November 2013 - mpiwg-rma - mailman3.cels.anl.gov

making dynamic windows suck less
by Jeff Hammond 27 Nov '13

27 Nov '13

Lots of people think - I take position on their correctness - that dynamic windows are doomed to substandard performance. I can certainly see some potential for this on networks that require memory registration. Can we try to add an opaque object that can encapsulate memory registration such that dynamic windows 2.0 would not have so many problems? If one were to send-recv not just the virtual address but MPI_Win_memobj (to be named properly later), then maybe we could deal with the shortcomings of IB, etc. w.r.t. RDMA and memory registration. This is really just a thought. I have no concrete proposal. Maybe people will have ideas in December. Jeff -- Jeff Hammond jeff.science(a)gmail.com

1 0

SC13 Best Paper Award for MPI-3 RMA paper
by Rajeev Thakur 22 Nov '13

22 Nov '13

FYI, this paper Enabling Highly-Scalable Remote Memory Access Programming with MPI-3 One Sided by Robert Gerstenberger, Maciej Besta, Torsten Hoefler http://htor.inf.ethz.ch/publications//img/gerstenberger-fompi-sc13.pdf won the Best Paper Award at SC13 yesterday. Congrats to Robert, Maciej, and Torsten, and it is also great that a paper on MPI-3 RMA won the best paper award! Rajeev

1 0

Re: [mpiwg-rma] MPI RMA telecon
by Jeff Hammond 17 Nov '13

17 Nov '13

Pavan and Bill rejected all possible options for the week after SC. I see no point in trying to schedule a telecon for the week before the Forum meets so I retract my Doodle poll. If people think that we need to have a telecon before the Forum, then please feel free to try to schedule one. Jeff On Sun, Nov 17, 2013 at 4:54 PM, Jeff Hammond (via Doodle) < mailer(a)doodle.com> wrote: > Jeff Hammond invites you to participate in the Doodle poll "MPI RMA > telecon." > > <https://doodle.com/?tmail=poll_invitecontact_participant_invitation&tlink=l…> > Hi there, > > Jeff Hammond (jeff.science(a)gmail.com) invites you to participate > in the Doodle poll "MPI RMA telecon". > Jeff Hammond says: > I suspect this week is no good for a telecon due to SC. I setup a Doodle > for the week after. The telecon is not going to be 4 hours long but I > didn't want to punch in 40 different time slots. Let's narrow down to AM or > PM first and then I'll fire off a fine-grain doodle for the blocks that > work. > > See http://lists.mpi-forum.org/mpiwg-rma/2013/11/1306.php for details. > > Jeff > Participate now<https://doodle.com/ie3phsmen5e76tqe?tmail=poll_invitecontact_participant_in…> > What is Doodle? Doodle is a web service that helps Jeff Hammond to find a > suitable date for meeting with a group of people. Learn more about how > Doodle works.<https://doodle.com/main.html?tlink=checkOutLink&tmail=poll_invitecontact_pa…> > You have received this e-mail because "Jeff Hammond" has invited > you to participate in the Doodle poll "MPI RMA telecon." > Doodle AG, Werdstrasse 21, 8021 Zürich > > > _______________________________________________ > mpiwg-rma mailing list > mpiwg-rma(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma > -- Jeff Hammond jeff.science(a)gmail.com

