[Mpi-forum] September MPI Forum / Agenda and Reminder to Register

Tue Sep 2 18:35:06 CDT 2014

Hi Jim,

I have to concur with Jeff - we need to stick with the 2 week deadline
based on the procedures document we agreed on.

I would suggest to have a session on this ticket (we can fit this in on
Monday, I am sure the small tickets won’t take that long) and we can talk
about the ticket 0 changes. If people agree that they are ticket 0, we can
still schedule a vote on this next time based on that.

If you want to go that route, please let me know how much time you think
you need.

Thanks,

Martin

________________________________________________________________________
Martin Schulz, schulzm at llnl.gov, http://scalability.llnl.gov/
CASC @ Lawrence Livermore National Laboratory, Livermore, USA

On 9/2/14, 3:23 PM, "Jeff Squyres (jsquyres)" <jsquyres at cisco.com> wrote:

>Jim --
>
>I'm sorry, but your mail missed the deadline (I also see you just added a
>new PDF to the ticket about 15 minutes ago).
>
>I think you have the following options:
>
>- proceed with the last version of the ticket that was announced to the
>Forum (as of last meeting?)
>- submit mpi-report.8.pdf for the December meeting
>
>
>
>On Sep 2, 2014, at 5:27 PM, Jim Dinan <james.dinan at gmail.com> wrote:
>
>> Hi Martin,
>> 
>> Apologies for missing the 9/1 deadline.  Endpoints #380 should be up
>>for a first vote in Japan.  The working group has made several changes
>>to the ticket based on feedback from the prior reading.  We feel that
>>these changes are ticket 0, and would like to read them prior to a vote.
>> If the forum feels the changes are not ticket 0, we would like to
>>re-read at the Japan meeting.
>> 
>> We do have an unmet implementation requirement, which we will satisfy
>>prior to the second vote.
>> 
>> Cheers,
>>  ~Jim.
>> 
>> 
>> On Tue, Sep 2, 2014 at 11:44 AM, Balaji, Pavan <balaji at anl.gov> wrote:
>> Hi Martin,
>> 
>> This is not a vote, just a formal reading.  We were asked to make some
>>changes and read it again at the last meeting.
>> 
>>   — Pavan
>> 
>> On Sep 2, 2014, at 10:30 AM, Schulz Martin <schulzm at llnl.gov> wrote:
>> 
>> > Hi Pavan,
>> >
>> > I think we can still add this, despite the two week deadline, since
>>we did
>> > talk about this in the last meeting. What I don’t remember, though,
>>is the
>> > state of the ticket: it’s listed as scheduled for vote and also
>>listed as
>> > a reading in the June notes, but the ticket has two options on it and
>>it’s
>> > not clear which one is the intended one. What do you want to do with
>>it?
>> >
>> > Thanks!
>> >
>> > Martin
>> >
>> >
>> > 
>>________________________________________________________________________
>> > Martin Schulz, schulzm at llnl.gov, http://scalability.llnl.gov/
>> > CASC @ Lawrence Livermore National Laboratory, Livermore, USA
>> >
>> >
>> >
>> >
>> >
>> > On 9/2/14, 6:12 AM, "Balaji, Pavan" <balaji at anl.gov> wrote:
>> >
>> >> Hi Martin,
>> >>
>> >> You missed 411:
>> >> https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/411
>> >>
>> >> We had decided to postpone it to this meeting last time.
>> >>
>> >> Can you add it too?
>> >>
>> >> Thanks,
>> >>
>> >> — Pavan
>> >>
>> >> On Sep 2, 2014, at 3:59 AM, Schulz Martin <schulzm at llnl.gov> wrote:
>> >>
>> >>> Hi Rolf, all,
>> >>>
>> >>> I have updated the agenda with all tickets to be read and voted on
>>- I
>> >>> only got Rolf’s request below, please let me know asap in case I
>>missed
>> >>> something (plus I moved the first votes from last time up to the
>>second
>> >>> vote stage). With that, the agenda should be complete.
>> >>>
>> >>> Rolf: all the tickets are on the agenda. There are a few, though,
>>as you
>> >>> also noted, that are ticket 0 changes - I grouped them separately
>>with
>> >>> the
>> >>> intent that we don’t vote on them (but if we decide otherwise, they
>> >>> count
>> >>> as announced). I know we are a bit sparse in the chapter committees
>> >>> right
>> >>> now, but we need to fix this anyway to make MPI 3.1 happen. I am
>> >>> planning
>> >>> on pinging people separately about this in the next days - if they
>>are
>> >>> any
>> >>> volunteers right away, please let me know.
>> >>>
>> >>> Thanks!
>> >>>
>> >>> Martin
>> >>>
>> >>> 
>>________________________________________________________________________
