On Fri, Oct 22, 2021 at 9:13 PM Barry Smith <[email protected]> wrote:
> > One technical reason is that PetscHeaderDestroy_Private() may call > PetscCommDestroy() which may call MPI_Comm_free() which is defined by the > standard to be collective. Though PETSc tries to limit its use of new MPI > communicators (for example generally many objects shared the same > communicator) if we did not free those we no longer need when destroying > objects we could run out. > PetscCommDestroy() might call MPI_Comm_free() , but it is very unlikely. Petsc uses reference counting on communicators, so in PetscCommDestroy(), it likely just decreases the count. In other words, PetscCommDestroy() is cheap and in effect not collective. > > I cannot off-hand think of another specific technical reason they must > be collective besides this good housekeeping. > > In what use case can you not call them collectively? > > Barry > > > > On Oct 22, 2021, at 9:21 PM, Alberto F. Martín <[email protected]> > wrote: > > > > Dear PETSc users, > > > > What is the main reason underlying PetscDestroy subroutines having > global collective semantics? Is this actually true for all PETSc objects? > Can this be relaxed/deactivated by, e.g., compilation macros/configuration > options? > > > > Thanks in advance! > > > > Best regards, > > > > Alberto. > > > >
