[petsc-users] PetscSF Fortran interface

Nicholas Arnold-Medabalimi narnoldm at umich.edu
Thu Jan 12 23:10:16 CST 2023 

That is what I tried initially however, I get a segmentation fault. I can
confirm it's due to the remote offsets because if I try and output
remoteoffsets between the Distribute Section and Create Section it throws
the same segmentation fault.

Thanks for the help
Nicholas

On Thu, Jan 12, 2023 at 11:56 PM Junchao Zhang <junchao.zhang at gmail.com>
wrote:

>
>
> On Thu, Jan 12, 2023 at 1:28 PM Nicholas Arnold-Medabalimi <
> narnoldm at umich.edu> wrote:
>
>> Hi Junchao
>>
>> Going back to this merge request. I'm not sure I follow exactly the usage
>> described in the commit history. I have a c prototype of what I am trying
>> to do which is
>>
>>     PetscSectionCreate(PETSC_COMM_WORLD, &leafSection);
>>     PetscSFDistributeSection(redistributionSF, filteredSection_local, &remoteOffsets,
>> leafSection);
>>     PetscSFCreateSectionSF(redistributionSF, filteredSection_local,
>> remoteOffsets, leafSection, &redistributionSF_dof);
>>
>> But something seems unclear with the usage in fortran around the
>> remoteoffsets. Do I have to insert the CreateRemoteOffsetsF90 like so? Any
>> clarification would be greatly appreciated.
>>
>> call PetscSFDistributeSectionF90(distributionSF, section_filt_l,
>> remoteoffsets, leafSection, ierr)
>> call PetscSFCreateRemoteOffsetsf90(distributionSF, section_filt_l,
>> leafSection, remoteoffsets, ierr )
>> call PetscSFCreateSectionSFF90(distributionSF, section_filt_l,
>> remoteoffsets, leafSection, distributionSF_dof, ierr)
>>
>> Hi, Nicholas,
>   Reading through comments at
> https://gitlab.com/petsc/petsc/-/merge_requests/5386#note_1022942470, I
> feel it should look like
>
> PetscInt, pointer :: remoteOffsets(:)
> call PetscSFDistributeSectionF90(distributionSF, section_filt_l,
> remoteoffsets, leafSection, ierr)  // allocate remoteoffsets
> call PetscSFCreateSectionSFF90(distributionSF, section_filt_l,
> remoteoffsets, leafSection, distributionSF_dof, ierr)
> call PetscIntArray1dDestroyF90(remoteOffsets,ierr) // free remoteoffsets
> when not needed
>
> Could you try it?
>
>
> Sincerely
>> Nicholas
>>
>> On Tue, Jan 10, 2023 at 4:42 PM Junchao Zhang <junchao.zhang at gmail.com>
>> wrote:
>>
>>> Hi, Nicholas,
>>>    It seems we have implemented it, but with another name,
>>> PetscSFCreateSectionSFF90, see
>>> https://gitlab.com/petsc/petsc/-/merge_requests/5386
>>>    Try it to see if it works!
>>>
>>> --Junchao Zhang
>>>
>>>
>>> On Tue, Jan 10, 2023 at 11:45 AM Nicholas Arnold-Medabalimi <
>>> narnoldm at umich.edu> wrote:
>>>
>>>> Er to be honest I still can't get my stub to compile properly, and I
>>>> don't know how to go about making a merge request. But here is what I am
>>>> attempting right now. Let me know how best to proceed
>>>>
>>>>
>>>> Its not exactly clear to me how to setup up the remote offset properly.
>>>>
>>>> in src/vec/is/sf/interface/ftn-custom/zsf.c
>>>>
>>>> PETSC_EXTERN void petscsfcreatesectionsf(PetscSF *sf, PetscSection
>>>> *rootSection, F90Array1d *aremoteOffsets, PetscSection *leafSection,
>>>> PetscSF *sectionSF, int * ierr PETSC_F90_2PTR_PROTO(remoteoffsetsd))
>>>> {
>>>>
>>>>   int * remoteOffsets;
>>>>   *ierr = F90Array1dAccess(aremoteOffsets, PETSC_INT, (void**)
>>>> &remoteOffsets PETSC_F90_2PTR_PARAM(remoteoffsetsd));if (*ierr) return;
>>>>   *ierr = PetscSFCreateSectionSF(*sf,*rootSection,
>>>> &remoteOffsets,*leafSection,*sectionSF);if (*ierr) return;
>>>>
>>>> }
>>>>
>>>> This is the sticking point.
>>>>
>>>> Sincerely
>>>> Nicholas
>>>>
>>>>
