[petsc-users] PCMG with shell restrictions

Tue Dec 10 07:37:13 CST 2024

On Tue, Dec 10, 2024 at 7:03 AM Matteo Semplice via petsc-users <
petsc-users at mcs.anl.gov> wrote:

> Dear petsc-users,
>
>     I am trying with a student to modify the MG example 65 to use
> mat-shells instead of assembled matrices. Our use case is for a method that
> will use custom shell operators and shell interpolation/restrictions.
>
> To start we have modified ex65 and tried to replace the standard
> restriction/interpolation with a shell matrix that performs the same
> operations on the DMDA grids.
>
> I attach our modifications to ex65.
>
> I will find time to look at the code. Right now, you might look at this
tutorial I setup for GMG with a user-defined operator.

  https://urldefense.us/v3/__https://gitlab.com/petsc/petsc/-/commits/knepley/tutorial-gmg__;!!G_uCfscf7eWS!adh8Ragc8P_2BSKs-gf4XfTABoIU_2W2Fn447_lRBIjR7zJMd9UsbMXr3idQcRxB_wNg8MS-WuJHGa-DiVe-$ 

  Thanks!

     Matt

> The problem is that the code sometimes completes execution and sometimes
> errors out like
>
> $ ./ex65shell -ksp_monitor -pc_type mg -da_refine 2 -ksp_rtol 1e-1
> >> Created DMshell 0x55bbc83c91a0 (0x55bbc8390a80)
> Calling KSPSolve from main
> computeRHS on grid 513
> computeMatrix on grid 513
> Inside Coarsen
> >> Created DMshell 0x55bbc84c5270 (0x55bbc84b2ff0)
> Inside Coarsen
> >> Created DMshell 0x55bbc84e0a30 (0x55bbc84bdb60)
> >> Create interpolation from 0x55bbc84c5270(0x55bbc84b2ff0) to
> 0x55bbc83c91a0(0x55bbc8390a80)
> [0]PETSC ERROR: --------------------- Error Message
> --------------------------------------------------------------
> [0]PETSC ERROR: Invalid pointer
> [0]PETSC ERROR: Invalid Pointer to PetscObject: Argument 'obj' (parameter
> # 1)
> [0]PETSC ERROR: See https://urldefense.us/v3/__https://petsc.org/release/faq/__;!!G_uCfscf7eWS!adh8Ragc8P_2BSKs-gf4XfTABoIU_2W2Fn447_lRBIjR7zJMd9UsbMXr3idQcRxB_wNg8MS-WuJHGbc9Nz8M$ 
> <https://urldefense.us/v3/__https://petsc.org/release/faq/__;!!G_uCfscf7eWS!Y29qRoOD71oxcd-BQFTUohkIToMJNW3puZ3CerpZ6pBnecqhBQ8TBwhSnikLt4BYQTmsBgtKvoCFi-JuLXQAE_Oe_a_d2XOL8KvOmw$>
> for trouble shooting.
> [0]PETSC ERROR: Petsc Release Version 3.22.0, unknown
> [0]PETSC ERROR: ./ex65shell with 1 MPI process(es) and PETSC_ARCH  on
> signalkuppe by matteo Tue Dec 10 12:55:12 2024
> [0]PETSC ERROR: Configure options:
> --prefix=/home/matteo/software/petscsaved/3.22-opt/
> PETSC_DIR=/home/matteo/software/petsc --PETSC_ARCH=dbg --with-debugging=1
> --with-st
> rict-petscerrorcode --download-hdf5 --download-ml --with-metis
> --with-parmetis --with-gmsh --with-triangle --with-zlib
> --with-p4est-dir=~/software/p4est/local/
> [0]PETSC ERROR: #1 PetscObjectReference() at
> /home/matteo/software/petsc/src/sys/objects/inherit.c:620
> [0]PETSC ERROR: #2 PCMGSetRScale() at
> /home/matteo/software/petsc/src/ksp/pc/impls/mg/mgfunc.c:394
> [0]PETSC ERROR: #3 PCSetUp_MG() at
> /home/matteo/software/petsc/src/ksp/pc/impls/mg/mg.c:998
> [0]PETSC ERROR: #4 PCSetUp() at
> /home/matteo/software/petsc/src/ksp/pc/interface/precon.c:1071
> [0]PETSC ERROR: #5 KSPSetUp() at
> /home/matteo/software/petsc/src/ksp/ksp/interface/itfunc.c:415
> [0]PETSC ERROR: #6 KSPSolve_Private() at
