Intermittent failure of fem Assembler python unit tests when petsc is compiled with debugging
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
DOLFIN |
New
|
Undecided
|
Unassigned |
Bug Description
Apologies if this is not a bug but a problem with my set-up but I've been seeing intermittent failures of the fem Assembler python unit test when using petsc with debugging enabled and thought it might be of interest.
The buildbot queue showing this can be found at:
http://
or for a specific failure instance:
http://
(this is a nightly build of lp:dolfin with no local modifications on ubuntu natty 11.04 64 bit).
I can reproduce manually but only when using petsc compiled with debugging enabled.
The error message begins:
One or more unit tests failed for fem (Assembler, Python):
Testing class Assembler
-------
*** Warning: Form::coloring does not properly consider form type.
Coloring mesh.
Start Boost coloring.
Boost coloring finished.
Number of colors: 28
*** Warning: Form::coloring does not properly consider form type.
.Coloring mesh.
Start Boost coloring.
Boost coloring finished.
Number of colors: 28
Renumbering mesh by cell colors.
<Mesh of topological dimension 3 (tetrahedra) with 125 vertices and 384 cells, ordered>
*** Warning: Form::coloring does not properly consider form type.
Coloring mesh.
Start Boost coloring.
Boost coloring finished.
Number of colors: 28
*** Warning: Form::coloring does not properly consider form type.
.*** Warning: OpenMpAssembler
*** Warning: Form::coloring does not properly consider form type.
Coloring mesh.
Start Boost coloring.
Boost coloring finished.
Number of colors: 15
*** Warning: Form::coloring does not properly consider form type.
Coloring mesh.
Start Boost coloring.
Boost coloring finished.
Number of colors: 7
[0]PETSC ERROR: PetscMallocVali
[0]PETSC ERROR: Memory at address 0x219f530 is corrupted
[0]PETSC ERROR: Probably write past beginning or end of array
[0]PETSC ERROR: Last intact block allocated in EventRegLogCreate() line 36 in src/sys/
[0]PETSC ERROR: -------
[0]PETSC ERROR: Memory corruption!
[0]PETSC ERROR: !
[0]PETSC ERROR: -------
[0]PETSC ERROR: Petsc Release Version 3.1.0, Patch 8, unknown
[0]PETSC ERROR: See docs/changes/
[0]PETSC ERROR: See docs/faq.html for hints about trouble shooting.
[0]PETSC ERROR: See docs/index.html for manual pages.
[0]PETSC ERROR: -------
[0]PETSC ERROR: Unknown Name on a linux-gnu named diana by buildbot Tue Nov 22 03:02:01 2011
[0]PETSC ERROR: Libraries linked from /usr/local/
[0]PETSC ERROR: Configure run at Sat Nov 19 00:16:01 2011
[0]PETSC ERROR: Configure options --with-
[0]PETSC ERROR: -------
[0]PETSC ERROR: PetscMallocVali
[0]PETSC ERROR: PetscFListDestroy() line 267 in src/sys/dll/reg.c
[0]PETSC ERROR: PetscHeaderDest
[0]PETSC ERROR: MatDestroy() line 888 in src/mat/
[0]PETSC ERROR: PetscMallocVali
[0]PETSC ERROR: Memory at address 0x219f530 is corrupted
[0]PETSC ERROR: Probably write past beginning or end of array
[0]PETSC ERROR: Last intact block allocated in EventRegLogCreate() line 36 in src/sys/
[0]PETSC ERROR: -------
[0]PETSC ERROR: Memory corruption!
[0]PETSC ERROR: !
Does this only happen with the OpenMpAssembler? PETSc is not officially thread-safe, but we assume/hope that via colouring that we avoid problems.