[petsc-dev] Petsc has generated inconsistent data

[Pierre Jolivet]

On Thu, 16 Mar 2017 19:46:34 -0500, Barry Smith wrote:
>> On Mar 16, 2017, at 10:26 AM, Hong <hzhang at mcs.anl.gov> wrote:
>>
>> I fixed typos in maint branch:
>> 
>> https://bitbucket.org/petsc/petsc/commits/d4ed72bbe47df8d3eb83fd589394594abed431db
>>
>> I'll check the error produced by './ex12 -pc_type gasm -ksp_view'.
>>
>> Hong
>
>    Pierre and Hong,
>
>      This may not be fixable. In GASM submatrices lie on overlapping
> collections of MPI processes thus the -mat_view likely won't work
> because it is called also on these submatrices. Even if it did work I
> suspect the entries for the various submatrices would be printed out
> in unreadable random order.
>
>     Are you intentionally trying to see what the submatrices look 
> like?

No, I was just tinkering around with the options and found this 
undocumented "feature".
Of course, I can live without a fix for this.

Thanks,
Pierre

>    Barry
>
>>
>> On Thu, Mar 16, 2017 at 3:31 AM, Pierre Jolivet 
>> <pierre.jolivet at enseeiht.fr> wrote:
>> Hello,
>> Using master:
>> > cd $PETSC_DIR/src/ksp/ksp/examples/tutorials && make ex12 && 
>> mpirun -np 1 ./ex12 -pc_type gasm -mat_view -ksp_view
>> [..]
>> [0]PETSC ERROR: Petsc has generated inconsistent data
>> [0]PETSC ERROR: Called more times than 
>> PetscViewerASCIIPushSynchronized()
>>
>> Also, in mumps.c (cf. 
>> http://lists.mcs.anl.gov/pipermail/petsc-dev/2017-February/020440.html):
>> “ICNTL(7) (sequentia matrix ordering):” => should be spelled 
>> “sequential"
>> "ICNTL(8) (scalling strategy):” => should be spelled “scaling”
>> "ICNTL(19) (Shur complement info):” => should be spelled “Schur”
>>
>> Thanks,
>> Pierre
>>