May 2013 - mpi-forum - mailman3.cels.anl.gov

Re: [Mpi-forum] [mpich-core] Did we screw up the const bindings in MPI-3?
by William Gropp 23 May '13

23 May '13

IIRC, this did come up during the discussions in the Forum, with the conclusion that since it didn't affect most users, only tool builders, they'd have to change. As Marc-Andre notes, for the C wrappers and const, there is a very low tech solution which is not too bad. Fortran is another story. I'm very concerned about the amount of effort in both implementation and testing that appears to be required. Bill William Gropp Director, Parallel Computing Institute Deputy Director for Research Institute for Advanced Computing Applications and Technologies Thomas M. Siebel Chair in Computer Science University of Illinois Urbana-Champaign On May 23, 2013, at 4:55 AM, Marc-Andre Hermanns wrote: > Hi Pavan, > > sorry, didn't directly answer your last question: > >>> Am I missing something here? Are all tools now required to have an >>> #if MPI_VERSION >= 3 check? Does this mean that all existing tools >>> are now broken for MPI-3 implementations? > > I guess it only affects tools that do use pre-generated wrappers. Tools that parse mpi.h and generate wrappers from what they find there might still work. > > Cheers, > Marc-Andre > -- > Marc-Andre Hermanns > German Research School for > Simulation Sciences GmbH > c/o Laboratory for Parallel Programming > 52062 Aachen | Germany > > Tel +49 241 80 99753 > Fax +49 241 80 6 99753 > Web www.grs-sim.de > > Members: Forschungszentrum Jülich GmbH | RWTH Aachen University > Registered in the commercial register of the local court of > Düren (Amtsgericht Düren) under registration number HRB 5268 > Registered office: Jülich > Executive board: Prof. Marek Behr, Ph.D | Prof. Dr. Sebastian M. Schmidt

1 0

Re: [Mpi-forum] Did we screw up the const bindings in MPI-3?
by Marc-Andre Hermanns 23 May '13

23 May '13

Hi Pavan, sorry, didn't directly answer your last question: >> Am I missing something here? Are all tools now required to have an >> #if MPI_VERSION >= 3 check? Does this mean that all existing tools >> are now broken for MPI-3 implementations? I guess it only affects tools that do use pre-generated wrappers. Tools that parse mpi.h and generate wrappers from what they find there might still work. Cheers, Marc-Andre -- Marc-Andre Hermanns German Research School for Simulation Sciences GmbH c/o Laboratory for Parallel Programming 52062 Aachen | Germany Tel +49 241 80 99753 Fax +49 241 80 6 99753 Web www.grs-sim.de Members: Forschungszentrum Jï¿½lich GmbH | RWTH Aachen University Registered in the commercial register of the local court of Dï¿½ren (Amtsgericht Dï¿½ren) under registration number HRB 5268 Registered office: Jï¿½lich Executive board: Prof. Marek Behr, Ph.D | Prof. Dr. Sebastian M. Schmidt

1 0

Re: [Mpi-forum] Did we screw up the const bindings in MPI-3?
by Marc-Andre Hermanns 23 May '13

