May 2009 - Mpi-22 - mailman3.cels.anl.gov

Re: [Mpi-22] [Mpi-forum] The const proposal - Ticket 140
by unknown＠example.com 27 May '09

27 May '09

I would like to add Mellanox to the list of organizations that support this letter. Pasha. Richard Treumann wrote: > > All - > > The signatories of this letter represent the majority of MPI > implementors participating in the MPI Forum. We are concerned that > proposal #140 ("Add const Keyword to the C bindings") has a number of > issues which suggest delaying to MPI-3 would be appropriate. > > In particular, the proposal: > > - Is likely to pass a simple majority vote, but does not carry the > support of the majority of MPI implementors, suggesting consensus has > not been reached. > - Changes 90+ MPI API interfaces, which is not a "trivial" change and > therefore does not meet the intent of the MPI-2.2 process. > - Is not needed to fix any serious bug in the standard text or to > solve an issue that cannot easily be avoided by the MPI application. > - Does not offer any demonstrable optimization opportunities for > implementation or application, but may constrain future implementation > opportunities. > > Therefore, we ask for your assistance in deferring proposal #140 to > the MPI-3 process, where more time can be spent assessing its impact. > > Thank you, > > - Cisco: Jeff Squires > - Intel: Alexander Supalov & Keith Underwood > - Sandia: Brian Barrett > - IBM: Richard Treumann > - QLogic: Avneesh Pant > - UTenn: George Bosilca > - HP: David George Solt > - UHouston: Edgar Gabriel > - Myricom: Patrick Geoffray > - ORNL: Richard Graham > - Sun: Terry Dontje > - NEC: Hubert Ritzdorf & Jesper Traeff > > > > Dick Treumann - MPI Team > IBM Systems & Technology Group > Dept X2ZA / MS P963 -- 2455 South Road -- Poughkeepsie, NY 12601 > Tele (845) 433-7846 Fax (845) 433-8363 > > ------------------------------------------------------------------------ > > _______________________________________________ > mpi-forum mailing list > mpi-forum_at_[hidden] > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >

1 0

Re: [Mpi-22] [Mpi-forum] The const proposal - Ticket 140
by unknown＠example.com 27 May '09

27 May '09

This is an example of why we have two votes - to give everyone a chance to take a second look at the issues. I'll note that the 17-10 (counting no's and abstains together) is worrisome; while a majority vote is the rule, a good standard will make a compelling argument for each feature. The process here would normally be to have a debate and then the second vote. However, for MPI 2.2, we have the additional requirements of limited scope of change to implementations - we didn't define what that meant precisely (and that is a good thing), but there is a strong argument that limited scope of change would require at least a majority of implementations to agree that the change is minor. Bill On May 27, 2009, at 12:48 PM, Erez Haba wrote: > Hi all, Im not really sure how to respond to this email. I will > just note that this proposal has passed 1st vote last September with > the following results. > > 4. Vote topic: MPI-2.2 const for C bindings, 1st vote: > YES: > 17 > NO: > 4 > ABSTAIN: > 6 > MISSED: > 0 > Result: > Ballot passed > > http://meetings.mpi-forum.org/secretary/2008/09/votes.php > > since then we decided to postpone the 1st vote to add very few minor > correction to ticket #46 (the original proposal) and thus created > ticket #140. > > I believe that all the points you mention below have been discussed > in the forum meeting(s). > > Thanks, > .Erez > > From: mpi-forum-bounces_at_[hidden] [mailto:mpi-forum-bounces_at_[hidden] > ] On Behalf Of Richard Treumann > Sent: Wednesday, May 27, 2009 6:53 AM > To: Main MPI Forum mailing list > Cc: mpi-22_at_[hidden] > Subject: [Mpi-forum] The const proposal - Ticket 140 > > All - > > The signatories of this letter represent the majority of MPI > implementors participating in the MPI Forum. We are concerned that > proposal #140 ("Add const Keyword to the C bindings") has a number > of issues which suggest delaying to MPI-3 would be appropriate. > > In particular, the proposal: > > - Is likely to pass a simple majority vote, but does not carry the > support of the majority of MPI implementors, suggesting consensus > has not been reached. > - Changes 90+ MPI API interfaces, which is not a "trivial" change > and therefore does not meet the intent of the MPI-2.2 process. > - Is not needed to fix any serious bug in the standard text or to > solve an issue that cannot easily be avoided by the MPI application. > - Does not offer any demonstrable optimization opportunities for > implementation or application, but may constrain future > implementation opportunities. > > Therefore, we ask for your assistance in deferring proposal #140 to > the MPI-3 process, where more time can be spent assessing its impact. > > Thank you, > > - Cisco: Jeff Squires > - Intel: Alexander Supalov & Keith Underwood > - Sandia: Brian Barrett > - IBM: Richard Treumann > - QLogic: Avneesh Pant > - UTenn: George Bosilca > - HP: David George Solt > - UHouston: Edgar Gabriel > - Myricom: Patrick Geoffray > - ORNL: Richard Graham > - Sun: Terry Dontje > - NEC: Hubert Ritzdorf & Jesper Traeff > > > > Dick Treumann - MPI Team > IBM Systems & Technology Group > Dept X2ZA / MS P963 -- 2455 South Road -- Poughkeepsie, NY 12601 > Tele (845) 433-7846 Fax (845) 433-8363 > > <ATT00001.txt> William Gropp Deputy Director for Research Institute for Advanced Computing Applications and Technologies Paul and Cynthia Saylor Professor of Computer Science University of Illinois Urbana-Champaign *

