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

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

02 Sep '14

Hi Pavan, I found the same things confusing (for a reading) as Jeff. Nevertheless, it may be good to add this to agenda to discuss the two options, unless you want to lead that discussion yourself next time. Thanks, Martin ________________________________________________________________________ Martin Schulz, schulzm(a)llnl.gov, http://scalability.llnl.gov/ CASC @ Lawrence Livermore National Laboratory, Livermore, USA On 9/2/14, 3:30 PM, "Balaji, Pavan" <balaji(a)anl.gov> wrote: > >On Sep 2, 2014, at 5:27 PM, Jeff Squyres (jsquyres) <jsquyres(a)cisco.com> >wrote: > >> On Sep 2, 2014, at 6:21 PM, Balaji, Pavan <balaji(a)anl.gov> wrote: >> >>> The text has a single combined version. There’s no two options in the >>>text. >> >> From 411: >> >> "Proposed Solution: >> >> ... >> Two possible solutions are being proposed: >> >> • We reserve the "mpi_" namespace and forbid implementations from >>using keys in that name space. >> • Solution 1 + we rename existing info keys to start with "mpi_" and >>deprecate the older info key names. >> The attached PDF provides solution 2, but it can easily be modified to >>solution 1 by simply dropping some changes, since solution 1 is a proper >>subset of solution 2. Specifically, the only change for solution 1 will >>be the one sentence addition on the first page of Chapter 9 (page 365)." >> >> Does your PDF not agree with your ticket text? >> >> That is most confusing. > >OK. The pdf is OK, but I’ll update the ticket text. > >>>> +1. It looks like you missed the deadline due to inactivity. >>> >>> Did I miss an email asking for tickets to be listed? I thought we >>>just upload the document on the ticket, which I did over a month ago. >>>But I don’t want to fight for it. If folks feel this can’t be a formal >>>reading in September, we can do this in December. >> >> Per the rules document: >> >> "General text proposals use the following process to be accepted into >>an MPI >> standards document: >> >> 1. Have a formal reading at a physical MPI Forum meeting where the >>meeting >> quorum has been met. >> >> (a) The final text of the proposal to be read must be made publicly >> available via the general MPI Forum broadcast email list at least >> two weeks prior to the start date of the physical MPI Forum meeting >> at which it is to be formally read." >> >> I don't see an email from you on the mpi-forum mailing list announcing >>411 until today. > >OK, thanks. I’ll keep that in mind for the future. > > — Pavan > >-- >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 Schulz Martin 02 Sep '14

02 Sep '14

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(a)llnl.gov, http://scalability.llnl.gov/ CASC @ Lawrence Livermore National Laboratory, Livermore, USA On 9/2/14, 3:23 PM, "Jeff Squyres (jsquyres)" <jsquyres(a)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(a)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(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 >> >> _______________________________________________ >> mpi-forum mailing list >> mpi-forum(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > >-- >Jeff Squyres >jsquyres(a)cisco.com >For corporate legal information go to: >http://www.cisco.com/web/about/doing_business/legal/cri/ > >_______________________________________________ >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] URGENT: Register for MPI Forum
by Jeff Squyres (jsquyres) 02 Sep '14

02 Sep '14

We now have enough for meeting quorum. We needed >=12 qualifying organizations to register. As of this moment, we have 15. On Sep 2, 2014, at 5:36 PM, Schulz Martin <schulzm(a)llnl.gov> wrote: > Hi all, > > If you haven’t 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 didn’t 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 > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0

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

02 Sep '14

On Sep 2, 2014, at 5:27 PM, Jeff Squyres (jsquyres) <jsquyres(a)cisco.com> wrote: > On Sep 2, 2014, at 6:21 PM, Balaji, Pavan <balaji(a)anl.gov> wrote: > >> The text has a single combined version. There’s no two options in the text. > > From 411: > > "Proposed Solution: > > ... > Two possible solutions are being proposed: > > • We reserve the "mpi_" namespace and forbid implementations from using keys in that name space. > • Solution 1 + we rename existing info keys to start with "mpi_" and deprecate the older info key names. > The attached PDF provides solution 2, but it can easily be modified to solution 1 by simply dropping some changes, since solution 1 is a proper subset of solution 2. Specifically, the only change for solution 1 will be the one sentence addition on the first page of Chapter 9 (page 365)." > > Does your PDF not agree with your ticket text? > > That is most confusing. OK. The pdf is OK, but I’ll update the ticket text. >>> +1. It looks like you missed the deadline due to inactivity. >> >> Did I miss an email asking for tickets to be listed? I thought we just upload the document on the ticket, which I did over a month ago. But I don’t want to fight for it. If folks feel this can’t be a formal reading in September, we can do this in December. > > Per the rules document: > > "General text proposals use the following process to be accepted into an MPI > standards document: > > 1. Have a formal reading at a physical MPI Forum meeting where the meeting > quorum has been met. > > (a) The final text of the proposal to be read must be made publicly > available via the general MPI Forum broadcast email list at least > two weeks prior to the start date of the physical MPI Forum meeting > at which it is to be formally read." > > I don't see an email from you on the mpi-forum mailing list announcing 411 until today. OK, thanks. I’ll keep that in mind for the future. — Pavan -- Pavan Balaji ✉️ http://www.mcs.anl.gov/~balaji

