Currently, external linear algebra support is limited to sequential cases. The idea is therefore, to expand it towards distributed-memory (MPI) solvers. Trilinos is good option, since it contains several packages as well as provides interfaces to third-party solvers.
In future, the sequential third-party solvers integrated in falcon may be deprecated.
Caveat(s):
- Jive (and therefore, falcon) does not have any global dof numbering.
- Unless Trilinos is added to Jive backend, the implementation would involve duplication of data (e.g., same matrix, vector as Jive and Trilinos types)
Assumption(s): Some basic assumptions that will be relaxed in later developments
- Node numbering starts from 1
- Nodes numbering is contiguous
- All nodes have same number of dofs
- Non-overlapping domain decomposition
Here is the workflow:
Currently, external linear algebra support is limited to sequential cases. The idea is therefore, to expand it towards distributed-memory (MPI) solvers. Trilinos is good option, since it contains several packages as well as provides interfaces to third-party solvers.
In future, the sequential third-party solvers integrated in falcon may be deprecated.
Caveat(s):
Assumption(s): Some basic assumptions that will be relaxed in later developments
Here is the workflow: