[petsc-dev] good partitioning packages?

[Fande Kong]

On Fri, Jun 1, 2018 at 5:40 PM, Matthew Knepley <knepley at gmail.com> wrote:

> On Fri, Jun 1, 2018 at 7:09 PM, Smith, Barry F. <bsmith at mcs.anl.gov>
> wrote:
>
>>
>>    Fande,
>>
>>        This is a great question. I am forwarding it to Mike Heroux who
>> has a high level position in the ECP; because I have similar concerns and
>> also don't have a good answer. ParMetis does indeed have a poor license and
>> essentially no support. Perhaps Mike has some ideas.
>>
>
> Also, the parallel scalability is crappy (sorry George).
>

Do we know the underneath reason why the parallels scalability is poor?
Poor algorithm? Poor implementation?

Here is a list of partitioning packages
https://en.wikipedia.org/wiki/Graph_partition#cite_note-patoh-16

Anybody has experiences on any of the listed packages?

Fande,


> Bill Gropp has proposed in the past developing a new
> partitioner along more scalable lines, such as the Teng algorithm used in
> Padma's 2013 SC paper
> (https://dl.acm.org/citation.cfm?id=2503280). Jed favors a multilevel
> approach which I do not understand.
> Its a shame that all the development time that went into PT-Scotch could
> not produce a scalable, open
> partitioner. Also, the label-push stuff seems only to work well for highly
> connected graphs, not meshes.
>
>    Matt
>
>
>>
>>    Barry
>>
>>
>>
>>
>> On Jun 1, 2018, at 5:46 PM, Kong, Fande <fande.kong at inl.gov> wrote:
>>
>> Hi Developers,
>>
>> I have introduced MatPartitioning interface to MOOSE. It is working
>> great, and we can use all external partitioning packages via a simple
>> interface.
>>
>> But here is a concern. Almost all the packages are not under development
>> any more. Does this make a bug fix more difficult in the future.  Also some
>> of them have bad licenses.
>>
>> I was wondering there is any other partitioning package in the community?
>>
>> Thanks,
>>
>> Fande,
>>
>>
>>
>
>
> --
> 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
>
> https://www.cse.buffalo.edu/~knepley/ <http://www.caam.rice.edu/~mk51/>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20180601/b7f90a3a/attachment.html>