to the<br>
original Dirichlet-imposed DOFs.<br>
<br>
To my surprise, even if I initialize the correction vector by filling it with<br>
zeros, I get the Dirichlet-imposed entries (and then a few more) set to<br>
NaN-s.<br>
<br>
The bigger surprise is that, even if I set manually to zero the<br>
Dirichlet-imposed entries in the correction vector _after_ the solver<br>
finishes, I still have the "few more" entries set to NaN.<br>
<br>
All the values in the correction vector that are not nan-s are the exact<br>
solution of the system, as expected. Thus, I conclude that the solver<br>
assembles the matrix and the residual vector correctly and returns the right<br>
solution of the matricial system.<br>
<br>
Any idea what should I look at for debugging this?<br>
<br>
NOTE: I solve the system by PREONLY (LU preconditioner, of course).<br>
<br>
Thanks in advance for any help.<br>
<font color="#888888"><br>
--<br>
Cristian Tibirna (1-418-) 656-2131 / 4340<br>
Laval University - Quebec, CAN ... <a href="http://www.giref.ulaval.ca/%7Ectibirna" target="_blank">http://www.giref.ulaval.ca/~ctibirna</a><br>
Research professional at GIREF ... <a href="mailto:ctibirna@giref.ulaval.ca">ctibirna@giref.ulaval.ca</a><br>
<br>
</font></blockquote></div><br><br clear="all"><br>-- <br>What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>-- Norbert Wiener<br>
</div>