Refactor mesh parallelization

Description

The parallelization is simplified: the boundaries are now not distributed, but every rank has its own boundary points it needs locally. Moreover, the ghost point ordering is different, which means that we need to pack and unpack the values for communication the ghost points (before only packing was needed). However, this can be done with simple maps and makes the ghost exchange more versatile because the ghost points could be reordered in an arbitrary way.

To do:

• GPU support

News snippet

Refactor mesh parallelization

Checklist

• I have checked that my code follows the Octopus coding standards
• I have added tests for all the new features added in this request.