[petsc-dev] proposed minor PetscPartitioner changes

[Hapla Vaclav]

Yes, MatPartitioning is a much more general thing and should stay in PETSc in long term. PetscPartitioner works just with DMPlex. My PETSCPARTITIONERMATPARTITIONING is a wrapper of MatPartitioner into PetscPartitioner so any MatPartitioning can already be used also for DMPlex. Hence, it also shows that this class is really redundant.

It was my first contribution in this direction and then I had to switch to other topics for several months. However, I'm still interested in it although it's not a thing I could do full time. I will look at it and think about a next achievable step.

Fande, if you build on MatPartitioning, you do it right and keep going this way.

Vaclav

> 20. 11. 2018 v 18:26, Smith, Barry F. via petsc-dev <petsc-dev at mcs.anl.gov>:
> 
> 
>   MatPartitioning is the future. It is just a question of stripping out the PetscPartitioner and replacing it with the MatPartitioner.   Yes, once this was determined we lost interest in actually doing the work of stripping it out.
> 
> 
>     Barry
> 
> 
>> On Nov 20, 2018, at 11:12 AM, Fande Kong via petsc-dev <petsc-dev at mcs.anl.gov> wrote:
>> 
>> I was wondering what is the conclusion on this?
>> 
>> PetscPartitioner and MatPartitioning, which one will be kept? There is the code duplication. I am asking because I an working on a hierarchical  partitioning based MatPartitioning Interface and we also use MatPartitioning Interface in MOOSE to access all external partitioners.
>> 
>> Just want to make sure I am using the right partitioning interface.
>> 
>> Thanks,
>> 
>> Fande,
>> 
>> 
>> 
>> On Thu, Dec 7, 2017 at 8:33 AM Matthew Knepley <knepley at gmail.com> wrote:
>> On Thu, Dec 7, 2017 at 10:26 AM, Vaclav Hapla <vaclav.hapla at erdw.ethz.ch> wrote:
>> 
>> 
>>> 9. 11. 2017 v 12:53, Matthew Knepley <knepley at gmail.com>:
>>> 
>>> I think I need to create a proof-of-concept. I would start by employing MatPartitioning in PetscPartitionerPartition with anything outside of this function untouched for now (as already suggested in #192), if you agree.
>>> 
>>> Or what about implementing a special temporary PetscPartitioner implementation wrapping MatPartitioning?
>>> PETSCPARTITIONERMATPARTITIONING sounds crazy, though :) But could be a good starting point.
>>> 
>>> This is probably easier, and allows nice regression testing, namely run all tests using EXTRA_OPTIONS="-petscparitioner_type matpartitioning". I think
>>> that should be alright for correctness. There are some parallel redistribution tests in Plex.
>>> 
>>> We will need at least one performance regression. Look at how I do it here:
>>> 
>>>  https://bitbucket.org/petsc/petsc/src/312beb00c9b3e1e8ec8fac64a948a1af779da02f/src/dm/impls/plex/examples/tests/ex9.c?at=master&fileviewer=file-view-default
>>> 
>>> You can make custom events, directly access the times, and compare. You could run the two versions
>>> and check for degradation for a sequence of meshes in parallel.
>>> 
>>>  Thanks,
>>> 
>>>     Matt
>> 
>> I have made some progress in this, see https://bitbucket.org/haplav/petsc/branch/haplav/feature-petscpartitionermatpartitioning
>> There's a new test ex23 which shows basically that
>>  -petscparitioner_type parmetis
>> and
>>  -petscparitioner_type matpartitioning -mat_partitioning_type parmetis
>> give exactly the same results.
>> 
>> I have not created a PR yet since the regression testing you propose fails in some cases. When running
>>  make -f gmakefile.test test EXTRA_OPTIONS="-petscpartitioner_type matpartitioning -options_left 0" search='dm_impls_plex_tests%'
>> it seems majority of tests pass but e.g. dm_impls_plex_tests-ex7_7 fails since its reference output has been saved for -petscpartitioner_type simple. How to deal with that, please?
>> 
>> Ignore the ones with simple. I think all the SNES tests are now simple, so for ex12, ex62, and ex77 can you just do a few runs
>> and confirm that both ways give the same results?
>> 
>>  Thanks,
>> 
>>     Matt
>> 
>> As can be seen in my PetscPartitionerPartition_MatPartitioning, the needed manipulations with IS are slightly more complicated than what I inferred from the Jed's comments. But still just the existing IS methods suffice. See bitbucket.org/haplav/petsc/src/ab7d5f43fd87d1d57b51b6e7ff8de0ef3e904673/src/dm/impls/plex/petscpartmatpart.c?at=haplav%2Ffeature-petscpartitionermatpartitioning#petscpartmatpart.c-92
>> 
>> Vaclav
>> 
>> 
>> 
>> -- 
>> 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/
>