>>>> On Tue, Jan 10, 2023 at 12:38 PM Junchao Zhang <junchao.zhang at gmail.com>
>>>> wrote:
>>>>
>>>>> Hi, Nicholas,
>>>>>   Could you make a merge request to PETSc and then our Fortran experts
>>>>> can comment on your MR?
>>>>>   Thanks.
>>>>>
>>>>> --Junchao Zhang
>>>>>
>>>>>
>>>>> On Tue, Jan 10, 2023 at 11:10 AM Nicholas Arnold-Medabalimi <
>>>>> narnoldm at umich.edu> wrote:
>>>>>
>>>>>> Hi Junchao
>>>>>>
>>>>>> I think I'm almost there, but I could use some insight into how to
>>>>>> use the PETSC_F90_2PTR_PROTO and F90Array1dAccess for the remoteOffset
>>>>>> parameter input so if another function comes up, I can add it myself
>>>>>> without wasting your time.
>>>>>> I am very grateful for your help and time.
>>>>>>
>>>>>> Sincerely
>>>>>> Nicholas
>>>>>>
>>>>>> On Tue, Jan 10, 2023 at 10:55 AM Junchao Zhang <
>>>>>> junchao.zhang at gmail.com> wrote:
>>>>>>
>>>>>>> Hi, Nicholas,
>>>>>>>    I am not a fortran guy, but I will try to add
>>>>>>> petscsfcreatesectionsf.
>>>>>>>
>>>>>>>    Thanks.
>>>>>>> --Junchao Zhang
>>>>>>>
>>>>>>>
>>>>>>> On Tue, Jan 10, 2023 at 12:50 AM Nicholas Arnold-Medabalimi <
>>>>>>> narnoldm at umich.edu> wrote:
>>>>>>>
>>>>>>>> I think it should be something like this, but I'm not very fluent
>>>>>>>> in Fortran C interop syntax. Any advice would be appreciated. Thanks
>>>>>>>>
>>>>>>>> PETSC_EXTERN void petscsfcreatesectionsf(PetscSF *sf, PetscSection
>>>>>>>> *rootSection, F90Array1d *aremoteOffsets, PetscSection *leafSection,
>>>>>>>> PetscSF *sectionSF, int * ierr PETSC_F90_2PTR_PROTO
>>>>>>>> (remoteoffsetsd))
>>>>>>>> {
>>>>>>>>
>>>>>>>>   int * remoteOffsets;
>>>>>>>>   *ierr = F90Array1dAccess(aremoteOffsets, PETSC_INT, (void**) &
>>>>>>>> remoteOffsets PETSC_F90_2PTR_PARAM(remoteoffsetsd));if (*ierr)
>>>>>>>> return;
>>>>>>>>   *ierr = PetscSFCreateSectionSF(*sf,*rootSection, &remoteOffsets,*
>>>>>>>> leafSection,*sectionSF);if (*ierr) return;
>>>>>>>>
>>>>>>>> }
>>>>>>>>
>>>>>>>> On Mon, Jan 9, 2023 at 11:41 PM Nicholas Arnold-Medabalimi <
>>>>>>>> narnoldm at umich.edu> wrote:
>>>>>>>>
>>>>>>>>> Hi Junchao
>>>>>>>>>
>>>>>>>>> Thanks again for your help in November. I've been using the your
>>>>>>>>> merge request branch quite heavily. Would it be possible to add a
>>>>>>>>> petscsfcreatesectionsf interface as well?
>>>>>>>>> I'm trying to write it myself using your commits as a guide but I
>>>>>>>>> have been struggling with handling the section parameter properly.
>>>>>>>>>
>>>>>>>>> Sincerely
>>>>>>>>> Nicholas
>>>>>>>>>
>>>>>>>>> On Sat, Nov 19, 2022 at 9:44 PM Junchao Zhang <
>>>>>>>>> junchao.zhang at gmail.com> wrote:
>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> On Sat, Nov 19, 2022 at 8:05 PM Nicholas Arnold-Medabalimi <
>>>>>>>>>> narnoldm at umich.edu> wrote:
>>>>>>>>>>
>>>>>>>>>>> Hi
>>>>>>>>>>>
>>>>>>>>>>> Thanks, this is awesome. Thanks for the very prompt fix. Just
>>>>>>>>>>> one question: will the array outputs on the fortran side copies (and need
>>>>>>>>>>> to be deallocated) or direct access to the dmplex?
>>>>>>>>>>>
>>>>>>>>>> Direct access to internal data;  no need to deallocate
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Sincerely
>>>>>>>>>>> Nicholas
>>>>>>>>>>>
>>>>>>>>>>> On Sat, Nov 19, 2022 at 8:21 PM Junchao Zhang <
>>>>>>>>>>> junchao.zhang at gmail.com> wrote:
>>>>>>>>>>>
>>>>>>>>>>>> Hi, Nicholas,
>>>>>>>>>>>>   See this MR,
>>>>>>>>>>>> https://gitlab.com/petsc/petsc/-/merge_requests/5860
>>>>>>>>>>>>   It is in testing, but you can try branch
>>>>>>>>>>>> jczhang/add-petscsf-fortran to see if it works for you.
>>>>>>>>>>>>
>>>>>>>>>>>>   Thanks.
