Draft Agenda for MPI RMA Working Group: March 2016
Rajeev Thakur edited this page Mar 1, 2016
·
10 revisions
Draft Agenda for MPI RMA Working Group: March 2016
(Notes from the meeting are at the bottom of this page.)
Copied over from the December 2015 meeting, with notes added during that meeting. Needs update for December.
The goal for the March meeting is to clean up all of the leftovers from the previous meetings, and move the active items to the new wiki for continued discussion.
Issues that have been migrated to the new wiki.
-
[#24](https://github.com/mpi-forum/mpi-issues/issues/24) - RMA needs new assertions for passive-target epochs. Current status (9/15): Proposed text available. Please review. Also needs input from implementations about how they might make use of the assertions. Only assertions that some implementation can exploit will be considered. -
[#23](https://github.com/mpi-forum/mpi-issues/issues/23) - extend the use of MPI_WIN_SHARED_QUERY to all windows. Text proposed, straw vote 4/2/0. Current status (9/15): Additional rationale added about why Win_create_dynamic is excluded.
Issues that refer to tickets in the old wiki. These will be considered dead at the end of the March 2016 meeting unless explicitly moved to the new wiki as an rma issue (either in the wg-rma issues on in the MPI standard issues).
-
#416 - MPI Accumulate operation restrictions. Status (3/15): No Incorrect statement about MPI-2 removed. If the RMA group is favorable to this proposal, we need the exact text, including the advice to users and implementers to ensure that the performance opportunities are not lost. Current status (9/15): Needs specific proposal text and discussion of potential performance issues. -
#434 - MPI 3.0 Errata: MPI Active Target synchronization requirements. See #456. Current status (9/15): needs to be rewritten to focus on clarity issue and consider whether the chapter should be reorganized to make this clearer. -
#436 - MPI 3.0 Errata: MPI Unified window clarification. See #456. Current status (9/15): Add a Rationale or Advice to Users on how/why this "eventually synchronized" occurs. -
#438 - Interpretation of MPI_WIN_BASE for shared memory windows. Current status (9/15): From the June meeting, new text is needed before this proposal can be further considered. -
#439 - RMA Notification. There is an alternative for this without a ticket; the alternative is linked from this ticket. A new paper added to the ticket; More discussion needed. Current status (9/15): Need a clear use case and a discussion of why send-recv is not enough for this. -
#459 - RMA sync ops with vector of windows. Current status (9/15): Some information added on performance impact. Also need to consider whether a global sync/flush_all across all created windows could be proposed instead. -
#475 - Allow MPI_Win_sync at any time. New since 3/15. Current status (9/15): Needs additional discussion. -
Nonblocking RMA epochs (see http://sc14.supercomputing.org/schedule/event_detail?evid=pap705 ) -
We still need the handle the issue raised by #456. -
There are also remaining items from the 3.1 discussion; see https://docs.google.com/document/d/1MX6NRDaq6GxslM06q_COu5uFtJjPyCqtlh09g1pc-L8/edit -
Require only one synchronization mode for a window (e.g., only passive, or only active). This would not be a backward compatible change, but it is not clear that any users need it, and it is difficult to implement correctly and efficiently.
Issues moved to new wiki
Tickets in the old wiki
- #416: Jeff will move to new wiki
- #434: Jeff will move to new wiki
- #436: Jeff will move to new wiki
- #438: Jeff will move to new wiki
- #439: Needs discussion. Jim will move to new wiki.
- #459: Jeff will either close or move to new wiki
- #475: Needs discussion and possible migration to new wiki
- Nonblocking RMA epochs paper: Pavan or Bill should create a ticket.
- Shared memory (#456): No resolution
- Old Google docs spreadsheet is done.
- Nathan will create a ticket.
- MPI_Win_sync: Lot of discussion. Jeff will create one or more tickets on Win_sync.