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

Ticket #421 (errata): Formal reading
by Rolf Rabenseifner 19 May '14

19 May '14

MPI Forum -- The RMA and Terms WG will be reading ticket #421 at the upcoming meeting in Chicago: https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/421 it is addendum and minor text change to the series of tickets 349+402+404. Therefore, it will be read and voted as minor text change to the ticket. The changes are needed for the consistency of the MPI document; they do not change the technical content of 349+402+404. Reading will be on Wednesday. On Thursday morning, first voting will be on adding #421 to the series #349+402+404 (which requires zero no-votes). After this, the first vote can be done on #349+402+404+421. If 421 would fail, then 1st vote would only be taken for #349+402+404. Best regards Rolf -- 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

Ticket #403 (errata): Formal reading
by Rolf Rabenseifner 19 May '14

19 May '14

MPI Forum -- The Fortran WG will be reading ticket #403 (errata) at the upcoming meeting in Chicago: https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/403 We missed to write "Fortran Compiler" when we discussed REAL*16 and so on. An additional advice to users explains the subtle differences that may exist between 16 byte floating point in Fortran and in C. The problem was detected, reported, and resolved with an solution by Bill Gropp Best regards Rolf -- 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] [MPI Forum] #420: Bug in neighbourhood collective example
by Rolf Rabenseifner 19 May '14

19 May '14

Dear all, the errata ticket #420 need to be read and voted for the case that there is no chapter committee for virtual topologies. Therefore I added the Forum list to take this as formal announcement. Otherwise, the chapter committee can pass this ticket. At the end of the next meeting, I hope that it is in the status "passed" irrespective of committee or forum vote. Rolf ----- Original Message ----- > From: "Martin Schulz" <schulzm(a)llnl.gov> > To: "Jeff Squyres (jsquyres)" <jsquyres(a)cisco.com>, "Rolf Rabenseifner" <rabenseifner(a)hlrs.de> > Cc: "Daniel Holmes" <dholmes(a)epcc.ed.ac.uk>, "Torsten Hoefler" <htor(a)inf.ethz.ch> > Sent: Sunday, May 18, 2014 8:47:05 AM > Subject: Re: [MPI Forum] #420: Bug in neighbourhood collective example > > According to Bill, any changes to examples are ticket 0 - I need to > push > on getting the chapter commities completed, but then it should go to > them > directly. > > Torsten, are you the chapter chair for this? If not, it would be good > to > get someone to commit to this role. > > Thanks, > > Martin > > > ________________________________________________________________________ > Martin Schulz, schulzm(a)llnl.gov, http://scalability.llnl.gov/ > CASC @ Lawrence Livermore National Laboratory, Livermore, USA > > > > > > > > > On 5/16/14, 11:42 AM, "Jeff Squyres (jsquyres)" <jsquyres(a)cisco.com> > wrote: > > >Daniel -- > > > >This is your ticket, right? You can go ahead and announce this on > >the > >MPI Forum list. > > > >I've just been announcing my/Fortran-related tickets (and one for > >Rolf > >because I thought he had disappeared on vacation already). I'm not > >announcing other tickets; people can generally announce their own > >tickets. :-) > > > > > >On May 16, 2014, at 1:39 PM, Rolf Rabenseifner > ><rabenseifner(a)hlrs.de> > >wrote: > > > >> Jeff and Daniel, > >> > >> for the case that Daniel does not come from UK to US for the > >> meeting, > >> Jeff, please can you also announce the reading of #420 > >> which corrects a small bug in the example on page 333: > >> > >> sndcounts(i) = 100 and rcvcounts(i) = 100 for i=1,2 > >> is substituted by > >> sndcounts(i) = 100 and rcvcounts(i) = 1 for i=1,2 > >> (don't worry about the changes in the syntax) > >> because the datatype is the whole vector with 100 elements > >> and not single MPI_REALs. > >> > >> In principle, this should be passed by the chapter committee, > >> without any Forum vote, but this may be non-existing > >> for Chapter 7 Topolgies. > >> > >> Best regards > >> Rolf > >> > >> ----- Forwarded Message ----- > >> From: "MPI Forum" <mpi-forum(a)lists.mpi-forum.org> > >> Sent: Friday, May 16, 2014 7:29:17 PM > >> Subject: Re: [MPI Forum] #420: Bug in neighbourhood collective > >> example > >> > >> #420: Bug in neighbourhood collective example > >> > >>-------------------------------------+----------------------------------- > >>-- > >> Reporter: dholmes | Owner: > >> dholmes > >> Type: Correction to standard | Status: new > >> (for upcoming release + errata) | Milestone: > >>2014/06/02 > >> Priority: Scheduled | Chicago, USA > >> Version: MPI 3.0 errata | Resolution: > >> Keywords: | Implementation status: > >>Unnecessary > >> > >>-------------------------------------+----------------------------------- > >>-- > >> Changes (by RolfRabenseifner): > >> > >> * priority: Waiting for proposal reviews => Scheduled > >> > >> > >> Comment: > >> > >> Review: Okay and necessary. > >> > >> The major change is "sndcounts(1:2) = 1".[[BR]] > >> All other changes are for better reading and better Fortran > >> typing. > >> > >> Therefore switched to SCHEDULED. > >> > >> -- > >> Ticket URL: > >><https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/420#comment:2> > >> MPI Forum <https://svn.mpi-forum.org/> > >> 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) > > > > > >-- > >Jeff Squyres > >jsquyres(a)cisco.com > >For corporate legal information go to: > >http://www.cisco.com/web/about/doing_business/legal/cri/ > > > > > -- 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

