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

Re: [Mpi3-rma] [EXTERNAL] Re: Call for errata
by Barrett, Brian W 28 Jan '13

28 Jan '13

I'm well aware. I was mostly suggesting that we not do something that's a total waste because Rolf suggested it. If the right thing is to add the Version, sending an e-mail to Jeff Squyres isn't that hard. Doing silly things to work around such an e-mail is, well, silly. Brian On 1/28/13 2:30 PM, "Jeff Hammond" <jhammond(a)alcf.anl.gov> wrote: >There is a reason and it is that there is no option for "MPI 3.0 >Errata" in the drop-down "Version" box. Obviously this is a fixable >problem but outside the scope of this WG. > >Jeff > >On Tue, Jan 29, 2013 at 6:26 AM, Barrett, Brian W <bwbarre(a)sandia.gov> >wrote: >> Rolf might like that, but there's really no reason for the title to >> include MPI-3.0 errata. If we want another selector, that makes sense, >> but having it in the title just burns good screen real estateďż˝ >> >> Brian >> >> On 1/28/13 2:23 PM, "Jim Dinan" <dinan(a)mcs.anl.gov> wrote: >> >>>Hi Pavan, >>> >>>Could you please log these as tickets, so we don't lose any items. >>>Please include the exact change, if it's ready. Also, Rolf would like >>>us to use the the following formatting for title/ticket type: >>> >>>Title: MPI-3.0 errata: xxxxxxxxxxx >>>Type: Correction to standard (for upcoming release + errata) >>> >>>Thanks, >>> ~Jim. >>> >>>On 1/28/13 2:33 PM, Pavan Balaji wrote: >>>> Jim, >>>> >>>> We talked about a few things in the last Forum. Slides are online on >>>> the Forum page. I also sent an email about it after the Forum; I can >>>> try to dig it up if you can't find it. >>>> >>>> -- Pavan >>>> >>>> On 01/28/2013 09:54 AM US Central Time, Jim Dinan wrote: >>>>> Hi All, >>>>> >>>>> We are assembling the list of RMA errata to be voted on at the March >>>>> meeting. Currently, we have tickets #347, #348, #350, and #355 (I'm >>>>> working on official wording for these changes and will circulate it >>>>>soon). >>>>> >>>>> Are there any other errata items that should be included on the >>>>>agenda >>>>> for the upcoming meeting? >>>>> >>>>> Thanks, >>>>> ~Jim. >>>>> _______________________________________________ >>>>> mpi3-rma mailing list >>>>> mpi3-rma(a)lists.mpi-forum.org >>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma >>>> >>>_______________________________________________ >>>mpi3-rma mailing list >>>mpi3-rma(a)lists.mpi-forum.org >>>http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma >>> >>> >> >> >> -- >> Brian W. Barrett >> Scalable System Software Group >> Sandia National Laboratories >> >> >> >> >> >> _______________________________________________ >> mpi3-rma mailing list >> mpi3-rma(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma > > > >-- >Jeff Hammond >Argonne Leadership Computing Facility >University of Chicago Computation Institute >jhammond(a)alcf.anl.gov / (630) 252-5381 >http://www.linkedin.com/in/jeffhammond >https://wiki.alcf.anl.gov/parts/index.php/User:Jhammond > -- Brian W. Barrett Scalable System Software Group Sandia National Laboratories

1 0

Re: [Mpi3-rma] [EXTERNAL] Re: Call for errata
by Jeff Hammond 28 Jan '13

28 Jan '13

There is a reason and it is that there is no option for "MPI 3.0 Errata" in the drop-down "Version" box. Obviously this is a fixable problem but outside the scope of this WG. Jeff On Tue, Jan 29, 2013 at 6:26 AM, Barrett, Brian W <bwbarre(a)sandia.gov> wrote: > Rolf might like that, but there's really no reason for the title to > include MPI-3.0 errata. If we want another selector, that makes sense, > but having it in the title just burns good screen real estateŠ > > Brian > > On 1/28/13 2:23 PM, "Jim Dinan" <dinan(a)mcs.anl.gov> wrote: > >>Hi Pavan, >> >>Could you please log these as tickets, so we don't lose any items. >>Please include the exact change, if it's ready. Also, Rolf would like >>us to use the the following formatting for title/ticket type: >> >>Title: MPI-3.0 errata: xxxxxxxxxxx >>Type: Correction to standard (for upcoming release + errata) >> >>Thanks, >> ~Jim. >> >>On 1/28/13 2:33 PM, Pavan Balaji wrote: >>> Jim, >>> >>> We talked about a few things in the last Forum. Slides are online on >>> the Forum page. I also sent an email about it after the Forum; I can >>> try to dig it up if you can't find it. >>> >>> -- Pavan >>> >>> On 01/28/2013 09:54 AM US Central Time, Jim Dinan wrote: >>>> Hi All, >>>> >>>> We are assembling the list of RMA errata to be voted on at the March >>>> meeting. Currently, we have tickets #347, #348, #350, and #355 (I'm >>>> working on official wording for these changes and will circulate it >>>>soon). >>>> >>>> Are there any other errata items that should be included on the agenda >>>> for the upcoming meeting? >>>> >>>> Thanks, >>>> ~Jim. >>>> _______________________________________________ >>>> mpi3-rma mailing list >>>> mpi3-rma(a)lists.mpi-forum.org >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma >>> >>_______________________________________________ >>mpi3-rma mailing list >>mpi3-rma(a)lists.mpi-forum.org >>http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma >> >> > > > -- > Brian W. Barrett > Scalable System Software Group > Sandia National Laboratories > > > > > > _______________________________________________ > mpi3-rma mailing list > mpi3-rma(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma -- Jeff Hammond Argonne Leadership Computing Facility University of Chicago Computation Institute jhammond(a)alcf.anl.gov / (630) 252-5381 http://www.linkedin.com/in/jeffhammond https://wiki.alcf.anl.gov/parts/index.php/User:Jhammond