23 May '13

Pavan, all, we had a similar reports from Scalasca users on systems with MPICH3. I didn't quite follow the constness discussion in details, but I think modifying existing prototypes may always end up in some forward or backward incompatibility. We solved it for us with a preprocessor macro that is prepended to all relevant buffer parameters. Depending on the MPI version the macro resolves to a 'const' or not. Quite low-tech, I know. ;-) So far that has been working well and it wasn't much work to implement, as most of our wrappers are generated anyway. I am sure TAU will be able to adapt to this quickly as well. Therefore, I wouldn't speak of a complete 'screw-up'. Personally, I am more concerned about adapting to the MPI-3 Fortran changes, because those will probably be more disruptive than this const issue. As far as I understood Jeff's explanations in one of the last tools telcos, the world as we know it for intercepting Fortran has ended and any tools vendor will have to work on their wrappers and their internals. Implementing the 'const-fix' along the way is probably not such a big deal then. Cheers, Marc-Andre On 23.05.13 06:17, Pavan Balaji wrote: > All, > > Some of our users complained that Tau doesn't work with the latest > version of mpich. After some investigation we discovered that this was > an issue with the MPI-3 standard. Specifically, we added const bindings > for many of the MPI functions. The thinking in the Forum was that this > change is backward compatible since applications using MPI_Send don't > have to change at all. However, this seems to be causing problems for > tools that redefine MPI functions. Basically, a tool can no longer have > a function like this: > > int MPI_Send(void *buf, int count, MPI_Datatype datatype, int dest, int > tag, MPI_Comm comm) > { > /* Do magic */ > ret = PMPI_Send(...); > /* Do some more magic */ > > return ret; > } > > The compiler will complain that this doesn't match the prototype in > mpi.h which is: > > int MPI_Send(const void *buf, int count, MPI_Datatype datatype, int > dest, int tag, MPI_Comm comm); > > Am I missing something here? Are all tools now required to have an #if > MPI_VERSION >= 3 check? Does this mean that all existing tools are now > broken for MPI-3 implementations? > > -- Pavan > -- Marc-Andre Hermanns German Research School for Simulation Sciences GmbH c/o Laboratory for Parallel Programming 52062 Aachen | Germany Tel +49 241 80 99753 Fax +49 241 80 6 99753 Web www.grs-sim.de Members: Forschungszentrum Jï¿½lich GmbH | RWTH Aachen University Registered in the commercial register of the local court of Dï¿½ren (Amtsgericht Dï¿½ren) under registration number HRB 5268 Registered office: Jï¿½lich Executive board: Prof. Marek Behr, Ph.D | Prof. Dr. Sebastian M. Schmidt

1 0

Did we screw up the const bindings in MPI-3?
by Pavan Balaji 22 May '13

22 May '13

All, Some of our users complained that Tau doesn't work with the latest version of mpich. After some investigation we discovered that this was an issue with the MPI-3 standard. Specifically, we added const bindings for many of the MPI functions. The thinking in the Forum was that this change is backward compatible since applications using MPI_Send don't have to change at all. However, this seems to be causing problems for tools that redefine MPI functions. Basically, a tool can no longer have a function like this: int MPI_Send(void *buf, int count, MPI_Datatype datatype, int dest, int tag, MPI_Comm comm) { /* Do magic */ ret = PMPI_Send(...); /* Do some more magic */ return ret; } The compiler will complain that this doesn't match the prototype in mpi.h which is: int MPI_Send(const void *buf, int count, MPI_Datatype datatype, int dest, int tag, MPI_Comm comm); Am I missing something here? Are all tools now required to have an #if MPI_VERSION >= 3 check? Does this mean that all existing tools are now broken for MPI-3 implementations? -- Pavan -- Pavan Balaji http://www.mcs.anl.gov/~balaji

1 0

Re: [Mpi-forum] Fwd: Ticket 375: Message Queue side document
by Jeff Squyres (jsquyres) 22 May '13

22 May '13

Cisco has no problem with this, and will be happy to vote "yes" for extending the "MQS" message queue side document's deadline until this Friday. On May 22, 2013, at 10:56 AM, "Schulz, Martin" <schulzm(a)llnl.gov> wrote: > Hi all, > > Please find below the announcement for ticket #375 for a possible reading at the June meeting. Anh tried sending it in time yesterday, but they are apparently some glitches with the email list. Did anybody else try sending something to the list that they later did not see being posted? > > Thanks! > > Martin > > PS: Please see Anh's request below - according to the vote document we passed last time, the forum has the ability to be flexible with time constraints, but only if every voting eligible member institution agrees to this in a formal vote. Therefore, if there are *ANY* concerns, please let Anh and Kathryn (as the tools WG chair) know ahead of time, so they can plan accordingly. > >> >> >> From: Anh Vo >> Sent: Tuesday, May 21, 2013 2:03 PM >> To: Main MPI Forum mailing list (mpi-forum(a)lists.mpi-forum.org) >> Subject: Ticket 375: Message Queue side document >> >> >> >> Hi all, >> >> Currently the MQS is scheduled for official reading on Thursday June 6th >> >> http://meetings.mpi-forum.org/secretary/2013/06/agenda.php >> >> >> >> Iï¿½ve created a ticket (375) and attached the latest PDF: >> >> https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/375 >> >> >> >> However, we would like to ask if itï¿½d be okay if the final document for reading be a few days late (i.e., by Friday noon PST) for the following reasons: >> >> >> >> 1. It's actively being worked on >> >> 2. It's narrow in scope >> >> 3. It's a side document >> >> 4. We're waiting for and incorporating all the feedback, which has taken a while to gather >> >> 5. We're documenting current practice >> >> >> >> Thanks >> >> Anh >> >> >> >> >> >> >> > > ________________________________________________________________________ > 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 -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0

