processor agglomeration does not support non-processor coupled patches

Functionality to add/problem to solve

processor agglomeration does not support non-processor coupled patches e.g. cyclicAMI

Target audience

Parallel runs that

• benefit from processor agglomeration
• use cyclicAMI/cyclicACMI

Proposal

Add general I/O for sending/receiving GAMG interfaces & fields.

What does success look like, and how can we measure that?

Use processorAgglomerator masterCoarsest on cases with cyclicAMI.