1 0

Re: [Mpi-forum] September MPI Forum / Agenda and Reminder to Register
by Jeff Squyres (jsquyres) 02 Sep '14

02 Sep '14

On Sep 2, 2014, at 6:21 PM, Balaji, Pavan <balaji(a)anl.gov> wrote: > > On Sep 2, 2014, at 5:17 PM, Jeff Squyres (jsquyres) <jsquyres(a)cisco.com> wrote: > >> On Sep 2, 2014, at 5:39 PM, Schulz Martin <schulzm(a)llnl.gov> wrote: >> >>> As we can add this, but I don’t see how this can be a formal reading. >>> There is no single text - there are still two options. > > The text has a single combined version. There’s no two options in the text. >From 411: "Proposed Solution: ... Two possible solutions are being proposed: • We reserve the "mpi_" namespace and forbid implementations from using keys in that name space. • Solution 1 + we rename existing info keys to start with "mpi_" and deprecate the older info key names. The attached PDF provides solution 2, but it can easily be modified to solution 1 by simply dropping some changes, since solution 1 is a proper subset of solution 2. Specifically, the only change for solution 1 will be the one sentence addition on the first page of Chapter 9 (page 365)." Does your PDF not agree with your ticket text? That is most confusing. >> +1. It looks like you missed the deadline due to inactivity. > > Did I miss an email asking for tickets to be listed? I thought we just upload the document on the ticket, which I did over a month ago. But I don’t want to fight for it. If folks feel this can’t be a formal reading in September, we can do this in December. Per the rules document: "General text proposals use the following process to be accepted into an MPI standards document: 1. Have a formal reading at a physical MPI Forum meeting where the meeting quorum has been met. (a) The final text of the proposal to be read must be made publicly available via the general MPI Forum broadcast email list at least two weeks prior to the start date of the physical MPI Forum meeting at which it is to be formally read." I don't see an email from you on the mpi-forum mailing list announcing 411 until today. -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0

Re: [Mpi-forum] September MPI Forum / Agenda and Reminder to Register
by Jeff Squyres (jsquyres) 02 Sep '14

02 Sep '14

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(a)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(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 > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0

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

02 Sep '14

On Sep 2, 2014, at 5:17 PM, Jeff Squyres (jsquyres) <jsquyres(a)cisco.com> wrote: > On Sep 2, 2014, at 5:39 PM, Schulz Martin <schulzm(a)llnl.gov> wrote: > >> As we can add this, but I don’t see how this can be a formal reading. >> There is no single text - there are still two options. The text has a single combined version. There’s no two options in the text. > +1. It looks like you missed the deadline due to inactivity. Did I miss an email asking for tickets to be listed? I thought we just upload the document on the ticket, which I did over a month ago. But I don’t want to fight for it. If folks feel this can’t be a formal reading in September, we can do this in December. — Pavan -- Pavan Balaji ✉️ http://www.mcs.anl.gov/~balaji

1 0

Re: [Mpi-forum] September MPI Forum / Agenda and Reminder to Register
by Jeff Squyres (jsquyres) 02 Sep '14

02 Sep '14

On Sep 2, 2014, at 5:39 PM, Schulz Martin <schulzm(a)llnl.gov> wrote: > As we can add this, but I don’t see how this can be a formal reading. > There is no single text - there are still two options. +1. It looks like you missed the deadline due to inactivity. -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0

Voting at the Japan meeting
by Jeff Squyres (jsquyres) 02 Sep '14

02 Sep '14

Sadly, I will not be attending the upcoming meeting in Japan in a few weeks. Dave Goodell will be attending; he will perform the Forum secretary duties in my stead. -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0

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

02 Sep '14

Hi Pavan, As we can add this, but I don’t see how this can be a formal reading. There is no single text - there are still two options. Should this rather be a plenary discussion to decide which option to pursue as a reading in December? Martin ________________________________________________________________________ Martin Schulz, schulzm(a)llnl.gov, http://scalability.llnl.gov/ CASC @ Lawrence Livermore National Laboratory, Livermore, USA On 9/2/14, 8:44 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&sta >>>>>tu >>>>> s= >>>>> >>>>> >>>>>assigned&status=new&status=reopened&milestone=2014%2F09%2F15+Kobe%2C+J >>>>>ap >>>>> 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