1 0

Re: [mpiwg-rma] schedule telecon
by Pavan Balaji 17 Nov '13

17 Nov '13

And one more item — We need an MPI_IN_PLACE capability for MPI_GET_ACCUMULATE, i.e., the origin and result buffer is the same. — Pavan On Nov 16, 2013, at 1:28 PM, Pavan Balaji <balaji(a)mcs.anl.gov> wrote: > > Sorry to keep adding more items, but one more item that needs discussion is the statement in the MPI-3 standard that could potentially disallow loads and PUTs to nonoverlapping locations in SEPARATE windows. Specifically, see #3 on page 455. > > “A put or accumulate must not access a target window once a load/store update or a put or accumulate update to another (overlapping) target window has started on a location in the target window, until the update becomes visible in the public copy of the window." > > The first line should either say “load/store access” or “store updates”. “load/store updates” doesn’t make sense. > > Now which of the two it is, is unclear. Is load + PUTs valid? This will be fine for hardware managed caches (even noncoherent ones), but not for software managed caches. > > One way to decouple this issue is to add more memory models — one that allows this and another that disallows this. IMO, there should be 4 different memory models that cover all cases: > > 1. MPI_WIN_SUCKY_MEMORY — for software managed cache type of systems. Even load + PUTs to nonoverlapping memory regions is not valid. > > 2. MPI_WIN_SEPARATE — current SEPARATE memory model, but we allow load + PUTs. This will work as long as the cache is hardware managed. > > 3. MPI_WIN_UNIFIED — this is what we decided on for our UNIFIED model. WIN_SYNC is required for overlapping memory regions. Windows are not always and immediately unified. They’ll eventually become unified without MPI calls. > > 4. MPI_WIN_SUPERUNIFIED — no WIN_SYNC required. Application can assume that the windows are truly unified. This works on x86 type of architectures. > > Of course, all this will be heavily debated. I’m just queuing up the discussion item. :-) > > — Pavan > > On Nov 4, 2013, at 1:55 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: > >> 369 seems trivial. Can we just enumerate the changes on the ticket >> and vote yes/no on moving forward with a reading in December on the >> telecon? >> >> Jeff >> >> On Mon, Nov 4, 2013 at 12:51 PM, Jim Dinan <james.dinan(a)gmail.com> wrote: >>> We also need to make sure that #349 is ready for a reading in December: >>> >>> MPI_Aint addressing arithmetic: >>> https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/349 >>> >>> I think the main todo here is to update the dynamic window examples. I also >>> have another open ticket that we should move forward: >>> >>> RMA needs same_disp info key: >>> https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/369 >>> >>> ~Jim. >>> >>> >>> On Sat, Nov 2, 2013 at 8:40 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: >>>> >>>> Yeah, I'll setup a doodle but I wanted to confirm the scope. Since >>>> Pavan greatly expanded the possible scope, I will propose a doodle >>>> that is oriented at discovering more than one time-slot since I >>>> suspect we will need that. >>>> >>>> Jeff >>>> >>>> On Sat, Nov 2, 2013 at 5:04 PM, Rajeev Thakur <thakur(a)mcs.anl.gov> wrote: >>>>> OK, when?. Do you want to set up a Doodle poll or something? >>>>> >>>>> Rajeev >>>>> >>>>> On Nov 2, 2013, at 4:12 PM, Jeff Hammond wrote: >>>>> >>>>>> I would like to schedule a telecon for the RMA WG to discuss the >>>>>> following three tickets. We've discussed them all over email but I >>>>>> want to get more conclusive feedback before the Chicago meeting. >>>>>> >>>>>> - extend the use of MPI_WIN_SHARED_QUERY to all windows >>>>>> (https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/397) >>>>>> - request-based remote completion for RMA >>>>>> (https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/398) >>>>>> - generalize same_op_no_op and allow user to specify all ops to be >>>>>> used (https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/399) >>>>>> >>>>>> Thanks, >>>>>> >>>>>> Jeff >>>>>> >>>>>> -- >>>>>> Jeff Hammond >>>>>> jeff.science(a)gmail.com >>>>>> _______________________________________________ >>>>>> mpiwg-rma mailing list >>>>>> mpiwg-rma(a)lists.mpi-forum.org >>>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma >>>>> >>>>> _______________________________________________ >>>>> mpiwg-rma mailing list >>>>> mpiwg-rma(a)lists.mpi-forum.org >>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma >>>> >>>> >>>> >>>> -- >>>> Jeff Hammond >>>> jeff.science(a)gmail.com >>>> _______________________________________________ >>>> mpiwg-rma mailing list >>>> mpiwg-rma(a)lists.mpi-forum.org >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma >>> >>> >>> >>> _______________________________________________ >>> mpiwg-rma mailing list >>> mpiwg-rma(a)lists.mpi-forum.org >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma >> >> >> >> -- >> Jeff Hammond >> jeff.science(a)gmail.com >> _______________________________________________ >> mpiwg-rma mailing list >> mpiwg-rma(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma > > -- > Pavan Balaji > http://www.mcs.anl.gov/~balaji > -- Pavan Balaji http://www.mcs.anl.gov/~balaji