1 0

Re: [Mpi3-rma] [EXTERNAL] Re: Call for errata
by Barrett, Brian W 28 Jan '13

28 Jan '13

Rolf might like that, but there's really no reason for the title to include MPI-3.0 errata. If we want another selector, that makes sense, but having it in the title just burns good screen real estateï¿½ Brian On 1/28/13 2:23 PM, "Jim Dinan" <dinan(a)mcs.anl.gov> wrote: >Hi Pavan, > >Could you please log these as tickets, so we don't lose any items. >Please include the exact change, if it's ready. Also, Rolf would like >us to use the the following formatting for title/ticket type: > >Title: MPI-3.0 errata: xxxxxxxxxxx >Type: Correction to standard (for upcoming release + errata) > >Thanks, > ~Jim. > >On 1/28/13 2:33 PM, Pavan Balaji wrote: >> Jim, >> >> We talked about a few things in the last Forum. Slides are online on >> the Forum page. I also sent an email about it after the Forum; I can >> try to dig it up if you can't find it. >> >> -- Pavan >> >> On 01/28/2013 09:54 AM US Central Time, Jim Dinan wrote: >>> Hi All, >>> >>> We are assembling the list of RMA errata to be voted on at the March >>> meeting. Currently, we have tickets #347, #348, #350, and #355 (I'm >>> working on official wording for these changes and will circulate it >>>soon). >>> >>> Are there any other errata items that should be included on the agenda >>> for the upcoming meeting? >>> >>> Thanks, >>> ~Jim. >>> _______________________________________________ >>> mpi3-rma mailing list >>> mpi3-rma(a)lists.mpi-forum.org >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma >> >_______________________________________________ >mpi3-rma mailing list >mpi3-rma(a)lists.mpi-forum.org >http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma > > -- Brian W. Barrett Scalable System Software Group Sandia National Laboratories

1 0

Re: [Mpi3-rma] Call for errata
by Jim Dinan 28 Jan '13

28 Jan '13

Hi Pavan, Could you please log these as tickets, so we don't lose any items. Please include the exact change, if it's ready. Also, Rolf would like us to use the the following formatting for title/ticket type: Title: MPI-3.0 errata: xxxxxxxxxxx Type: Correction to standard (for upcoming release + errata) Thanks, ~Jim. On 1/28/13 2:33 PM, Pavan Balaji wrote: > Jim, > > We talked about a few things in the last Forum. Slides are online on > the Forum page. I also sent an email about it after the Forum; I can > try to dig it up if you can't find it. > > -- Pavan > > On 01/28/2013 09:54 AM US Central Time, Jim Dinan wrote: >> Hi All, >> >> We are assembling the list of RMA errata to be voted on at the March >> meeting. Currently, we have tickets #347, #348, #350, and #355 (I'm >> working on official wording for these changes and will circulate it soon). >> >> Are there any other errata items that should be included on the agenda >> for the upcoming meeting? >> >> Thanks, >> ~Jim. >> _______________________________________________ >> mpi3-rma mailing list >> mpi3-rma(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma >

1 0

Re: [Mpi3-rma] Call for errata
by Pavan Balaji 28 Jan '13

28 Jan '13

Jim, We talked about a few things in the last Forum. Slides are online on the Forum page. I also sent an email about it after the Forum; I can try to dig it up if you can't find it. -- Pavan On 01/28/2013 09:54 AM US Central Time, Jim Dinan wrote: > Hi All, > > We are assembling the list of RMA errata to be voted on at the March > meeting. Currently, we have tickets #347, #348, #350, and #355 (I'm > working on official wording for these changes and will circulate it soon). > > Are there any other errata items that should be included on the agenda > for the upcoming meeting? > > Thanks, > ~Jim. > _______________________________________________ > mpi3-rma mailing list > mpi3-rma(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma -- Pavan Balaji http://www.mcs.anl.gov/~balaji

