September 2014 - mpi-forum - mailman3.cels.anl.gov

URGENT: Register for MPI Forum
by Schulz Martin 02 Sep '14

02 Sep '14

Hi all, If you haven¹t done so, please register for the MPI forum as soon as possible, i.e., TODAY! We need a complete list of attendees ahead of time to enable everyone to get access to Riken. Also, if the registration is as it is right now, we are under quorum (!), but I didn¹t see a few institutions listed that I would expect to be present, so there is hope. It would be good to know for sure, though! Thanks! Martin ________________________________________________________________________ Martin Schulz, schulzm(a)llnl.gov, http://scalability.llnl.gov/ CASC @ Lawrence Livermore National Laboratory, Livermore, USA

1 0

Re: [Mpi-forum] September MPI Forum / Agenda and Reminder to Register
by Jim Dinan 02 Sep '14

02 Sep '14

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(a)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(a)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(a)llnl.gov, http://scalability.llnl.gov/ > > CASC @ Lawrence Livermore National Laboratory, Livermore, USA > > > > > > > > > > > > On 9/2/14, 6:12 AM, "Balaji, Pavan" <balaji(a)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(a)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(a)llnl.gov, http://scalability.llnl.gov/ > >>> CASC @ Lawrence Livermore National Laboratory, Livermore, USA > >>> > >>> > >>> > >>> > >>> > >>> On 8/28/14, 8:30 AM, "Rolf Rabenseifner" <rabenseifner(a)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(a)llnl.gov> > >>>>> To: "Main mailing list" <mpi-forum(a)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(a)llnl.gov, http://scalability.llnl.gov/ > >>>>> CASC @ Lawrence Livermore National Laboratory, Livermore, USA > >>>>> > >>>>> > >>>>> _______________________________________________ > >>>>> mpi-forum mailing list > >>>>> mpi-forum(a)lists.mpi-forum.org > >>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > >>>>> > >>>> > >>>> -- > >>>> Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)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(a)lists.mpi-forum.org > >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > >>> > >>> > >>> _______________________________________________ > >>> mpi-forum mailing list > >>> mpi-forum(a)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(a)lists.mpi-forum.org > >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > > > > > _______________________________________________ > > mpi-forum mailing list > > mpi-forum(a)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(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >

1 0

Re: [Mpi-forum] September MPI Forum / Agenda and Reminder to Register
by Balaji, Pavan 02 Sep '14

02 Sep '14

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(a)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(a)llnl.gov, http://scalability.llnl.gov/ > CASC @ Lawrence Livermore National Laboratory, Livermore, USA > > > > > > On 9/2/14, 6:12 AM, "Balaji, Pavan" <balaji(a)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(a)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(a)llnl.gov, http://scalability.llnl.gov/ >>> CASC @ Lawrence Livermore National Laboratory, Livermore, USA >>> >>> >>> >>> >>> >>> On 8/28/14, 8:30 AM, "Rolf Rabenseifner" <rabenseifner(a)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(a)llnl.gov> >>>>> To: "Main mailing list" <mpi-forum(a)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(a)llnl.gov, http://scalability.llnl.gov/ >>>>> CASC @ Lawrence Livermore National Laboratory, Livermore, USA >>>>> >>>>> >>>>> _______________________________________________ >>>>> mpi-forum mailing list >>>>> mpi-forum(a)lists.mpi-forum.org >>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>>>> >>>> >>>> -- >>>> Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)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(a)lists.mpi-forum.org >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>> >>> >>> _______________________________________________ >>> mpi-forum mailing list >>> mpi-forum(a)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(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum -- Pavan Balaji ✉️ http://www.mcs.anl.gov/~balaji

1 0

Re: [Mpi-forum] September MPI Forum / Agenda and Reminder to Register
by Schulz Martin 02 Sep '14

02 Sep '14

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(a)llnl.gov, http://scalability.llnl.gov/ CASC @ Lawrence Livermore National Laboratory, Livermore, USA On 9/2/14, 6:12 AM, "Balaji, Pavan" <balaji(a)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(a)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(a)llnl.gov, http://scalability.llnl.gov/ >> CASC @ Lawrence Livermore National Laboratory, Livermore, USA >> >> >> >> >> >> On 8/28/14, 8:30 AM, "Rolf Rabenseifner" <rabenseifner(a)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(a)llnl.gov> >>>> To: "Main mailing list" <mpi-forum(a)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(a)llnl.gov, http://scalability.llnl.gov/ >>>> CASC @ Lawrence Livermore National Laboratory, Livermore, USA >>>> >>>> >>>> _______________________________________________ >>>> mpi-forum mailing list >>>> mpi-forum(a)lists.mpi-forum.org >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>>> >>> >>> -- >>> Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)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(a)lists.mpi-forum.org >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >> >> >> _______________________________________________ >> mpi-forum mailing list >> mpi-forum(a)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(a)lists.mpi-forum.org >http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum

1 0

Re: [Mpi-forum] September MPI Forum / Agenda and Reminder to Register
by Balaji, Pavan 02 Sep '14

02 Sep '14

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(a)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(a)llnl.gov, http://scalability.llnl.gov/ > CASC @ Lawrence Livermore National Laboratory, Livermore, USA > > > > > > On 8/28/14, 8:30 AM, "Rolf Rabenseifner" <rabenseifner(a)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&status= >> assigned&status=new&status=reopened&milestone=2014%2F09%2F15+Kobe%2C+Japan >> &col=id&col=summary&col=milestone&col=status&col=type&col=priority&col=ver >> sion&order=priority >> >> Best regards >> Rolf >> >> ----- Original Message ----- >>> From: "Schulz Martin" <schulzm(a)llnl.gov> >>> To: "Main mailing list" <mpi-forum(a)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(a)llnl.gov, http://scalability.llnl.gov/ >>> CASC @ Lawrence Livermore National Laboratory, Livermore, USA >>> >>> >>> _______________________________________________ >>> mpi-forum mailing list >>> mpi-forum(a)lists.mpi-forum.org >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>> >> >> -- >> Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)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(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum -- Pavan Balaji ✉️ http://www.mcs.anl.gov/~balaji

1 0

Re: [Mpi-forum] September MPI Forum / Agenda and Reminder to Register
by Schulz Martin 02 Sep '14

02 Sep '14

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(a)llnl.gov, http://scalability.llnl.gov/ CASC @ Lawrence Livermore National Laboratory, Livermore, USA On 8/28/14, 8:30 AM, "Rolf Rabenseifner" <rabenseifner(a)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&status= >assigned&status=new&status=reopened&milestone=2014%2F09%2F15+Kobe%2C+Japan >&col=id&col=summary&col=milestone&col=status&col=type&col=priority&col=ver >sion&order=priority > >Best regards >Rolf > >----- Original Message ----- >> From: "Schulz Martin" <schulzm(a)llnl.gov> >> To: "Main mailing list" <mpi-forum(a)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(a)llnl.gov, http://scalability.llnl.gov/ >> CASC @ Lawrence Livermore National Laboratory, Livermore, USA >> >> >> _______________________________________________ >> mpi-forum mailing list >> mpi-forum(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >> > >-- >Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)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(a)lists.mpi-forum.org >http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum

1 0