1 0

Re: [Mpi-22] [Mpi-forum] The const proposal - Ticket 140
by unknown＠example.com 27 May '09

27 May '09

Hi all, I'm not really sure how to respond to this email. I will just note that this proposal has passed 1st vote last September with the following results. 4. Vote topic: MPI-2.2 const for C bindings, 1st vote: YES: 17 NO: 4 ABSTAIN: 6 MISSED: 0 Result: Ballot passed http://meetings.mpi-forum.org/secretary/2008/09/votes.php ince then we decided to postpone the 1st vote to add very few minor correction to ticket #46 (the original proposal) and thus created ticket #140. I believe that all the points you mention below have been discussed in the forum meeting(s). Thanks, .Erez From: mpi-forum-bounces_at_[hidden] [mailto:mpi-forum-bounces_at_[hidden]] On Behalf Of Richard Treumann Sent: Wednesday, May 27, 2009 6:53 AM To: Main MPI Forum mailing list Cc: mpi-22_at_[hidden] Subject: [Mpi-forum] The const proposal - Ticket 140 All - The signatories of this letter represent the majority of MPI implementors participating in the MPI Forum. We are concerned that proposal #140 ("Add const Keyword to the C bindings") has a number of issues which suggest delaying to MPI-3 would be appropriate. In particular, the proposal: - Is likely to pass a simple majority vote, but does not carry the support of the majority of MPI implementors, suggesting consensus has not been reached. - Changes 90+ MPI API interfaces, which is not a "trivial" change and therefore does not meet the intent of the MPI-2.2 process. - Is not needed to fix any serious bug in the standard text or to solve an issue that cannot easily be avoided by the MPI application. - Does not offer any demonstrable optimization opportunities for implementation or application, but may constrain future implementation opportunities. Therefore, we ask for your assistance in deferring proposal #140 to the MPI-3 process, where more time can be spent assessing its impact. Thank you, - Cisco: Jeff Squires - Intel: Alexander Supalov & Keith Underwood - Sandia: Brian Barrett - IBM: Richard Treumann - QLogic: Avneesh Pant - UTenn: George Bosilca - HP: David George Solt - UHouston: Edgar Gabriel - Myricom: Patrick Geoffray - ORNL: Richard Graham - Sun: Terry Dontje - NEC: Hubert Ritzdorf & Jesper Traeff Dick Treumann - MPI Team IBM Systems & Technology Group Dept X2ZA / MS P963 -- 2455 South Road -- Poughkeepsie, NY 12601 Tele (845) 433-7846 Fax (845) 433-8363 *

1 0

The const proposal - Ticket 140
by unknown＠example.com 27 May '09

27 May '09

All - The signatories of this letter represent the majority of MPI implementors participating in the MPI Forum. We are concerned that proposal #140 ("Add const Keyword to the C bindings") has a number of issues which suggest delaying to MPI-3 would be appropriate. In particular, the proposal: - Is likely to pass a simple majority vote, but does not carry the support of the majority of MPI implementors, suggesting consensus has not been reached. - Changes 90+ MPI API interfaces, which is not a "trivial" change and therefore does not meet the intent of the MPI-2.2 process. - Is not needed to fix any serious bug in the standard text or to solve an issue that cannot easily be avoided by the MPI application. - Does not offer any demonstrable optimization opportunities for implementation or application, but may constrain future implementation opportunities. Therefore, we ask for your assistance in deferring proposal #140 to the MPI-3 process, where more time can be spent assessing its impact. Thank you, - Cisco: Jeff Squires - Intel: Alexander Supalov & Keith Underwood - Sandia: Brian Barrett - IBM: Richard Treumann - QLogic: Avneesh Pant - UTenn: George Bosilca - HP: David George Solt - UHouston: Edgar Gabriel - Myricom: Patrick Geoffray - ORNL: Richard Graham - Sun: Terry Dontje - NEC: Hubert Ritzdorf & Jesper Traeff Dick Treumann - MPI Team IBM Systems & Technology Group Dept X2ZA / MS P963 -- 2455 South Road -- Poughkeepsie, NY 12601 Tele (845) 433-7846 Fax (845) 433-8363 *

1 0

Re: [Mpi-22] Another MPI-2.2 attribute ambiguity?
by unknown＠example.com 26 May '09

