In reading the error message I see that I did not clone A, to get P, so P was the wrong type with a device.
Thanks, Mark On Tue, Oct 31, 2023 at 2:24 PM Mark Adams <[email protected]> wrote: > Correction, I get the same message with -mat_type aijcusparse. > > Thanks, > Mark > > On Tue, Oct 31, 2023 at 9:29 AM Mark Adams <[email protected]> wrote: > >> I am getting this error. >> This is in GAMG/HEM setup. PtAP for the coarse grid construction works, >> but I call this in a graph routine >> (/global/u2/m/madams/petsc/src/mat/coarsen/impls/hem/hem.c:1043). >> >> Also, this PtAP does not need to be on the GPU anyway because P is >> extremely sparse ... can I pin, say P, to the CPU to keep this all on the >> host? >> >> Thanks, >> Mark >> >> >> [0]PETSC ERROR: --------------------- Error Message >> -------------------------------------------------------------- >> [0]PETSC ERROR: Petsc has generated inconsistent data >> [0]PETSC ERROR: Unspecified symbolic phase for product AB with A >> mpiaijkokkos, B mpiaij. Call MatProductSetFromOptions() first >> [0]PETSC ERROR: WARNING! There are unused option(s) set! Could be the >> program crashed before usage or a spelling mistake, etc! >> [0]PETSC ERROR: Option left: name:-ksp_converged_reason (no value) >> source: command line >> [0]PETSC ERROR: Option left: name:-ksp_viewxx (no value) source: >> command line >> [0]PETSC ERROR: Option left: name:-log_view_gpu_timexxx (no value) >> source: command line >> [0]PETSC ERROR: Option left: name:-options_left (no value) source: >> command line >> [0]PETSC ERROR: Option left: name:-pc_gamg_use_aggressive_square_graph >> value: true source: command line >> [0]PETSC ERROR: Option left: name:-pc_gamg_use_minimum_degree_ordering >> value: false source: command line >> [0]PETSC ERROR: See https://petsc.org/release/faq/ for trouble shooting. >> [0]PETSC ERROR: Petsc Development GIT revision: v3.20.0-168-ga7898f52c39 >> GIT Date: 2023-10-28 10:07:38 -0500 >> [0]PETSC ERROR: /global/u2/m/madams/petsc/src/snes/tests/./ex13 on a >> arch-perlmutter-dbg-gcc-kokkos-cuda named nid001680 by madams Tue Oct 31 >> 06:21:25 2023 >> [0]PETSC ERROR: Configure options --CFLAGS=" -g" --CXXFLAGS=" -g" >> --CUDAFLAGS="-g -Xcompiler -rdynamic" --with-cc=cc --with-cxx=CC >> --with-fc=ftn --LDFLAGS=-lmpifort_gnu_91 --FFLAGS=" -g " --COPTFLAGS=" >> -O0" --CXXOPTFLAGS=" -O0" --FOPTFLAGS=" -O0" --download-triangle=1 >> --with-debugging=1 --with-cuda=1 --with-cuda-arch=80 --with-mpiexec="srun >> -G4" --with-batch=0 --download-kokkos --download-kokkos-kernels >> --with-kokkos-kernels-tpl=0 --with-make-np=8 >> PETSC_ARCH=arch-perlmutter-dbg-gcc-kokkos-cuda >> [0]PETSC ERROR: #1 MatProductSymbolic() at >> /global/u2/m/madams/petsc/src/mat/interface/matproduct.c:807 >> [0]PETSC ERROR: #2 MatProductSymbolic_PtAP_Unsafe() at >> /global/u2/m/madams/petsc/src/mat/interface/matproduct.c:73 >> [0]PETSC ERROR: #3 MatProductSymbolic_Unsafe() at >> /global/u2/m/madams/petsc/src/mat/interface/matproduct.c:185 >> [0]PETSC ERROR: #4 MatProductSymbolic() at >> /global/u2/m/madams/petsc/src/mat/interface/matproduct.c:795 >> [0]PETSC ERROR: #5 MatPtAP() at >> /global/u2/m/madams/petsc/src/mat/interface/matrix.c:9938 >> [0]PETSC ERROR: #6 MatCoarsenApply_HEM_private() at >> /global/u2/m/madams/petsc/src/mat/coarsen/impls/hem/hem.c:1043 >> >
