On Nov 2, 2022, at 6:52 AM, Stephan Köhler <stephan.koehler@math.tu-freiberg.de> wrote:

Dear PETSc/Tao team,

it seems to be that there is a bug in the LMVM matrix class:

In the function MatCreateVecs_LMVM, see, e.g., https://petsc.org/release/src/ksp/ksp/utils/lmvm/lmvmimpl.c.html at line 214.
it is not checked if the vectors  *L, or *R are NULL.  This is, in particular, a problem if this matrix class is combined with the Schur complement matrix class, since MatMult_SchurComplement
calls this function with NULL as *R, see, e.g. https://petsc.org/release/src/ksp/ksp/utils/schurm/schurm.c.html at line 66.

I attach a minimal example.  You need to modify the paths to the PETSc installation in the makefile.

Best regards
Stephan Köhler

--
Stephan Köhler
TU Bergakademie Freiberg
Institut für numerische Mathematik und Optimierung

Akademiestraße 6
09599 Freiberg
Gebäudeteil Mittelbau, Zimmer 2.07

Telefon: +49 (0)3731 39-3173 (Büro)

<Minimal_example_schur_lmvm.tar.gz><OpenPGP_0xC9BF2C20DFE9F713.asc>