26 May '09

This got even more complex when I talked to Iain on the phone. We're going to move this conversation to the MPI Fortran working group list (mpi3-fortran). In case you didn't catch it before, this is quite definitely an MPI-3 item. On May 22, 2009, at 10:51 AM, Jeff Squyres wrote: > Bill and I chatted on the phone a bit -- this is rapidly evolving > into a complex issue. :-) > > Let me chat with Iain about this a bit more and get back to the > group -- this may turn into a topic of discussion at the next Forum > meeting. > > > On May 22, 2009, at 10:08 AM, William Gropp wrote: > >> Remember, the primary definition of the MPI routine is the language- >> neutral form. That specifies a pointer to some extra state, but does >> says nothing about the size of the data to which that points. I'm >> *sure* that the intention of the language-neutral design was that >> extra state would often be either null or a pointer to a structure >> containing more than one item; the Fortran equivalent to this is an >> array. In Fortran 77, there is no way to specify what in C is a void >> pointer, so the binding says "pass an address-sized integer" (e.g., >> something that could be a pointer) and hope for the best. >> >> The point about expressions being passed also applies to C - if you >> pass, as extra state, an address to data that may go out of scope, >> you >> have a similar problem to passing an expression as extra state in >> Fortran. More specifically, it is already the case that everything >> which is syntactically allowed is not semantically allowed. Just >> because it will compile doesn't make it correct MPI code. >> >> There is certainly ambiguity here. My preference is to start with >> the >> intent of the language neutral specification and make that work. >> That >> may require additional advice to users and implementors about what is >> valid and what is invalid (note that expressions are semantically >> invalid in other places, such as buffers to the non-blocking routines >> in Fortran). >> >> Bill >> >> On May 22, 2009, at 7:39 AM, Jeff Squyres wrote: >> >> > Imagine a Fortran communicator keyval copy function callback like >> this >> > (typed off the top of my head; pardon typos): >> > >> > SUBROUTINE COPY_FUNCTION1(OLDCOMM, KEYVAL, EXTRA_STATE, >> AVALIN, >> > AVALOUT, FLAG, IERR) >> > ... >> > INTEGER EXTRA_STATE >> > INTEGER IERR >> > EXTRA_STATE = EXTRA_STATE + 1 >> > WRITE *, "Extra state is: ", EXTRA_STATE >> > IERR = MPI_SUCCESS >> > END >> > >> > Now suppose MPI_COMM_CREATE_KEYVAL is invoked like this: >> > >> > INTEGER EV = 7 >> > CALL MPI_COMM_KEYVAL_CREATE(COPY_FUNCTION1, ..., EV, IERR) >> > >> > I think we all agree that the output the first time that >> > COPY_FUNCTION1 is invoked should be 7. But what is the output the >> > second time it is invoked -- 7 or 8? That is the question. >> > >> > I think Iain has conclusively show that in this case: >> > >> > INTEGER EV = 7 >> > CALL MPI_COMM_KEYVAL_CREATE(COPY_FUNCTION1, ..., EV + 3, >> IERR) >> > EV = 42 >> > >> > the output from the first call to COPY_FUNCTION1 still needs to >> be 10. >> > >> > The point is that the MPI implementation needs to *copy* the >> INTEGER >> > value into internal storage and use *that* storage when calling >> > COPY_FUNCTION1. If for no other reason than the value may be the >> > result of an expression, and that storage disappears after the >> call to >> > MPI_COMM_KEYVAL_CREATE completes (as in the above example). >> > >> > I'm not sure what you're referring to with the array stuff -- the >> > Fortran prototype for the callback function says that the extra >> value >> > can only be a single INTEGER...? >> > >> > >> > >> > On May 21, 2009, at 4:41 PM, William Gropp wrote: >> > >> >> I'm not sure what "change the value" means in #1. I think that >> you >> >> mean "if we interpret extra_data as a pointer to a Fortran data >> item, >> >> the MPI implementation will pass this pointer (by value) to the >> >> callback. If the callback stores through this pointer, then >> >> (naturally) the afore