Use the (restricted) additive Schwarz method, each block is (approximately) solved with its own
Options Database Keys#
-Sets total blocks. Defaults to one block per MPI rank.
If you run with, for example, 3 blocks on 1 processor or 3 blocks on 3 processors you will get a different convergence rate due to the default option of -pc_asm_type restrict. Use -pc_asm_type basic to get the same convergence behavior
Each processor can have one or more blocks, but a block cannot be shared by more
than one processor. Use
PCGASM for subdomains shared by multiple processes.
**** -*** M Dryja, OB Widlund, An additive variant of the Schwarz alternating method for the case of many subregions Courant Institute, New York University Technical report
**** -*** Barry Smith, Petter Bjorstad, and William Gropp, Domain Decompositions: Parallel Multilevel Methods for Elliptic Partial Differential Equations, Cambridge University Press.