Class RunContext

Take the Mesh instance `mesh' and distribute it according to `partition'.

If partition is an integer, invoke PyMetis to partition the mesh into this many parts, distributing over the first `partition' ranks.

If partition is None, act as if partition was the integer corresponding to the current number of ranks on the job.

If partition is not an integer, it must be a mapping from element number to rank. (A list or tuple of rank numbers will do, for example, or so will a full-blown dict.)

Returns a mesh chunk.

We deliberately do not define the term `mesh chunk'. The return value of this function is to be treated as opaque by the user, only to be used as an argument to make_discretization().

This routine may only be invoked on the head rank.

Wait for a mesh chunk to be sent by the head rank.

We deliberately do not define the term `mesh chunk'. The return value of this function is to be treated as opaque by the user, only to be used as an argument to make_discretization().

This routine should only be invoked on non-head ranks.

Construct a Discretization instance.

`mesh_data' is whatever gets returned from distribute_mesh and receive_mesh(). Any extra arguments are directly forwarded to the respective Discretization constructor.

rank

Get Method:

unreachable.rank(self)

ranks

Get Method:

unreachable.ranks(self)

head_rank

Get Method:

unreachable.head_rank(self)

is_head_rank

Get Method:

unreachable.is_head_rank(self)