mentioned data item will be changed. If >> that's >> >> what #1 means, then yes, that's the right answer. Note that the >> >> extra_data could be a pointer to an array of arbitrary size, and >> the >> >> call back could update any part of the array. Saving a copy of >> what >> >> extra_data points at (is that what #2 is doing?) is going to be >> >> tricky, since you don't know the size of the thing to copy >> (without a >> >> lot of help from the compiler in Fortran). >> >> >> >> Bill >> >> >> >> On May 19, 2009, at 7:55 PM, Jeff Squyres wrote: >> >> >> >>> So then you're in favor of #1 from my original post, right? >> >>> >> >>> http://lists.mpi-forum.org/mpi-22/2009/05/0407.php >> >>> >> >>> >> >>> On May 19, 2009, at 4:39 PM, William Gropp wrote: >> >>> >> >>>> Silly me, I missed that. >> >>>> >> >>>> In Fortran (particularly thinking in terms of Fortran 77, since >> >> that >> >>>> was the original design), the Fortran callback should be able >> to do >> >>>> things like increment the "extra data" - i.e., if the extra data >> >> is a >> >>>> counter, the Fortran callback should be able to increment it. >> >> Since >> >>>> Fortran 77 has no pointers (and pointers in Fortran 90 aren't >> the >> >>>> same >> >>>> as addresses), having an address is worthless in a Fortran 77 >> >>>> program. So if the MPI implementation stores the address of the >> >>>> extra >> >>>> data and passes that to the call back (by value, so Fortran sees >> >> the >> >>>> address of the extra data), users should get what they expect in >> >>>> Fortran. >> >>>> >> >>>> Bill >> >>>> >> >>>> On May 19, 2009, at 12:22 PM, Iain Bason wrote: >> >>>> >> >>>>> >> >>>>> On May 19, 2009, at 1:16 PM, William Gropp wrote: >> >>>>> >> >>>>>> MPICH2 passes the value of the extra_state that was saved in >> the >> >>>>>> internal storage to the callback routine, which is case 2 (the >> >> copy >> >>>>>> is implicit as the value is passed to the routine and C >> semantics >> >>>>>> ensures that the value isn't changed). >> >>>>> >> >>>>> But this is the Fortran interface, not the C interface. >> >>>>> >> >>>>> Iain >> >>>>> >> >>>> >> >>>> William Gropp >> >>>> Deputy Director for Research >> >>>> Institute for Advanced Computing Applications and Technologies >> >>>> Paul and Cynthia Saylor Professor of Computer Science >> >>>> University of Illinois Urbana-Champaign >> >>>> >> >>>> >> >>>> >> >>>> >> >>>> _______________________________________________ >> >>>> mpi-22 mailing list >> >>>> mpi-22_at_[hidden] >> >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 >> >>>> >> >>> >> >>> >> >>> -- >> >>> Jeff Squyres >> >>> Cisco Systems >> >>> >> >>> _______________________________________________ >> >>> mpi-22 mailing list >> >>> mpi-22_at_[hidden] >> >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 >> >> >> >> William Gropp >> >> Deputy Director for Research >> >> Institute for Advanced Computing Applications and Technologies >> >> Paul and Cynthia Saylor Professor of Computer Science >> >> University of Illinois Urbana-Champaign >> >> >> >> >> >> >> >> >> >> _______________________________________________ >> >> mpi-22 mailing list >> >> mpi-22_at_[hidden] >> >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 >> >> >> > >> > >> > -- >> > Jeff Squyres >> > Cisco Systems >> > >> > _______________________________________________ >> > mpi-22 mailing list >> > mpi-22_at_[hidden] >> > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 >> >> William Gropp >> Deputy Director for Research >> Institute for Advanced Computing Applications and Technologies >> Paul and Cynthia Saylor Professor of Computer Science >> University of Illinois Urbana-Champaign >> >> >> >> >> _______________________________________________ >> mpi-22 mailing list >> mpi-22_at_[hidden] >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 >> > > > -- > Jeff Squyres > Cisco Systems > > _______________________________________________ > mpi-22 mailing list > mpi-22_at_[hidden] > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 -- Jeff Squyres Cisco Systems

