[petsc-dev] who ever broke next with DMCreateDomainDecomposition_DA stuff please fix

[Barry Smith]

   Having crashes in next week after week, month after month looks bad.


On Sep 15, 2014, at 2:45 PM, Matthew Knepley <knepley at gmail.com> wrote:

> On Mon, Sep 15, 2014 at 2:36 PM, Barry Smith <bsmith at mcs.anl.gov> wrote:
> 
>    Ok, get it out of next then, whatever it is.
> 
>    Yes, everyone but you is checking the next builds before moving anything to master.
> 
> I mean I did not think that test was meaningful. When I looked at the numbers to try and use it
> for verification, it did not make sense.
> 
>   Matt
>  
> 
>    Barry
> 
> On Sep 15, 2014, at 2:31 PM, Matthew Knepley <knepley at gmail.com> wrote:
> 
> > On Mon, Sep 15, 2014 at 2:11 PM, Barry Smith <bsmith at mcs.anl.gov> wrote:
> >
> >   Whoever broke
> >
> > > [1]PETSC ERROR: #1 DMCreateDomainDecomposition_DA() line 425 in /sandbox/petsc/petsc.clone/src/dm/impls/da/dadd.c
> > > Configure options --with-debugging=1 --download-f2cblaslapack=1 --with-precision=__float128 --with-clanguage=cxx --with-debugging=0 --with-no-output -PETSC_ARCH=arch-linux-opt-cxx-quad -PETSC_DIR=/sandbox/petsc/petsc.clone
> > > [2]PETSC ERROR: #1 DMCreateDomainDecomposition_DA() line 425 in /sandbox/petsc/petsc.clone/src/dm/impls/da/dadd.c
> > > Configure options --with-debugging=1 --download-f2cblaslapack=1 --with-precision=__float128 --with-clanguage=cxx --with-debugging=0 --with-no-output -PETSC_ARCH=arch-linux-opt-cxx-quad -PETSC_DIR=/sandbox/petsc/petsc.clone
> > > [3]PETSC ERROR: #1 DMCreateDomainDecomposition_DA() line 425 in /sandbox/petsc/petsc.clone/src/dm/impls/da/dadd.c
> > > [3]PETSC ERROR: #2 DMCreateDomainDecomposition() line 1357 in /sandbox/petsc/petsc.clone/src/dm/interface/dm.c
> >
> >   http://ftp.mcs.anl.gov/pub/petsc/nightlylogs/archive/2014/09/15/examples_next_arch-linux-opt-cxx-quad_cg.log
> >
> > please fix it! It has been broken for a while now.
> >
> > I am not completely sure it was doing what we expect in the first place, and I think the interface is
> > bad and does not generalize to unstructured. I want to fix it (fairly badly), but I got sidetracked on
> > things that people are demanding. Its fine with me if we revert that branch from next, but I did not
> > think anyone was looking at this.
> >
> >   Thanks,
> >
> >      Matt
> >
> >    Thanks
> >
> >     Barry
> >
> >
> >
> >
> > --
> > What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.
> > -- Norbert Wiener
> 
> 
> 
> 
> -- 
> What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.
> -- Norbert Wiener