1 0

Re: [Mpi3-rma] Call for errata
by Hubert Ritzdorf 28 Jan '13

28 Jan '13

Hi, Page 429, Line 41-42 specifies for MPI_Fetch_and_op: The datatype argument must be a predefined datatype. I think it should be: The datatype argument must be a predefined or MPI basic datatype (cf. Section 5.9.2 on page 176). Reasons: (*) This would be consistent to the reduce functions and MPI_Compare_and_swap and would allow handles returned by MPI_TYPE_CREATE_F90_INTEGER, MPI_TYPE_CREATE_F90_REAL, or MPI_TYPE_F90_COMPLEX. (which are currently excluded since they are not predefined). (*) MPI_Compare_and_swap accepts handles created by MPI_TYPE_CREATE_F90_INTEGER (they are contained in Fortran integer category). (*) This proposed change support also atomic gets of MPI_WCHAR. Additionally, on Page 430, Line 39: predefined datatypes should be replaced by basic datatypes since Section 5.9.2 defines the MPI basic datatypes and not the predefined datatypes (which are defined in Tables 3.x). Ticket 350 (MPI_NO_OP in new accumulate functions, origin buffer arguments are ignored) should also state that the number of entries transferred is defined by target_count Hubert ________________________________________ From: mpi3-rma-bounces(a)lists.mpi-forum.org [mpi3-rma-bounces(a)lists.mpi-forum.org] on behalf of Jim Dinan [dinan(a)mcs.anl.gov] Sent: Monday, January 28, 2013 4:54 PM To: MPI 3.0 Remote Memory Access working group Subject: [Mpi3-rma] Call for errata Hi All, We are assembling the list of RMA errata to be voted on at the March meeting. Currently, we have tickets #347, #348, #350, and #355 (I'm working on official wording for these changes and will circulate it soon). Are there any other errata items that should be included on the agenda for the upcoming meeting? Thanks, ~Jim. _______________________________________________ mpi3-rma mailing list mpi3-rma(a)lists.mpi-forum.org http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma Click https://www.mailcontrol.com/sr/cqZpjZJWCcHGX2PQPOmvUsQN7C0BBstkddyjT4P!KFPM… to report this email as spam.

1 0

[Mpi3-rma] Call for errata
by Jim Dinan 28 Jan '13

28 Jan '13

Hi All, We are assembling the list of RMA errata to be voted on at the March meeting. Currently, we have tickets #347, #348, #350, and #355 (I'm working on official wording for these changes and will circulate it soon). Are there any other errata items that should be included on the agenda for the upcoming meeting? Thanks, ~Jim.

1 0

Re: [Mpi3-rma] MPI_Win_allocate_shared and Win_free
by Hubert Ritzdorf 18 Jan '13

18 Jan '13

Hi Pavan, thanks for the fast answer. Hubert ________________________________________ From: Pavan Balaji [balaji(a)mcs.anl.gov] Sent: Friday, January 18, 2013 2:58 PM To: mpi3-rma(a)lists.mpi-forum.org Cc: Hubert Ritzdorf Subject: Re: [Mpi3-rma] MPI_Win_allocate_shared and Win_free Isn't this the same problem as the user calling "free" on a buffer given to MPI_Win_create before doing an MPI_Win_free? This should be a user error, though I'm not sure we call that out in the standard. -- Pavan On 01/18/2013 07:26 AM US Central Time, Hubert Ritzdorf wrote: > Hi, > > I have a question to the memory allocated in MPI_Win_allocate_shared. > What happens if (see also pseudo code below) > > (*) the allocated memory is used in another MPI_Win_create and > (*) the window created by MPI_Win_allocate_shared is freed and > (*) the user program still works with the allocates shared memory > before the window created by MPI_Win_create is freed ? > > MPI 3.0, page 415, lines 6-8 specifies: > > If the window was created with MPI_WIN_ALLOCATE_SHARED, > MPI_WIN_FREE will free the window memory that was allocated > in MPI_WIN_ALLOCATE_SHARED. > > Thus main question: > > (a) Should the corresponding program fail or > (b) should the program work correctly (i..e. the MPI library recognizes > that the memory is still used in another window) and the shared memory > is freed when the window created by MPI_Win_create is freed ? > > Best regards > > Hubert > > ------- Pseudo code ------- > > MPI_Win_allocate_shared (shared_size, disp_unit, shared_info, comm, > &shared_baseptr, &shared_win); > > MPI_Win_create (shared_baseptr, shared_size, disp_unit, > info2, comm, &win2); > > /* Non-concurrent work/communication window shared_win */ > ... > > /* Non-concurrent communication on window win2 */ > ... > > /* Free window shared_win; > Is shared_baseptr really deallocated in MPI_Win_free > as defined on page 415 lines 6-8 ? */ > > MPI_Win_free (&shared_win); > > /* Work on window win2 */ > ... > > /* Free window win2; > Is shared_baseptr deallocated in following MPI_Win_free */ > > MPI_Win_free (&win2); > > > > > _______________________________________________ > mpi3-rma mailing list > mpi3-rma(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma > -- Pavan Balaji http://www.mcs.anl.gov/~balaji Click https://www.mailcontrol.com/sr/jsfgRi1ZxR!GX2PQPOmvUuzn0lj6veCA855EakXxQgJE… to report this email as spam.

