mpi-forum / mpi-issues

Tickets for the MPI Forum
http://www.mpi-forum.org/
67 stars 8 forks source link

MPI_COMM_TYPE_HW_UNGUIDED description #326

Open VictorEijkhout opened 4 years ago

VictorEijkhout commented 4 years ago

(my first ticket. let me know if I should have done this differently)

MPI_COMM_TYPE_HW_UNGUIDED — the group of MPI processes associated with newcomm must be a strict subset of the group associated with comm and each
newcomm corresponds to a single instance of a hardware resource type (for example a computing core or an L3 cache).
  1. This is not a standalone description. Shouldn't this start with a sentence like " this value specifies that the communicator \mpiarg{comm} is split according to a" for the guided cases?
  2. Also note the strange line break after "each" which I don't understand, looking at the TeX source.

Proposal: include same sentence as for MPI_COMM_TYPE_HW_GUIDED. Fix line break.

Impact on users: improved intelligibility.

GuillaumeMercier commented 2 years ago

As for your first point: the description of MPI_COMM_TYPE_SHARED does not follow this pattern either. Instead of starting each description with "this value specifies ...", I would rather remove it from the description of MPI_COMM_TYPE_HW_GUIDED. Something like:

MPI_COMM_TYPE_HW_GUIDED---the communicator `comm` is split according to a
  hardware resource type (for example a computing core or an L3 cache) specified by 
  the `mpi_hw_resource_type` info key.

As for your second point: the line break is not in the document anymore it seems.

wesbland commented 1 year ago

This wasn't read at the December 2022 meeting. The last opportunity for MPI 4.1 is to have it ready at the March 2022 meeting (and it needs to "pass" the reading).

@GuillaumeMercier Are you (or is someone else) planning to push this forward or should we move it out of the plan for MPI 4.1?

GuillaumeMercier commented 1 year ago

I will push it forward. It shouldn't be an issue, actually.