>> >>> Martin Schulz, schulzm at llnl.gov, http://scalability.llnl.gov/
>> >>> CASC @ Lawrence Livermore National Laboratory, Livermore, USA
>> >>>
>> >>>
>> >>>
>> >>>
>> >>>
>> >>> On 8/28/14, 8:30 AM, "Rolf Rabenseifner" <rabenseifner at hlrs.de>
>>wrote:
>> >>>
>> >>>> Dear Martin,
>> >>>>
>> >>>> please add the following errata tickets to the agenda :
>> >>>>
>> >>>> #456 MPI 3.0 Errata: MPI RMA synchronization for MPI shared memory
>> >>>>
>> >>>> This is a large ticket that needs discussion
>> >>>> first in the RMA-WG, second in the plenary, and
>> >>>> hopefully some voting (corrctions and final)
>> >>>>
>> >>>> ____
>> >>>> Small Fortran corrections are:
>> >>>>
>> >>>> #433 MPI 3.0 Errata: MPI_WIN_ATTACH must only be for simply
>>contiguous
>> >>>> buffers
>> >>>> #440 Fortran cross-reference correction (ticket 0 / errata)
>> >>>> #441 Missing text for Fortran MPI_SUBARRAYS_SUPPORTED equals
>>.FALSE.
>> >>>> (errata)
>> >>>> #442 Fortran "vector subscipts" and MPI_BOTTOM (ticket 0 / errata)
>> >>>> #443 Fortran MPI_BOTTOM not for initialization or assignment
>>(ticket 0
>> >>>> /
>> >>>> errata)
>> >>>> #444 Fortran and Persistent Requests (ticket 0 / errata)
>> >>>> #445 Fortran and Permament Data Movement: "All" nonblocking
>>(ticket 0 /
>> >>>> errata)
>> >>>> #446 Fortran cross-reference update (ticket-0/errata)
>> >>>>
>> >>>> ____
>> >>>> Fortran and absolute address text corrections:
>> >>>> #431 MPI 3.0 Errata: MPI Address Errata
>> >>>>
>> >>>> ____
>> >>>> Errata and ticket-0 additions to tickets #349, #402, #404
>> >>>>     (current status of the triple 1st vote):
>> >>>> #421 MPI_Aint addressing arithmetic (ticket-0 additions)
>> >>>>
>> >>>> ____
>> >>>> Small errata corrections to Chap.2:
>> >>>>
>> >>>> #447 Chap.2 errata: Tools section is an exception: no Fortran
>>interface
>> >>>> #448 Chap.2 errata: Semantic term "matching" is missing
>> >>>> #449 Chap.2 errata: Missing "in C" in wording about
>>handle-comparison
>> >>>> #450 Chap.2 errata: Outdated cross-reference to MPI-1
>> >>>> #451 Chap.2 errata: Correcting description of nonblocking
>> >>>> #452 Chap.2 errata: Wording correction about deprecated and removed
>> >>>> functions
>> >>>> #453 Chap.2 errata: Avoidance of mpi_ / pmpi_ in C due to Fortran
>> >>>> reasons
>> >>>>
>> >>>> #455 Chap.2 errata: Removal of duplicated text on MPI_Aint and
>> >>>> MPI_Offset
>> >>>>
>> >>>>
>> >>>> _____________________________________
>> >>>> The rest of this mail are only hints.
>> >>>> ____
>> >>>> I also expect that Jeff may want to update, discuss and vote:
>> >>>> #428 Add more non-profile-able MPI routines
>> >>>>
>> >>>> ____
>> >>>> I also expect that Kathryn may want to update, discuss and vote:
>> >>>> #383 Consistency of tools information interface variable
>>information
>> >>>> #368 MPI_T Minor Errata
>> >>>>
>> >>>> ____
>> >>>> I also expect that Bill may want to update, discuss and vote:
>> >>>> #438 Interpretation of MPI_WIN_BASE for shared memory windows
>> >>>>
>> >>>> ____
>> >>>> Other MPI-3.0 errata tickets that may or should be handled before
>> >>>> MPI-3.1
>> >>>> is finalized:
>> >>>>
>> >>>> #420 Bug in neighbourhood collective example
>> >>>>   - due to missing chapter committee action,
>> >>>>     it shoud be voted in
>> >>>>
>> >>>> ____
>> >>>> The only MPI-3.0-errata ticket that may need to go MPI-4.0 is:
>> >>>> #374 Comm_free/Type_free/Win_free: clarify attribute destructor
>> >>>> semantics
>> >>>>
>> >>>>
>> >>>> ____
>> >>>> Tickets #428, #368, #368, #420 may need a milestone-change to
>>"Kobe".
>> >>>> Otherwise, they are invisible in the Kobe tickets list:
>> >>>>
>> >>>> 
>>https://svn.mpi-forum.org/trac/mpi-forum-web/query?status=accepted&statu
>> >>>> s=
>> >>>>
>> >>>> 
>>assigned&status=new&status=reopened&milestone=2014%2F09%2F15+Kobe%2C+Jap
>> >>>> an
>> >>>>
>> >>>> 
>>&col=id&col=summary&col=milestone&col=status&col=type&col=priority&col=v
>> >>>> er
>> >>>> sion&order=priority
>> >>>>
>> >>>> Best regards
>> >>>> Rolf
>> >>>>
>> >>>> ----- Original Message -----
>> >>>>> From: "Schulz Martin" <schulzm at llnl.gov>
>> >>>>> To: "Main mailing list" <mpi-forum at lists.mpi-forum.org>
>> >>>>> Sent: Thursday, August 28, 2014 9:53:01 AM
>> >>>>> Subject: [Mpi-forum] September MPI Forum / Agenda and Reminder to
>> >>>>> Register