1 0

Re: [Mpi3-rma] MPI_Win_allocate_shared and Win_free
by Pavan Balaji 18 Jan '13

18 Jan '13

Isn't this the same problem as the user calling "free" on a buffer given to MPI_Win_create before doing an MPI_Win_free? This should be a user error, though I'm not sure we call that out in the standard. -- Pavan On 01/18/2013 07:26 AM US Central Time, Hubert Ritzdorf wrote: > Hi, > > I have a question to the memory allocated in MPI_Win_allocate_shared. > What happens if (see also pseudo code below) > > (*) the allocated memory is used in another MPI_Win_create and > (*) the window created by MPI_Win_allocate_shared is freed and > (*) the user program still works with the allocates shared memory > before the window created by MPI_Win_create is freed ? > > MPI 3.0, page 415, lines 6-8 specifies: > > If the window was created with MPI_WIN_ALLOCATE_SHARED, > MPI_WIN_FREE will free the window memory that was allocated > in MPI_WIN_ALLOCATE_SHARED. > > Thus main question: > > (a) Should the corresponding program fail or > (b) should the program work correctly (i..e. the MPI library recognizes > that the memory is still used in another window) and the shared memory > is freed when the window created by MPI_Win_create is freed ? > > Best regards > > Hubert > > ------- Pseudo code ------- > > MPI_Win_allocate_shared (shared_size, disp_unit, shared_info, comm, > &shared_baseptr, &shared_win); > > MPI_Win_create (shared_baseptr, shared_size, disp_unit, > info2, comm, &win2); > > /* Non-concurrent work/communication window shared_win */ > ... > > /* Non-concurrent communication on window win2 */ > ... > > /* Free window shared_win; > Is shared_baseptr really deallocated in MPI_Win_free > as defined on page 415 lines 6-8 ? */ > > MPI_Win_free (&shared_win); > > /* Work on window win2 */ > ... > > /* Free window win2; > Is shared_baseptr deallocated in following MPI_Win_free */ > > MPI_Win_free (&win2); > > > > > _______________________________________________ > mpi3-rma mailing list > mpi3-rma(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma > -- Pavan Balaji http://www.mcs.anl.gov/~balaji

1 0

[Mpi3-rma] MPI_Win_allocate_shared and Win_free
by Hubert Ritzdorf 18 Jan '13

18 Jan '13

Hi, I have a question to the memory allocated in MPI_Win_allocate_shared. What happens if (see also pseudo code below) (*) the allocated memory is used in another MPI_Win_create and (*) the window created by MPI_Win_allocate_shared is freed and (*) the user program still works with the allocates shared memory before the window created by MPI_Win_create is freed ? MPI 3.0, page 415, lines 6-8 specifies: If the window was created with MPI_WIN_ALLOCATE_SHARED, MPI_WIN_FREE will free the window memory that was allocated in MPI_WIN_ALLOCATE_SHARED. Thus main question: (a) Should the corresponding program fail or (b) should the program work correctly (i..e. the MPI library recognizes that the memory is still used in another window) and the shared memory is freed when the window created by MPI_Win_create is freed ? Best regards Hubert ------- Pseudo code ------- MPI_Win_allocate_shared (shared_size, disp_unit, shared_info, comm, &shared_baseptr, &shared_win); MPI_Win_create (shared_baseptr, shared_size, disp_unit, info2, comm, &win2); /* Non-concurrent work/communication window shared_win */ ... /* Non-concurrent communication on window win2 */ ... /* Free window shared_win; Is shared_baseptr really deallocated in MPI_Win_free as defined on page 415 lines 6-8 ? */ MPI_Win_free (&shared_win); /* Work on window win2 */ ... /* Free window win2; Is shared_baseptr deallocated in following MPI_Win_free */ MPI_Win_free (&win2);

1 0