1 0

Re: [Mpi-22] Another MPI-2.2 attribute ambiguity?
by unknown＠example.com 22 May '09

22 May '09

Bill and I chatted on the phone a bit -- this is rapidly evolving into a complex issue. :-) Let me chat with Iain about this a bit more and get back to the group -- this may turn into a topic of discussion at the next Forum meeting. On May 22, 2009, at 10:08 AM, William Gropp wrote: > Remember, the primary definition of the MPI routine is the language- > neutral form. That specifies a pointer to some extra state, but does > says nothing about the size of the data to which that points. I'm > *sure* that the intention of the language-neutral design was that > extra state would often be either null or a pointer to a structure > containing more than one item; the Fortran equivalent to this is an > array. In Fortran 77, there is no way to specify what in C is a void > pointer, so the binding says "pass an address-sized integer" (e.g., > something that could be a pointer) and hope for the best. > > The point about expressions being passed also applies to C - if you > pass, as extra state, an address to data that may go out of scope, you > have a similar problem to passing an expression as extra state in > Fortran. More specifically, it is already the case that everything > which is syntactically allowed is not semantically allowed. Just > because it will compile doesn't make it correct MPI code. > > There is certainly ambiguity here. My preference is to start with the > intent of the language neutral specification and make that work. That > may require additional advice to users and implementors about what is > valid and what is invalid (note that expressions are semantically > invalid in other places, such as buffers to the non-blocking routines > in Fortran). > > Bill > > On May 22, 2009, at 7:39 AM, Jeff Squyres wrote: > > > Imagine a Fortran communicator keyval copy function callback like > this > > (typed off the top of my head; pardon typos): > > > > SUBROUTINE COPY_FUNCTION1(OLDCOMM, KEYVAL, EXTRA_STATE, > AVALIN, > > AVALOUT, FLAG, IERR) > > ... > > INTEGER EXTRA_STATE > > INTEGER IERR > > EXTRA_STATE = EXTRA_STATE + 1 > > WRITE *, "Extra state is: ", EXTRA_STATE > > IERR = MPI_SUCCESS > > END > > > > Now suppose MPI_COMM_CREATE_KEYVAL is invoked like this: > > > > INTEGER EV = 7 > > CALL MPI_COMM_KEYVAL_CREATE(COPY_FUNCTION1, ..., EV, IERR) > > > > I think we all agree that the output the first time that > > COPY_FUNCTION1 is invoked should be 7. But what is the output the > > second time it is invoked -- 7 or 8? That is the question. > > > > I think Iain has conclusively show that in this case: > > > > INTEGER EV = 7 > > CALL MPI_COMM_KEYVAL_CREATE(COPY_FUNCTION1, ..., EV + 3, IERR) > > EV = 42 > > > > the output from the first call to COPY_FUNCTION1 still needs to be > 10. > > > > The point is that the MPI implementation needs to *copy* the INTEGER > > value into internal storage and use *that* storage when calling > > COPY_FUNCTION1. If for no other reason than the value may be the > > result of an expression, and that storage disappears after the > call to > > MPI_COMM_KEYVAL_CREATE completes (as in the above example). > > > > I'm not sure what you're referring to with the array stuff -- the > > Fortran prototype for the callback function says that the extra > value > > can only be a single INTEGER...? > > > > > > > > On May 21, 2009, at 4:41 PM, William Gropp wrote: > > > >> I'm not sure what "change the value" means in #1. I think that you > >> mean "if we interpret extra_data as a pointer to a Fortran data > item, > >> the MPI implementation will pass this pointer (by value) to the > >> callback. If the callback stores through this pointer, then > >> (naturally) the afore mentioned data item will be changed. If > that's > >> what #1 means, then yes, that's the right answer. Note that the > >> extra_data could be a pointer to an array of arbitrary size, and > the > >> call back could update any part of the array. Saving a copy of what > >> extra_data points at (is that what #2 is doing?) is going to be > >> tricky, since you don't know the size of the thing to copy > (without a > >> lot of help from the compiler in Fortran). > >> > >> Bill > >> > >> On May 19, 2009, at 7:55 PM, Jeff Squyres wrote: > >> > >>> So then you're in favor of #1 from my original post, right? > >>> > >>> http://lists.mpi-forum.org/mpi-22/2009/05/0407.php > >>> > >>> > >>> On May 19, 2009, at 4:39 PM, William Gropp wrote: > >>> > >>>> Silly me, I missed that. > >>>> > >>>> In Fortran (particularly thinking in terms of Fortran 77, since > >> that > >>>> was the original design), the Fortran callback should be able > to do > >>>> things like increment the "extra data" - i.e., if the extra data > >> is a > >>>> counter, the Fortran callback should be able to increment it. > >> Since > >>>> Fortran 77 has no pointers (and pointers in Fortran 90 aren't the > >>>> same > >>>> as addresses), having an address is worthless in a Fortran 77 > >>>> program. So if the MPI implementation stores the address of the > >>>> extra > >>>> data and passes that to the call back (by value, so Fortran sees > >> the > >>>> address of the extra data), users should get what they expect in > >>>> Fortran. > >>>> > >>>> Bill > >>>> > >>>> On May 19, 2009, at 12:22 PM, Iain Bason wrote: > >>>> > >>>>> > >>>>> On May 19, 2009, at 1:16 PM, William Gropp wrote: > >>>>> > >>>>>> MPICH2 passes the value of the extra_state that was saved in > the > >>>>>> internal storage to the callback routine, which is case 2 (the > >> copy > >>>>>> is implicit as the value is passed to the routine and C > semantics > >>>>>> ensures that the value isn't changed). > >>>>> > >>>>> But this is the Fortran interface, not the C interface. > >>>>> > >>>>> Iain > >>>>> > >>>> > >>>> William Gropp > >>>> Deputy Director for Research > >>>> Institute for Advanced Computing Applications and Technologies > >>>> Paul and Cynthia Saylor Professor of Computer Science > >>>> University of Illinois Urbana-Champaign > >>>> > >>>> > >>>> > >>>> > >>>> _______________________________________________ > >>>> mpi-22 mailing list > >>>> mpi-22_at_[hidden] > >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 > >>>> > >>> > >>> > >>> -- > >>> Jeff Squyres > >>> Cisco Systems > >>> > >>> _______________________________________________ > >>> mpi-22 mailing list > >>> mpi-22_at_[hidden] > >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 > >> > >> William Gropp > >> Deputy Director for Research > >> Institute for Advanced Computing Applications and Technologies > >> Paul and Cynthia Saylor Professor of Computer Science > >> University of Illinois Urbana-Champaign > >> > >> > >> > >> > >> _______________________________________________ > >> mpi-22 mailing list > >> mpi-22_at_[hidden] > >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 > >> > > > > > > -- > > Jeff Squyres > > Cisco Systems > > > > _______________________________________________ > > mpi-22 mailing list > > mpi-22_at_[hidden] > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 > > William Gropp > Deputy Director for Research > Institute for Advanced Computing Applications and Technologies > Paul and Cynthia Saylor Professor of Computer Science > University of Illinois Urbana-Champaign > > > > > _______________________________________________ > mpi-22 mailing list > mpi-22_at_[hidden] > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 > -- Jeff Squyres Cisco Systems

