Barry: Thanks for the quick response!
I assumed that the Mat object created by MatCreateSchurComplement can somehow be solved by a KSP option (like plain CG). I am aware of the fieldsplit pc. My situation is that I have to solve a problem with S = L - D K^-1 D^T, wherein L, D, K are all sparse. Of course, the last option for me is to write a simple CG solver for S myself. Do you have any recommendations? Thanks, Luis 2018-01-23 11:25 GMT-08:00 Smith, Barry F. <[email protected]>: > > Luis, > > This doesn't really have any thing to do with the Schur complement per > se, you are asking do an ILU on this abstract Schur complement matrix which > is not possible. > > Almost for sure you want to use PCFIELDSPLIT and not directly create > the MatCreateSchurComplement() yourself. PCFIELDSPLIT has many possible > variants for solving systems with segregated Schur complement methods. > > Barry > > > > On Jan 23, 2018, at 1:13 PM, saturday luis <[email protected]> > wrote: > > > > Hi PETSc team: > > > > I am trying to solve a Stokes problem in a segregated manner, in which > the pressure field is obtained by solving the pressure Schur complement. > This is what I did. > > > > I created a Schur complement by calling > > > > Mat S; > > MatCreateSchurComplement(gloAssem_ptr->K_11, gloAssem_ptr->K_11, > gloAssem_ptr->K_10, gloAssem_ptr->K_01, gloAssem_ptr->K_00, &S); > > > > Then I use the Mat S as an input for KSP: > > KSPSetOperators(ksp, S, S); > > > > I got error message as follows. > > [0]PETSC ERROR: See http://www.mcs.anl.gov/petsc/documentation/ > linearsolvertable.html for possible LU and Cholesky solvers > > [0]PETSC ERROR: Could not locate a solver package. Perhaps you must > ./configure with --download-<package> > > [0]PETSC ERROR: See http://www.mcs.anl.gov/petsc/documentation/faq.html > for trouble shooting. > > [0]PETSC ERROR: Petsc Release Version 3.8.3, Dec, 09, 2017 > > [0]PETSC ERROR: ./bipn3d on a arch-linux2-cxx-debug named bacon by luis > Tue Jan 23 11:02:34 2018 > > [0]PETSC ERROR: Configure options --prefix=~/lib/petsc-3.8.3-debug > --with-x=0 -with-pic --with-mpi-compilers=1 > --with-mpi-dir=/home/lib/mpich-3.2/ > --with-clanguage=C++ --with-scalar-type=real --with-shared-libraries=1 > --with-precision=double --with-hypre=1 --download-hypre --with-spai=1 > --download-spai --with-superlu_dist=1 --download-superlu_dist > --with-superlu=1 --download-superlu --with-parmetis=1 --download-parmetis > --with-metis=1 --download-metis --with-mumps=1 --download-mumps > --with-parmetis=1 --download-parmetis --with-metis=1 --download-metis > --with-scalapack=1 --download-scalapack --with-blacs=1 --download-blacs > --with-spooles=1 --download-spooles --with-debugging=yes > --download-fblaslapack > > [0]PETSC ERROR: #1 MatGetFactor() line 4346 in > /home/lib/petsc-3.8.3/src/mat/interface/matrix.c > > [0]PETSC ERROR: #2 PCSetUp_ILU() line 142 in > /home/lib/petsc-3.8.3/src/ksp/pc/impls/factor/ilu/ilu.c > > [0]PETSC ERROR: #3 PCSetUp() line 924 in /home/lib/petsc-3.8.3/src/ksp/ > pc/interface/precon.c > > [0]PETSC ERROR: #4 KSPSetUp() line 381 in /home/lib/petsc-3.8.3/src/ksp/ > ksp/interface/itfunc.c > > [0]PETSC ERROR: #5 KSPSolve() line 612 in /home/lib/petsc-3.8.3/src/ksp/ > ksp/interface/itfunc.c > > > > Let me know if you can see any issue in the usage of Schur complement > creation. > > > > Thanks, > > > > Luis > >
