[petsc-dev] petsc-dev: ASM with sbaij matrix and block size=1 broken?

[Jed Brown]

On Fri, Nov 4, 2011 at 23:20, Brad Aagaard <baagaard at usgs.gov> wrote:

> I assume you mean src/ksp/ksp/examples/**tutorials/ex2
> I have not yet been able to duplicate the error with this example. Do you
> know of any other test cases that might be appropriate?
>

Any scalar symmetric problem can be done this way, but this problem is
fishy and doesn't make sense in any "standard" case.


>
> My PyLith simulation with -malloc and -malloc_debug flags chokes on line
> 933 of src/mat/impls/baij/mpi/**baijjob.c with a EXEC_BAD_ACCESS.
>

What do I need to do to reproduce this on my machine?


>
>
> 933               for (l=0; l<nzA; l++)   cols[imark+l] = cstart +
> cworkA[l];
>

I need to hunt around to find out why nzA is bigger than it should be (or
why some array is the wrong size).
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20111104/fc676550/attachment.html>