[petsc-dev] MatSetColoring

Peter Brune prbrune at gmail.com
Fri May 17 15:01:47 CDT 2013


https://bitbucket.org/petsc/petsc/commits/all/tip/prbrune/snes-jacobiancolorfixchanges
setting user coloring back to how it was in 3.3; I'll change the
manual to describe why this is bad (doesn't work with grid sequencing or
FAS or a number of other things) and then merge into maint if it's OK.

After that, we should take MatGetColoring and MatSetColoring and make them
be a convenient interface.  Can we remove the remaining Adifor stuff and
start from scratch on this?

Having a good parallel MatColoring would make a lot of these problems less
severe.  We should write a simple one and then go from there.

- Peter


On Fri, May 17, 2013 at 9:37 AM, Peter Brune <prbrune at gmail.com> wrote:

> So I'm going to be putting this back into maint to unbreak user colorings
> for good.  It appears that there's still MatSetValuesAdifor in there, and
> that it still refers to MatSetColoring()... was this just missed and should
> it be removed?
>
> - Peter
>
>
> On Thu, May 16, 2013 at 5:34 PM, Barry Smith <bsmith at mcs.anl.gov> wrote:
>
>>
>> On May 16, 2013, at 5:09 PM, Jed Brown <jedbrown at mcs.anl.gov> wrote:
>>
>> > This interface was removed, but MatSetColoring_MPIAIJ and
>> > MatSetColoring_SeqAIJ remain.
>> >
>> > commit bcaeba4d41d6ca6c6dc4189db20683073a9959ce
>> > Author: Barry Smith <bsmith at mcs.anl.gov>
>> > Date:   Sun Dec 30 18:31:13 2012 -0600
>> >
>> >    removed adic stuff
>> >
>> >    Hg-commit: 99933b6f66a48c78b00a3952edfaf6191df65340
>> >
>> >
>> > Peter and I think that MatSetColoring would be useful in case the user
>> > can compute a parallel coloring that they would like SNES to use.  Shall
>> > we revive the interface and make MatGetColoring use it if available?
>>
>>    I can't remember anything about this stuff. I likely was overzealous
>> in removing stuff when I bagged some of the adic stuff.
>>
>>    Barry
>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20130517/cd700120/attachment.html>


More information about the petsc-dev mailing list