[Mpi3-rma] FW: MPI 3 RMA Examples needed

[Jeff Hammond]

> However, if your problem is static (enough), you could establish
> channels upfront (e.g., sparse matrix vector in Krylov subspace
> methods).

I am not a PDE person, but my intuition is that those sparse matvec
solvers often run in conjunction with AMR or FEM where static,
one-time initialization is impossible.  However, I'll trust the PETSc
team and similar-minded folks to figure out those issues when they
come.

Perhaps the LLNL folks know something about the multi-physics codes
which solve continuum equations using FEM and can confirm or deny the
possible utility of RMI in those codes at exascale.

Jeff

-- 
Jeff Hammond
Argonne Leadership Computing Facility
jhammond at mcs.anl.gov / (630) 252-5381
http://www.linkedin.com/in/jeffhammond