1 0

Re: [mpiwg-rma] [Mpi-forum] Deadline for readings, votes, and errata items: 11/25!
by Jeff Hammond 17 Nov '13

17 Nov '13

Please make sure that RMA and Hybrid do not overlap. I know many of us need to attend both. Jeff On Sun, Nov 17, 2013 at 4:58 PM, Schulz, Martin <schulzm(a)llnl.gov> wrote: > > On Nov 17, 2013, at 3:55 PM, Pavan Balaji <balaji(a)mcs.anl.gov> > wrote: > >> Hi Martin, >> >> Can you add some WG time for RMA? We have a bunch of discussion items that we want to get resolved. > > Sure - one session, one half day? How much do you want? > > Martin > > >> >> — Pavan >> >> On Nov 17, 2013, at 3:20 PM, Schulz, Martin <schulzm(a)llnl.gov> wrote: >> >>> Hi all, >>> >>> Just as a reminder - based on the rules we voted in a few meetings ago, all readings, votes, and errata items must be announced two weeks prior to the start of the meeting though the MPI forum mailing list. The deadline for the December meeting is therefore November 25th, i.e., next Monday! >>> >>> Also, all working group chairs, please send me any requests for working group and/or plenary time, so I can post a complete schedule. If you don't want any time during this meeting, please let me know as well, so I can keep track. So far, I have only heard from a single group (thanks, Kathryn!). Also, as usual, please be prepared to give a brief working group update/summary on Wednesday morning when we start with the plenary time. >>> >>> Thanks! >>> >>> Martin >>> >>> >>> ________________________________________________________________________ >>> Martin Schulz, schulzm(a)llnl.gov, http://people.llnl.gov/schulzm >>> 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 >> >> -- >> Pavan Balaji >> http://www.mcs.anl.gov/~balaji >> > > ________________________________________________________________________ > Martin Schulz, schulzm(a)llnl.gov, http://people.llnl.gov/schulzm > CASC @ Lawrence Livermore National Laboratory, Livermore, USA > > > > _______________________________________________ > mpiwg-rma mailing list > mpiwg-rma(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma -- Jeff Hammond jeff.science(a)gmail.com

1 0

Re: [mpiwg-rma] [Mpi-forum] Deadline for readings, votes, and errata items: 11/25!
by Schulz, Martin 17 Nov '13

17 Nov '13

OK, no problem. Martin On Nov 17, 2013, at 3:59 PM, Pavan Balaji <balaji(a)mcs.anl.gov> wrote: > > On Nov 17, 2013, at 4:58 PM, Schulz, Martin <schulzm(a)llnl.gov> wrote: >> On Nov 17, 2013, at 3:55 PM, Pavan Balaji <balaji(a)mcs.anl.gov> >> wrote: >> >>> Hi Martin, >>> >>> Can you add some WG time for RMA? We have a bunch of discussion items that we want to get resolved. >> >> Sure - one session, one half day? How much do you want? > > Half day, I think. We have a big list of things to go through. > > — Pavan > > -- > Pavan Balaji > http://www.mcs.anl.gov/~balaji ________________________________________________________________________ Martin Schulz, schulzm(a)llnl.gov, http://people.llnl.gov/schulzm CASC @ Lawrence Livermore National Laboratory, Livermore, USA

1 0

Re: [mpiwg-rma] [Mpi-forum] Deadline for readings, votes, and errata items: 11/25!
by Pavan Balaji 17 Nov '13

17 Nov '13

On Nov 17, 2013, at 4:58 PM, Schulz, Martin <schulzm(a)llnl.gov> wrote: > On Nov 17, 2013, at 3:55 PM, Pavan Balaji <balaji(a)mcs.anl.gov> > wrote: > >> Hi Martin, >> >> Can you add some WG time for RMA? We have a bunch of discussion items that we want to get resolved. > > Sure - one session, one half day? How much do you want? Half day, I think. We have a big list of things to go through. — Pavan -- Pavan Balaji http://www.mcs.anl.gov/~balaji

