[petsc-users] question about PC_BJACOBI

[Leo van Kampenhout]

Hi all,

With p number of processors in the communicator, the block preconditioner
PC_BJACOBI will by default use p blocks. So far, so good. However, in order
to compare this algorithmic efficiency decrease (since the bigger p, the
less efficient the preconditioner), i ran the commands

mpirun -n 1 ./program -pc_bjacobi_blocks 8
mpirun -n 8 ./program -pc_bjacobi_blocks 8

I expected the preconditioning to be equally efficient in this case.
However, GMRES makes more iterations in the first case (30 against 28) which
I cannot explain. Are there more subtle differences about the preconditioner
or the KSP that i'm overlooking here?

regards,

Leo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20100923/fb73df3a/attachment.htm>