>> >>>>>
>> >>>>> Hi all,
>> >>>>>
>> >>>>> Based on the (sparse) feedback that I have gotten, I have
>>assembled
>> >>>>> an
>> >>>>> initial agenda and uploaded it to the server. Please check the
>> >>>>> agenda,
>> >>>>> especially if you sent me a request, and double check that I
>>didn¹t
>> >>>>> miss
>> >>>>> anything or anybody. If something is missing, please let me know
>> >>>>> asap.
>> >>>>>
>> >>>>> As usual, I tried to split working groups and plenary on separate
>> >>>>> days,
>> >>>>> with votes on day three. However, so far I have only received one
>>WG
>> >>>>> meeting request, which would make Monday relatively sparse. We
>>could
>> >>>>> move
>> >>>>> parts of Tuesday to that day and probably finish on Tuesday
>>evening.
>> >>>>> Some
>> >>>>> people have voiced interest in doing this. However, this would
>>move
>> >>>>> votes
>> >>>>> onto the same day as the second half of the plenary and I would
>>like
>> >>>>> some
>> >>>>> feedback on this (since the voting times are typically fixed to
>>the
>> >>>>> last
>> >>>>> day). So the options are:
>> >>>>>
>> >>>>> Monday WG time (with open, general WG time), Tuesday plenary,
>> >>>>> Wednesday
>> >>>>> votes
>> >>>>>
>> >>>>> -or-
>> >>>>>
>> >>>>> Monday WG and plenary time (errata), Tuesday plenary and votes,
>> >>>>> Wednesday
>> >>>>> off
>> >>>>>
>> >>>>> Please let me know if you have a preference one way or the other.
>>If
>> >>>>> I
>> >>>>> don¹t hear anything or if there are objections to moving the
>>voting
>> >>>>> block,
>> >>>>> we¹ll leave things as is (first option).
>> >>>>>
>> >>>>> Thanks!
>> >>>>>
>> >>>>> Martin
>> >>>>>
>> >>>>> PS: Please don¹t forget to register, if you haven¹t done so, by
>>the
>> >>>>> end of
>> >>>>> August (just a few days left). We need an exact attendee list, so
>> >>>>> people
>> >>>>> can get access to Riken.
>> >>>>>
>> >>>>>
>> >>>>> 
>>_______________________________________________________________________
>> >>>>> _
>> >>>>> Martin Schulz, schulzm at llnl.gov, http://scalability.llnl.gov/
>> >>>>> CASC @ Lawrence Livermore National Laboratory, Livermore, USA
>> >>>>>
>> >>>>>
>> >>>>> _______________________________________________
>> >>>>> mpi-forum mailing list
>> >>>>> mpi-forum at lists.mpi-forum.org
>> >>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum
>> >>>>>
>> >>>>
>> >>>> --
>> >>>> Dr. Rolf Rabenseifner . . . . . . . . . .. email
>>rabenseifner at hlrs.de
>> >>>> High Performance Computing Center (HLRS) . phone
>>++49(0)711/685-65530
>> >>>> University of Stuttgart . . . . . . . . .. fax ++49(0)711 /
>>685-65832
>> >>>> Head of Dpmt Parallel Computing . . .
>>www.hlrs.de/people/rabenseifner
>> >>>> Nobelstr. 19, D-70550 Stuttgart, Germany . . . . (Office: Room
>>1.307)
>> >>>> _______________________________________________
>> >>>> mpi-forum mailing list
>> >>>> mpi-forum at lists.mpi-forum.org
>> >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum
>> >>>
>> >>>
>> >>> _______________________________________________
>> >>> mpi-forum mailing list
>> >>> mpi-forum at lists.mpi-forum.org
>> >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum
>> >>
>> >> --
>> >> Pavan Balaji  ✉️
>> >> http://www.mcs.anl.gov/~balaji
>> >>
>> >> _______________________________________________
>> >> mpi-forum mailing list
>> >> mpi-forum at lists.mpi-forum.org
>> >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum
>> >
>> >
>> > _______________________________________________
>> > mpi-forum mailing list
>> > mpi-forum at lists.mpi-forum.org
>> > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum
>> 
>> --
>> Pavan Balaji  ✉️
>> http://www.mcs.anl.gov/~balaji
>> 
>> _______________________________________________
>> mpi-forum mailing list
>> mpi-forum at lists.mpi-forum.org
>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum
>> 
>> _______________________________________________
>> mpi-forum mailing list
>> mpi-forum at lists.mpi-forum.org
>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum
>
>
>-- 
>Jeff Squyres
>jsquyres at cisco.com
>For corporate legal information go to:
>http://www.cisco.com/web/about/doing_business/legal/cri/
>
>_______________________________________________
>mpi-forum mailing list
>mpi-forum at lists.mpi-forum.org
>http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum