[petsc-users] Error to TaoSolve() call

[Khai Hong Pham]

Jed,

Thank you for your advise. I'm solving multi-field problem which were
solved by static-splitting schema. I tried to reused my code, that's why I
tried to use the nested matrix and vector. I'm trying to use the contiguous
vector and matrix storage now. I appreciate if you could give me any advise
on managing the matrix and vector (based on the skeleton of the existed
code for separate problems) to get good performance.

Khai


On Tue, Aug 12, 2014 at 11:54 AM, Jed Brown <jed at jedbrown.org> wrote:

> Khai Hong Pham <khaipham at utexas.edu> writes:
>
> > Hello,
> >
> > I call TaosSolve() in my code. The solution vector is a nested and
> Jacobian
> > matrix is nested too. It shows me the error:
> >
> > [0]PETSC ERROR:
> > ------------------------------------------------------------------------
> > [0]PETSC ERROR: Caught signal number 11 SEGV: Segmentation Violation,
> > probably memory access out of range
> > [0]PETSC ERROR: Try option -start_in_debugger or
> -on_error_attach_debugger
> > [0]PETSC ERROR: or see
> > http://www.mcs.anl.gov/petsc/documentation/faq.html#valgrind[0]PETSC
> ERROR:
> > or try http://valgrind.org on GNU/linux and Apple Mac OS X to find
> memory
> > corruption errors
> > [1]PETSC ERROR: [0]PETSC ERROR: likely location of problem given in stack
> > below
> > [0]PETSC ERROR: ---------------------  Stack Frames
> > ------------------------------------
> > [0]PETSC ERROR: Note: The EXACT line numbers in the stack are not
> available,
> > [0]PETSC ERROR:       INSTEAD the line number of the start of the
> function
> > [0]PETSC ERROR:       is given.
> > [0]PETSC ERROR: [0] VecRestoreArray_Nest line 731
> > /home/khp/bin/petsc-3.5.1/src/vec/vec/impls/nest/vecnest.c
> > [0]PETSC ERROR: [0] VecRestoreArray line 1656
> > /home/khp/bin/petsc-3.5.1/src/vec/vec/interface/rvector.c
> > [0]PETSC ERROR: [0] VecMedian line 806
> > /home/khp/bin/petsc-3.5.1/src/vec/vec/utils/projection.c
> > [0]PETSC ERROR: [0] TaoSolve_SSILS line 52
> > /home/khp/bin/petsc-3.5.1/src/tao/complementarity/impls/ssls/ssils.c
> > [0]PETSC ERROR: [0] TaoSolve line 171
> > /home/khp/bin/petsc-3.5.1/src/tao/interface/taosolver.c
> >
> > I debugged with valgrind and received the following message:
> >
> > ==19424== Invalid read of size 8
> > ==19424==    at 0x54089E4: VecRestoreArray_Nest (vecnest.c:746)
> > ==19424==    by 0x53F5F6F: VecRestoreArray (rvector.c:1666)
> > ==19424==    by 0x54BE874: VecMedian (projection.c:855)
>
> Can you try 'next'?  I think this bad restore has been fixed.
>
> I *strongly* recommend *not* using VecNest.  99% of people that think
> they should use it definitely should not use it.  Just use contiguous
> storage for your vectors.  VecNest creates a lot of problems that are
> almost always larger than the perceived concern that people try to avoid
> by using VecNest.
>
> > ==19424==    by 0x6247B74: TaoSolve_SSILS (ssils.c:59)
> > ==19424==    by 0x620739E: TaoSolve (taosolver.c:188)
> > ==19424==    by 0x40491C: main (main.cpp:265)
> > ==19424==  Address 0x0 is not stack'd, malloc'd or (recently) free'd
> >
> > It looks like the bug relates to the nested vector. Could you please give
> > me a hint to locate the bug? Thanks.
> >
> > Khai
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20140812/8f24f684/attachment-0001.html>