1 0

Re: [Mpi-22] Another MPI-2.2 attribute ambiguity?
by unknown＠example.com 22 May '09

22 May '09

Remember, the primary definition of the MPI routine is the language- neutral form. That specifies a pointer to some extra state, but does says nothing about the size of the data to which that points. I'm *sure* that the intention of the language-neutral design was that extra state would often be either null or a pointer to a structure containing more than one item; the Fortran equivalent to this is an array. In Fortran 77, there is no way to specify what in C is a void pointer, so the binding says "pass an address-sized integer" (e.g., something that could be a pointer) and hope for the best. The point about expressions being passed also applies to C - if you pass, as extra state, an address to data that may go out of scope, you have a similar problem to passing an expression as extra state in Fortran. More specifically, it is already the case that everything which is syntactically allowed is not semantically allowed. Just because it will compile doesn't make it correct MPI code. There is certainly ambiguity here. My preference is to start with the intent of the language neutral specification and make that work. That may require additional advice to users and implementors about what is valid and what is invalid (note that expressions are semantically invalid in other places, such as buffers to the non-blocking routines in Fortran). Bill On May 22, 2009, at 7:39 AM, Jeff Squyres wrote: > Imagine a Fortran communicator keyval copy function callback like this > (typed off the top of my head; pardon typos): > > SUBROUTINE COPY_FUNCTION1(OLDCOMM, KEYVAL, EXTRA_STATE, AVALIN, > AVALOUT, FLAG, IERR) > ... > INTEGER EXTRA_STATE > INTEGER IERR > EXTRA_STATE = EXTRA_STATE + 1 > WRITE *, "Extra state is: ", EXTRA_STATE > IERR = MPI_SUCCESS > END > > Now suppose MPI_COMM_CREATE_KEYVAL is invoked like this: > > INTEGER EV = 7 > CALL MPI_COMM_KEYVAL_CREATE(COPY_FUNCTION1, ..., EV, IERR) > > I think we all agree that the output the first time that > COPY_FUNCTION1 is invoked should be 7. But what is the output the > second time it is invoked -- 7 or 8? That is the question. > > I think Iain has conclusively show that in this case: > > INTEGER EV = 7 > CALL MPI_COMM_KEYVAL_CREATE(COPY_FUNCTION1, ..., EV + 3, IERR) > EV = 42 > > the output from the first call to COPY_FUNCTION1 still needs to be 10. > > The point is that the MPI implementation needs to *copy* the INTEGER > value into internal storage and use *that* storage when calling > COPY_FUNCTION1. If for no other reason than the value may be the > result of an expression, and that storage disappears after the call to > MPI_COMM_KEYVAL_CREATE completes (as in the above example). > > I'm not sure what you're referring to with the array stuff -- the > Fortran prototype for the callback function says that the extra value > can only be a single INTEGER...? > > > > On May 21, 2009, at 4:41 PM, William Gropp wrote: > >> I'm not sure what "change the value" means in #1. I think that you >> mean "if we interpret extra_data as a pointer to a Fortran data item, >> the MPI implementation will pass this pointer (by value) to the >> callback. If the callback stores through this pointer, then >> (naturally) the afore mentioned data item will be changed. If that's >> what #1 means, then yes, that's the right answer. Note that the >> extra_data could be a pointer to an array of arbitrary size, and the >> call back could update any part of the array. Saving a copy of what >> extra_data points at (is that what #2 is doing?) is going to be >> tricky, since you don't know the size of the thing to copy (without a >> lot of help from the compiler in Fortran). >> >> Bill >> >> On May 19, 2009, at 7:55 PM, Jeff Squyres wrote: >> >>> So then you're in favor of #1 from my original post, right? >>> >>> http://lists.mpi-forum.org/mpi-22/2009/05/0407.php >>> >>> >>> On May 19, 2009, at 4:39 PM, William Gropp wrote: >>> >>>> Silly me, I missed that. >>>> >>>> In Fortran (particularly thinking in terms of Fortran 77, since >> that >>>> was the original design), the Fortran callback should be able to do >>>> things like increment the "extra data" - i.e., if the extra data >> is a >>>> counter, the Fortran callback should be able to increment it. >> Since >>>> Fortran 77 has no pointers (and pointers in Fortran 90 aren't the >>>> same >>>> as addresses), having an address is worthless in a Fortran 77 >>>> program. So if the MPI implementation stores the address of the >>>> extra >>>> data and passes that to the call back (by value, so Fortran sees >> the >>>> address of the extra data), users should get what they expect in >>>> Fortran. >>>> >>>> Bill >>>> >>>> On May 19, 2009, at 12:22 PM, Iain Bason wrote: >>>> >>>>> >>>>> On May 19, 2009, at 1:16 PM, William Gropp wrote: >>>>> >>>>>> MPICH2 passes the value of the extra_state that was saved in the >>>>>> internal storage to the callback routine, which is case 2 (the >> copy >>>>>> is implicit as the value is passed to the routine and C semantics >>>>>> ensures that the value isn't changed). >>>>> >>>>> But this is the Fortran interface, not the C interface. >>>>> >>>>> Iain >>>>> >>>> >>>> William Gropp >>>> Deputy Director for Research >>>> Institute for Advanced Computing Applications and Technologies >>>> Paul and Cynthia Saylor Professor of Computer Science >>>> University of Illinois Urbana-Champaign >>>> >>>> >>>> >>>> >>>> _______________________________________________ >>>> mpi-22 mailing list >>>> mpi-22_at_[hidden] >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 >>>> >>> >>> >>> -- >>> Jeff Squyres >>> Cisco Systems >>> >>> _______________________________________________ >>> mpi-22 mailing list >>> mpi-22_at_[hidden] >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 >> >> William Gropp >> Deputy Director for Research >> Institute for Advanced Computing Applications and Technologies >> Paul and Cynthia Saylor Professor of Computer Science >> University of Illinois Urbana-Champaign >> >> >> >> >> _______________________________________________ >> mpi-22 mailing list >> mpi-22_at_[hidden] >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 >> > > > -- > Jeff Squyres > Cisco Systems > > _______________________________________________ > mpi-22 mailing list > mpi-22_at_[hidden] > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 William Gropp Deputy Director for Research Institute for Advanced Computing Applications and Technologies Paul and Cynthia Saylor Professor of Computer Science University of Illinois Urbana-Champaign

