Hello,<br><br>I am on a Mac, and I didn't use "--with-shared=0" when I configured PETSc originally, but I just re-ran it with that option and got the same error. I even completely deleted and re-installed PETSc because I wasn't positive it had overwritten itself properly. It still crashed at exactly the same place. I've attached configure_vars for you.<br>
<br>The particle tracker slipped my mind in the last email, but it's definitely in the top three... this is why I want to get the profiler working!<br><br>Thanks again,<br><br>Neil<br><br><div class="gmail_quote">On Tue, Jan 13, 2009 at 3:49 PM, Walter Landry <span dir="ltr"><<a href="mailto:walter@geodynamics.org" target="_blank">walter@geodynamics.org</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;"><div>"Neil de Laplante" <<a href="mailto:neildlp@mit.edu" target="_blank">neildlp@mit.edu</a>> wrote:<br>
> Hello Walter,<br>
><br>
> I got the previous problem figured out; it was actually a PETSc issue hidden<br>
> in not having the correct fortran compiler installed, making for a pretty<br>
> cryptic error message. Anyway, I have a new problem:<br>
><br>
> When I run "make -d" in the Gale home directory, it crashes when it can't find<br>
> a StGermain file deep in the code (I think):<br>
<br>
</div>I do not think the "-d" option has any effect, since it is not passed<br>
on to scons.<br>
<div><br>
> — unknown option: -rpath=/Users/oeit/Documents/NdLp_CUDA/PETSc/petsc-2.3.2-p10<br>
> /lib/darwin9.6.0-c-debug<br>
> collect2: ld returned 1 exit status<br>
> scons: *** [build/StGermain/Base/FlattenXML/src/FlattenXML] Error 1<br>
> scons: building terminated because of errors.<br>
> Reaping losing child 0x001095f0 PID 28833<br>
> make: *** [all] Error 2<br>
> Removing child 0x001095f0 PID 28833 from chain.<br>
><br>
> Sure enough, there's no /src/FlattenXML directory or file, so I don't know<br>
> what it was looking for or why. Any suggestions?<br>
<br>
</div>It looks like it does not like the "-rpath" option. I am guessing<br>
that you are using a Mac? Did you configure petsc with the option<br>
"--with-shared=0"? Can I see a copy of "configure_vars" in your Gale<br>
directory?<br>
<div><br>
> For a bit of context, I'm trying to recompile it myself to add debugging flags<br>
> so that I can profile the code. Once I've got that figured out I'm going to<br>
> port the bottlenecks, which I'm guessing are the finite element integrals and<br>
> matrix solvers, to compute on Graphics Processing Units (GPUs) using CUDA.<br>
<br>
</div>Actually, I think the biggest bottlenecks are particle calculations<br>
and matrix assembly. But I have not done a rigorous analysis.<br>
<div><div></div><div><br>
Cheers,<br>
Walter Landry<br>
<a href="mailto:walter@geodynamics.org" target="_blank">walter@geodynamics.org</a><br>
</div></div></blockquote></div><br><br clear="all"><br>-- <br>Neil de Laplante<br><br>PhD Candidate in Geophysics<br>Earth, Atmospheric and Planetary Sciences<br>Massachusetts Institute of Technology<br>77 Massachusetts Ave 54-615<br>
Cambridge, MA, USA 02139<br>(617) 999-6528<br><a href="mailto:neildlp@mit.edu" target="_blank">neildlp@mit.edu</a><br>