> /home/matteo/software/petsc/src/ksp/ksp/interface/itfunc.c:826
> [0]PETSC ERROR: #7 KSPSolve() at
> /home/matteo/software/petsc/src/ksp/ksp/interface/itfunc.c:1075
> [0]PETSC ERROR: #8 main() at ../src/ex65shell.c:79
> [0]PETSC ERROR: PETSc Option Table entries:
> [0]PETSC ERROR: -da_refine 2 (source: command line)
> [0]PETSC ERROR: -ksp_monitor (source: command line)
> [0]PETSC ERROR: -ksp_rtol 1e-1 (source: command line)
> [0]PETSC ERROR: -pc_type mg (source: command line)
> [0]PETSC ERROR: ----------------End of Error Message -------send entire
> error message to petsc-maint at mcs.anl.gov----------
>
> In all cases valgrind complains like
>
> $ valgrind ./ex65shell -ksp_monitor -pc_type mg -da_refine 2 -ksp_rtol
> 1e-1 -int_view ascii::ascii_info
> ==2130767== Memcheck, a memory error detector
> ==2130767== Copyright (C) 2002-2022, and GNU GPL'd, by Julian Seward et
> al.
> ==2130767== Using Valgrind-3.19.0 and LibVEX; rerun with -h for copyright
> info
> ==2130767== Command: ./ex65shell -ksp_monitor -pc_type mg -da_refine 2
> -ksp_rtol 1e-1 -int_view ascii::ascii_info
> ==2130767==
> hwloc x86 backend cannot work under Valgrind, disabling.
> May be reenabled by dumping CPUIDs with hwloc-gather-cpuid
> and reloading them under Valgrind with HWLOC_CPUID_PATH.
> >> Created DMshell 0xfbf50f0 (0xfb49480)
> Calling KSPSolve from main
> computeRHS on grid 513
> computeMatrix on grid 513
> Inside Coarsen
> >> Created DMshell 0xff404a0 (0xff062d0)
> Inside Coarsen
> >> Created DMshell 0xff78ca0 (0xff4a530)
> >> Create interpolation from 0xff404a0(0xff062d0) to 0xfbf50f0(0xfb49480)
> Mat Object: 1 MPI process
>  type: shell
>  rows=513, cols=257
> ==2130767== Conditional jump or move depends on uninitialised value(s)
> ==2130767==    at 0x83FDAAE: PCSetUp_MG (mg.c:998)
> ==2130767==    by 0x8620C04: PCSetUp (precon.c:1071)
> ==2130767==    by 0x7DA78D5: KSPSetUp (itfunc.c:415)
> ==2130767==    by 0x7DB2093: KSPSolve_Private (itfunc.c:826)
> ==2130767==    by 0x7DB7A53: KSPSolve (itfunc.c:1075)
> ==2130767==    by 0x10D2F7: main (ex65shell.c:79)
> ==2130767==
> ==2130767== Conditional jump or move depends on uninitialised value(s)
> ==2130767==    at 0x5794061: VecDestroy (vector.c:570)
> ==2130767==    by 0x83FDC36: PCSetUp_MG (mg.c:999)
> ==2130767==    by 0x8620C04: PCSetUp (precon.c:1071)
> ==2130767==    by 0x7DA78D5: KSPSetUp (itfunc.c:415)
> ==2130767==    by 0x7DB2093: KSPSolve_Private (itfunc.c:826)
> ==2130767==    by 0x7DB7A53: KSPSolve (itfunc.c:1075)
> ==2130767==    by 0x10D2F7: main (ex65shell.c:79)
> ==2130767==
>
> We are clearly doing something wrong since PCSetUp_MG (mg.c:998)  is an
> area of code where I wouldn't expect we would enter.
>
> Can you advise on the proper way to achieve our goal?
>
> Best regards
>
>     Matteo
>

-- 
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://urldefense.us/v3/__https://www.cse.buffalo.edu/*knepley/__;fg!!G_uCfscf7eWS!adh8Ragc8P_2BSKs-gf4XfTABoIU_2W2Fn447_lRBIjR7zJMd9UsbMXr3idQcRxB_wNg8MS-WuJHGdoVXWI2$  <https://urldefense.us/v3/__http://www.cse.buffalo.edu/*knepley/__;fg!!G_uCfscf7eWS!adh8Ragc8P_2BSKs-gf4XfTABoIU_2W2Fn447_lRBIjR7zJMd9UsbMXr3idQcRxB_wNg8MS-WuJHGdRGFKZR$ >
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20241210/7efa83d8/attachment.html>