>>>>>>>>>>>> --Junchao Zhang
>>>>>>>>>>>>
>>>>>>>>>>>> On Sat, Nov 19, 2022 at 4:16 PM Nicholas Arnold-Medabalimi <
>>>>>>>>>>>> narnoldm at umich.edu> wrote:
>>>>>>>>>>>>
>>>>>>>>>>>>> Hi Junchao
>>>>>>>>>>>>>
>>>>>>>>>>>>> Thanks. I was wondering if there is any update on this. I may
>>>>>>>>>>>>> write a small interface for those two routines myself in the interim but
>>>>>>>>>>>>> I'd appreciate any insight you have.
>>>>>>>>>>>>>
>>>>>>>>>>>>> Sincerely
>>>>>>>>>>>>> Nicholas
>>>>>>>>>>>>>
>>>>>>>>>>>>> On Wed, Nov 16, 2022 at 10:39 PM Junchao Zhang <
>>>>>>>>>>>>> junchao.zhang at gmail.com> wrote:
>>>>>>>>>>>>>
>>>>>>>>>>>>>> Hi, Nicholas,
>>>>>>>>>>>>>>   I will have a look and get back to you.
>>>>>>>>>>>>>>   Thanks.
>>>>>>>>>>>>>> --Junchao Zhang
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> On Wed, Nov 16, 2022 at 9:27 PM Nicholas Arnold-Medabalimi <
>>>>>>>>>>>>>> narnoldm at umich.edu> wrote:
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> Hi Petsc Users
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> I'm in the process of adding some Petsc for mesh management
>>>>>>>>>>>>>>> into an existing Fortran Solver. It has been relatively straightforward so
>>>>>>>>>>>>>>> far but I am running into an issue with using PetscSF routines. Some like
>>>>>>>>>>>>>>> the PetscSFGetGraph work no problem but a few of my routines require the
>>>>>>>>>>>>>>> use of PetscSFGetLeafRanks and PetscSFGetRootRanks and those don't seem to
>>>>>>>>>>>>>>> be in the fortran interface and I just get a linking error. I also don't
>>>>>>>>>>>>>>> seem to see a PetscSF file in the finclude. Any clarification or assistance
>>>>>>>>>>>>>>> would be appreciated.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> Sincerely
>>>>>>>>>>>>>>> Nicholas
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> --
>>>>>>>>>>>>>>> Nicholas Arnold-Medabalimi
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> Ph.D. Candidate
>>>>>>>>>>>>>>> Computational Aeroscience Lab
>>>>>>>>>>>>>>> University of Michigan
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> --
>>>>>>>>>>>>> Nicholas Arnold-Medabalimi
>>>>>>>>>>>>>
>>>>>>>>>>>>> Ph.D. Candidate
>>>>>>>>>>>>> Computational Aeroscience Lab
>>>>>>>>>>>>> University of Michigan
>>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> --
>>>>>>>>>>> Nicholas Arnold-Medabalimi
>>>>>>>>>>>
>>>>>>>>>>> Ph.D. Candidate
>>>>>>>>>>> Computational Aeroscience Lab
>>>>>>>>>>> University of Michigan
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>> --
>>>>>>>>> Nicholas Arnold-Medabalimi
>>>>>>>>>
>>>>>>>>> Ph.D. Candidate
>>>>>>>>> Computational Aeroscience Lab
>>>>>>>>> University of Michigan
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>> Nicholas Arnold-Medabalimi
>>>>>>>>
>>>>>>>> Ph.D. Candidate
>>>>>>>> Computational Aeroscience Lab
>>>>>>>> University of Michigan
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>> --
>>>>>> Nicholas Arnold-Medabalimi
>>>>>>
>>>>>> Ph.D. Candidate
>>>>>> Computational Aeroscience Lab
>>>>>> University of Michigan
>>>>>>
>>>>>
>>>>
>>>> --
>>>> Nicholas Arnold-Medabalimi
>>>>
>>>> Ph.D. Candidate
>>>> Computational Aeroscience Lab
>>>> University of Michigan
>>>>
>>>
>>
>> --
>> Nicholas Arnold-Medabalimi
>>
>> Ph.D. Candidate
>> Computational Aeroscience Lab
>> University of Michigan
>>
>

-- 
Nicholas Arnold-Medabalimi

Ph.D. Candidate
Computational Aeroscience Lab
University of Michigan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20230113/f260cc73/attachment-0001.html>

More information about the petsc-users mailing list