Fwd: Ticket 375: Message Queue side document
by Schulz, Martin 22 May '13

22 May '13

Hi all, Please find below the announcement for ticket #375 for a possible reading at the June meeting. Anh tried sending it in time yesterday, but they are apparently some glitches with the email list. Did anybody else try sending something to the list that they later did not see being posted? Thanks! Martin PS: Please see Anh's request below - according to the vote document we passed last time, the forum has the ability to be flexible with time constraints, but only if every voting eligible member institution agrees to this in a formal vote. Therefore, if there are *ANY* concerns, please let Anh and Kathryn (as the tools WG chair) know ahead of time, so they can plan accordingly. > > > From: Anh Vo > Sent: Tuesday, May 21, 2013 2:03 PM > To: Main MPI Forum mailing list (mpi-forum(a)lists.mpi-forum.org) > Subject: Ticket 375: Message Queue side document > > > > Hi all, > > Currently the MQS is scheduled for official reading on Thursday June 6th > > http://meetings.mpi-forum.org/secretary/2013/06/agenda.php > > > > I’ve created a ticket (375) and attached the latest PDF: > > https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/375 > > > > However, we would like to ask if it’d be okay if the final document for reading be a few days late (i.e., by Friday noon PST) for the following reasons: > > > > 1. It's actively being worked on > > 2. It's narrow in scope > > 3. It's a side document > > 4. We're waiting for and incorporating all the feedback, which has taken a while to gather > > 5. We're documenting current practice > > > > Thanks > > Anh > > > > > > > ________________________________________________________________________ Martin Schulz, schulzm(a)llnl.gov, http://people.llnl.gov/schulzm CASC @ Lawrence Livermore National Laboratory, Livermore, USA

1 0

Re: [Mpi-forum] Last call for agenda items for the June meeting
by Anthony Skjellum 21 May '13

21 May '13

Martin, sending my student this time ... Amin Hassanj .. I will not come . Tony Anthony Skjellum, PhD 205-807-4968 On May 21, 2013, at 8:28 AM, "Schulz, Martin" <schulzm(a)llnl.gov> wrote: > Hi all, > > Just as a reminder, please send me any official items for the agenda of the June forum by 1pm PT today. In particular, please let me know if there are any errata items (or other items that require votes). > > So far, I have only received one such item, which seems rather trivial. If this stays this way, I am thinking about moving this single vote to Thursday afternoon (and skip the usual overnight wait time), since I had requests for this to accommodate travel schedules. However, if there are *ANY* objections against that, please let me know and we'll keep the usual and agreed upon Thursday reading / Friday vote schedule. > > 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

1 0

Re: [Mpi-forum] Last call for agenda items for the June meeting
by Kathryn Mohror 21 May '13