1 0

Re: [mpiwg-rma] [Mpi-forum] Deadline for readings, votes, and errata items: 11/25!
by Schulz, Martin 17 Nov '13

17 Nov '13

On Nov 17, 2013, at 3:55 PM, Pavan Balaji <balaji(a)mcs.anl.gov> wrote: > Hi Martin, > > Can you add some WG time for RMA? We have a bunch of discussion items that we want to get resolved. Sure - one session, one half day? How much do you want? Martin > > — Pavan > > On Nov 17, 2013, at 3:20 PM, Schulz, Martin <schulzm(a)llnl.gov> wrote: > >> Hi all, >> >> Just as a reminder - based on the rules we voted in a few meetings ago, all readings, votes, and errata items must be announced two weeks prior to the start of the meeting though the MPI forum mailing list. The deadline for the December meeting is therefore November 25th, i.e., next Monday! >> >> Also, all working group chairs, please send me any requests for working group and/or plenary time, so I can post a complete schedule. If you don't want any time during this meeting, please let me know as well, so I can keep track. So far, I have only heard from a single group (thanks, Kathryn!). Also, as usual, please be prepared to give a brief working group update/summary on Wednesday morning when we start with the plenary time. >> >> Thanks! >> >> Martin >> >> >> ________________________________________________________________________ >> Martin Schulz, schulzm(a)llnl.gov, http://people.llnl.gov/schulzm >> 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 > > -- > Pavan Balaji > http://www.mcs.anl.gov/~balaji > ________________________________________________________________________ Martin Schulz, schulzm(a)llnl.gov, http://people.llnl.gov/schulzm CASC @ Lawrence Livermore National Laboratory, Livermore, USA

1 0

Re: [mpiwg-rma] [Mpi-forum] Deadline for readings, votes, and errata items: 11/25!
by Pavan Balaji 17 Nov '13

17 Nov '13

Hi Martin, Can you add some WG time for RMA? We have a bunch of discussion items that we want to get resolved. — Pavan On Nov 17, 2013, at 3:20 PM, Schulz, Martin <schulzm(a)llnl.gov> wrote: > Hi all, > > Just as a reminder - based on the rules we voted in a few meetings ago, all readings, votes, and errata items must be announced two weeks prior to the start of the meeting though the MPI forum mailing list. The deadline for the December meeting is therefore November 25th, i.e., next Monday! > > Also, all working group chairs, please send me any requests for working group and/or plenary time, so I can post a complete schedule. If you don't want any time during this meeting, please let me know as well, so I can keep track. So far, I have only heard from a single group (thanks, Kathryn!). Also, as usual, please be prepared to give a brief working group update/summary on Wednesday morning when we start with the plenary time. > > Thanks! > > Martin > > > ________________________________________________________________________ > Martin Schulz, schulzm(a)llnl.gov, http://people.llnl.gov/schulzm > 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 -- Pavan Balaji http://www.mcs.anl.gov/~balaji

1 0

MPI RMA telecon
by Jeff Hammond (via Doodle) 17 Nov '13

17 Nov '13

Hi there, Jeff Hammond (jeff.science(a)gmail.com) invites you to participate in the Doodle poll "MPI RMA telecon". Jeff Hammond says: I suspect this week is no good for a telecon due to SC. I setup a Doodle for the week after. The telecon is not going to be 4 hours long but I didn't want to punch in 40 different time slots. Let's narrow down to AM or PM first and then I'll fire off a fine-grain doodle for the blocks that work. See http://lists.mpi-forum.org/mpiwg-rma/2013/11/1306.php for details. Jeff Participate now https://doodle.com/ie3phsmen5e76tqe?tmail=poll_invitecontact_participant_in… What is Doodle? Doodle is a web service that helps Jeff Hammond to find a suitable date for meeting with a group of people. Learn more about how Doodle works. (https://doodle.com/main.html?tlink=checkOutLink&tmail=poll_invitecontact_pa…) ---------------------------------------------------------------------- You have received this e-mail because "Jeff Hammond" has invited you to participate in the Doodle poll "MPI RMA telecon." ---- Doodle AG, Werdstrasse 21, 8021 Zürich

1 0