Ticket #420 (ticket 0?): formal reading, if needed
by Daniel Holmes 18 May '14

18 May '14

MPI Forum, There is some debate over whether ticket #420 is a ticket 0 change or an errata. If it is the latter, then it will be read at the June meeting in Chicago. https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/420 It is a bug fix for one of the examples of the new neighbourhood collectives. Cheers, Dan. -- Dan Holmes Applications Consultant in HPC Research EPCC, The University of Edinburgh James Clerk Maxwell Building The Kings Buildings Mayfield Road Edinburgh, UK EH9 3JZ T: +44(0)131 651 3465 E: dholmes(a)epcc.ed.ac.uk *Please consider the environment before printing this email.* The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.

1 0

Ticket #323 (fault tolerance): formal reading
by Aurélien Bouteiller 18 May '14

18 May '14

MPI Forum — The Fault Tolerance WG will be reading ticket #323 (fault tolerance) at the upcoming meeting in Chicago: https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/323 Aurelien -- ~~~ Aurélien Bouteiller, Ph.D. ~~~ ~ Research Scientist @ ICL ~ The University of Tennessee, Innovative Computing Laboratory 1122 Volunteer Blvd, suite 309, Knoxville, TN 37996 tel: +1 (865) 974-9375 fax: +1 (865) 974-8296 https://icl.cs.utk.edu/~bouteill/

1 0

Ticket #424 (errata): Formal reading
by Jeff Squyres (jsquyres) 17 May '14

17 May '14

MPI Forum -- The Fortran WG will be reading ticket #424 (errata) at the upcoming meeting in Chicago: https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/424 It is a trivial fix for a parameter type in MPI_IMPROBE (from INTEGER to LOGICAL). FWIW: both Open MPI and MPICH implemented the parameter type "correctly" (i.e., with LOGICAL, not with the INTEGER type stated in the specification). -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0

Ticket #427 (errata): Formal reading
by Jeff Squyres (jsquyres) 16 May '14

16 May '14

MPI Forum -- The Fortran WG will be reading ticket #427 (errata) at the upcoming meeting in Chicago: https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/427 This ticket is two things: 1. Amends some ticket-0 level changes that were made to the Language Bindings chapter. These nominally don't need Forum approval, but they are mixed in with... 2. Some blanket language that, while they don't change any semantics, are worthy of review by the greater Forum to ensure that they are consistent with the rest of the whole standard (e.g., usage of the term "nonblocking operations" to cover a wide array of things). -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0

Ticket #422 (errata): formal reading
by Rolf Rabenseifner 16 May '14

16 May '14

MPI Forum -- The Terms Chapter WG will be reading ticket #422 (errata) at the upcoming meeting in Chicago: https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/422 It contains errata to correct inconsitencies in the Chapter 2 Terms It would be possible to decide it by the chapter committee (because of ticket-0 quality), but this committee is currently very small, and therefore errata-vote is better. Rolf -- 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

Ticket #415 (errata): formal reading
by Jeff Squyres (jsquyres) 16 May '14

16 May '14

MPI Forum -- The Fortran WG will be reading ticket #415 (errata) at the upcoming meeting in Chicago: https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/415 It is a trivial addition of the word "INTEGER" into the definition of the mpi_f08 MPI_Status type. -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0

Ticket #393 (errata): formal reading
by Jeff Squyres (jsquyres) 16 May '14

16 May '14

MPI Forum -- The Fortran WG will be reading ticket #393 (errata) at the upcoming meeting in Chicago: https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/393 It is a new, short Advice to Implementors clarifying the requirement of providing interceptable routines (since modern Fortran lets you inline). While this is already an implicit requirement of the profiling chapter, we added this non-normative AtoI because the question about Fortran inlining *keeps coming up again and again and again* (i.e., the Fortran "CONTAINS" keyword). So we propose adding this AtoI to be completely explicit about the issue. -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0