21 May '13

Hi Rolf, Thanks for letting me know about that. I assigned it to myself. Kathryn On May 21, 2013, at 7:56 AM, Rolf Rabenseifner <rabenseifner(a)hlrs.de> wrote: > Kathryn and Martin, > > one of you should own this ticket (it was only reported by Bill) > --> ACTION "assign to" ..... > > You should set the "Type" to "Correction to standard (for upcoming release + errata)", > otherwise users can use it in Fortran until MPI-3.1 and will detect > non-backward-compatible change with the next release. > > Best regards > Rolf > > ----- Original Message ----- >> From: "Pavan Balaji" <balaji(a)mcs.anl.gov> >> To: "Martin Schulz" <schulzm(a)llnl.gov>, mpi-forum(a)lists.mpi-forum.org >> Sent: Tuesday, May 21, 2013 3:34:32 PM >> Subject: Re: [Mpi-forum] Last call for agenda items for the June meeting >> Martin, >> >> Can you confirm that this is the only ticket that is being voted on at >> this meeting? >> >> https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/354 >> >> -- Pavan >> >> On 05/21/2013 08:28 AM US Central Time, Schulz, Martin wrote: >>> Hi all, >>> >>> Just as a reminder, please send me any official items for the agenda >>> of the June forum by 1pm PT today. In particular, please let me know >>> if there are any errata items (or other items that require votes). >>> >>> So far, I have only received one such item, which seems rather >>> trivial. If this stays this way, I am thinking about moving this >>> single vote to Thursday afternoon (and skip the usual overnight wait >>> time), since I had requests for this to accommodate travel >>> schedules. However, if there are *ANY* objections against that, >>> please let me know and we'll keep the usual and agreed upon Thursday >>> reading / Friday vote schedule. >>> >>> 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 >> _______________________________________________ >> 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) ______________________________________________________________ Kathryn Mohror, kathryn(a)llnl.gov, http://people.llnl.gov/mohror1 CASC @ Lawrence Livermore National Laboratory, Livermore, CA, USA

1 0

Re: [Mpi-forum] Last call for agenda items for the June meeting
by Rolf Rabenseifner 21 May '13

21 May '13

Kathryn and Martin, one of you should own this ticket (it was only reported by Bill) --> ACTION "assign to" ..... You should set the "Type" to "Correction to standard (for upcoming release + errata)", otherwise users can use it in Fortran until MPI-3.1 and will detect non-backward-compatible change with the next release. Best regards Rolf ----- Original Message ----- > From: "Pavan Balaji" <balaji(a)mcs.anl.gov> > To: "Martin Schulz" <schulzm(a)llnl.gov>, mpi-forum(a)lists.mpi-forum.org > Sent: Tuesday, May 21, 2013 3:34:32 PM > Subject: Re: [Mpi-forum] Last call for agenda items for the June meeting > Martin, > > Can you confirm that this is the only ticket that is being voted on at > this meeting? > > https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/354 > > -- Pavan > > On 05/21/2013 08:28 AM US Central Time, Schulz, Martin wrote: > > Hi all, > > > > Just as a reminder, please send me any official items for the agenda > > of the June forum by 1pm PT today. In particular, please let me know > > if there are any errata items (or other items that require votes). > > > > So far, I have only received one such item, which seems rather > > trivial. If this stays this way, I am thinking about moving this > > single vote to Thursday afternoon (and skip the usual overnight wait > > time), since I had requests for this to accommodate travel > > schedules. However, if there are *ANY* objections against that, > > please let me know and we'll keep the usual and agreed upon Thursday > > reading / Friday vote schedule. > > > > 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 > _______________________________________________ > 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)

1 0

Re: [Mpi-forum] Last call for agenda items for the June meeting
by Pavan Balaji 21 May '13

21 May '13

Martin, Can you confirm that this is the only ticket that is being voted on at this meeting? https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/354 -- Pavan On 05/21/2013 08:28 AM US Central Time, Schulz, Martin wrote: > Hi all, > > Just as a reminder, please send me any official items for the agenda of the June forum by 1pm PT today. In particular, please let me know if there are any errata items (or other items that require votes). > > So far, I have only received one such item, which seems rather trivial. If this stays this way, I am thinking about moving this single vote to Thursday afternoon (and skip the usual overnight wait time), since I had requests for this to accommodate travel schedules. However, if there are *ANY* objections against that, please let me know and we'll keep the usual and agreed upon Thursday reading / Friday vote schedule. > > 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