1 0

Re: [Mpi-22] Another MPI-2.2 attribute ambiguity?
by unknown＠example.com 22 May '09

22 May '09

Imagine a Fortran communicator keyval copy function callback like this (typed off the top of my head; pardon typos): SUBROUTINE COPY_FUNCTION1(OLDCOMM, KEYVAL, EXTRA_STATE, AVALIN, AVALOUT, FLAG, IERR) ... INTEGER EXTRA_STATE INTEGER IERR EXTRA_STATE = EXTRA_STATE + 1 WRITE *, "Extra state is: ", EXTRA_STATE IERR = MPI_SUCCESS END Now suppose MPI_COMM_CREATE_KEYVAL is invoked like this: INTEGER EV = 7 CALL MPI_COMM_KEYVAL_CREATE(COPY_FUNCTION1, ..., EV, IERR) I think we all agree that the output the first time that COPY_FUNCTION1 is invoked should be 7. But what is the output the second time it is invoked -- 7 or 8? That is the question. I think Iain has conclusively show that in this case: INTEGER EV = 7 CALL MPI_COMM_KEYVAL_CREATE(COPY_FUNCTION1, ..., EV + 3, IERR) EV = 42 the output from the first call to COPY_FUNCTION1 still needs to be 10. The point is that the MPI implementation needs to *copy* the INTEGER value into internal storage and use *that* storage when calling COPY_FUNCTION1. If for no other reason than the value may be the result of an expression, and that storage disappears after the call to MPI_COMM_KEYVAL_CREATE completes (as in the above example). I'm not sure what you're referring to with the array stuff -- the Fortran prototype for the callback function says that the extra value can only be a single INTEGER...? On May 21, 2009, at 4:41 PM, William Gropp wrote: > I'm not sure what "change the value" means in #1. I think that you > mean "if we interpret extra_data as a pointer to a Fortran data item, > the MPI implementation will pass this pointer (by value) to the > callback. If the callback stores through this pointer, then > (naturally) the afore mentioned data item will be changed. If that's > what #1 means, then yes, that's the right answer. Note that the > extra_data could be a pointer to an array of arbitrary size, and the > call back could update any part of the array. Saving a copy of what > extra_data points at (is that what #2 is doing?) is going to be > tricky, since you don't know the size of the thing to copy (without a > lot of help from the compiler in Fortran). > > Bill > > On May 19, 2009, at 7:55 PM, Jeff Squyres wrote: > > > So then you're in favor of #1 from my original post, right? > > > > http://lists.mpi-forum.org/mpi-22/2009/05/0407.php > > > > > > On May 19, 2009, at 4:39 PM, William Gropp wrote: > > > >> Silly me, I missed that. > >> > >> In Fortran (particularly thinking in terms of Fortran 77, since > that > >> was the original design), the Fortran callback should be able to do > >> things like increment the "extra data" - i.e., if the extra data > is a > >> counter, the Fortran callback should be able to increment it. > Since > >> Fortran 77 has no pointers (and pointers in Fortran 90 aren't the > >> same > >> as addresses), having an address is worthless in a Fortran 77 > >> program. So if the MPI implementation stores the address of the > >> extra > >> data and passes that to the call back (by value, so Fortran sees > the > >> address of the extra data), users should get what they expect in > >> Fortran. > >> > >> Bill > >> > >> On May 19, 2009, at 12:22 PM, Iain Bason wrote: > >> > >>> > >>> On May 19, 2009, at 1:16 PM, William Gropp wrote: > >>> > >>>> MPICH2 passes the value of the extra_state that was saved in the > >>>> internal storage to the callback routine, which is case 2 (the > copy > >>>> is implicit as the value is passed to the routine and C semantics > >>>> ensures that the value isn't changed). > >>> > >>> But this is the Fortran interface, not the C interface. > >>> > >>> Iain > >>> > >> > >> William Gropp > >> Deputy Director for Research > >> Institute for Advanced Computing Applications and Technologies > >> Paul and Cynthia Saylor Professor of Computer Science > >> University of Illinois Urbana-Champaign > >> > >> > >> > >> > >> _______________________________________________ > >> mpi-22 mailing list > >> mpi-22_at_[hidden] > >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 > >> > > > > > > -- > > Jeff Squyres > > Cisco Systems > > > > _______________________________________________ > > mpi-22 mailing list > > mpi-22_at_[hidden] > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 > > William Gropp > Deputy Director for Research > Institute for Advanced Computing Applications and Technologies > Paul and Cynthia Saylor Professor of Computer Science > University of Illinois Urbana-Champaign > > > > > _______________________________________________ > mpi-22 mailing list > mpi-22_at_[hidden] > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 > -- Jeff Squyres Cisco Systems

