On Wed, 19 Mar 2008, Shengli Xu wrote:

> micro_fluid_loadcase11_assemble is the  performance log of calling the
> function of assembling stiffness matrix.

> | micro_fluid_loadcase11_assemble    1         5515.6511   5515.651089

You might want to do some more fine-grained performance logging.  The
whole point of PerfLog objects is to make it easier to pin down this
sort of problem.

You might also try running with oprofile, if you've got access to a
system on which that's convenient for you.

On the other hand, the information that subsequent assembles are fast
may be sufficient alone; it sounds to me like Ben's hunch is right and
you're trying to write sparse matrix values to a misallocated sparsity
pattern.  I'm not sure how that would occur, though, unless your
formulation is very non-typical.

What version of libMesh are you using?
---
Roy

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Libmesh-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-users

Reply via email to