1 0

Re: [Mpi-22] Another MPI-2.2 attribute ambiguity?
by unknown＠example.com 21 May '09

21 May '09

I'm not sure what "change the value" means in #1. I think that you mean "if we interpret extra_data as a pointer to a Fortran data item, the MPI implementation will pass this pointer (by value) to the callback. If the callback stores through this pointer, then (naturally) the afore mentioned data item will be changed. If that's what #1 means, then yes, that's the right answer. Note that the extra_data could be a pointer to an array of arbitrary size, and the call back could update any part of the array. Saving a copy of what extra_data points at (is that what #2 is doing?) is going to be tricky, since you don't know the size of the thing to copy (without a lot of help from the compiler in Fortran). Bill On May 19, 2009, at 7:55 PM, Jeff Squyres wrote: > So then you're in favor of #1 from my original post, right? > > http://lists.mpi-forum.org/mpi-22/2009/05/0407.php > > > On May 19, 2009, at 4:39 PM, William Gropp wrote: > >> Silly me, I missed that. >> >> In Fortran (particularly thinking in terms of Fortran 77, since that >> was the original design), the Fortran callback should be able to do >> things like increment the "extra data" - i.e., if the extra data is a >> counter, the Fortran callback should be able to increment it. Since >> Fortran 77 has no pointers (and pointers in Fortran 90 aren't the >> same >> as addresses), having an address is worthless in a Fortran 77 >> program. So if the MPI implementation stores the address of the >> extra >> data and passes that to the call back (by value, so Fortran sees the >> address of the extra data), users should get what they expect in >> Fortran. >> >> Bill >> >> On May 19, 2009, at 12:22 PM, Iain Bason wrote: >> >>> >>> On May 19, 2009, at 1:16 PM, William Gropp wrote: >>> >>>> MPICH2 passes the value of the extra_state that was saved in the >>>> internal storage to the callback routine, which is case 2 (the copy >>>> is implicit as the value is passed to the routine and C semantics >>>> ensures that the value isn't changed). >>> >>> But this is the Fortran interface, not the C interface. >>> >>> Iain >>> >> >> William Gropp >> Deputy Director for Research >> Institute for Advanced Computing Applications and Technologies >> Paul and Cynthia Saylor Professor of Computer Science >> University of Illinois Urbana-Champaign >> >> >> >> >> _______________________________________________ >> mpi-22 mailing list >> mpi-22_at_[hidden] >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 >> > > > -- > Jeff Squyres > Cisco Systems > > _______________________________________________ > mpi-22 mailing list > mpi-22_at_[hidden] > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-22 William Gropp Deputy Director for Research Institute for Advanced Computing Applications and Technologies Paul and Cynthia Saylor Professor of Computer Science University of Illinois Urbana-Champaign

1 0

Re: [Mpi-22] Another MPI-2.2 attribute ambiguity?
by unknown＠example.com 20 May '09

20 May '09

On May 20, 2009, at 10:53 AM, Jeff Squyres (jsquyres) wrote: > - Open MPI's upcoming v1.3.3 currently does #1 > - Platform (Scali) MPI does #1 > - HP MPI does #2 > - I *think* MPICH2 is doing #1 (Bill's text can be interpreted 2 ways, > but we already discussed that it's not valid to keep the address of > the value passed through MPI_*_CREATE_KEYVAL) > - Alexander said that Intel MPI does the same thing as Bill's original > reply, but I'm not sure what that means (because Bill's original reply > was about C) > - NEC MPI does #1. -- Jeff Squyres Cisco Systems

1 0