- mpiwg-fortran - mailman3.cels.anl.gov

Agenda for meeting today (in ~10 mins)
by Jeff Squyres (jsquyres) 05 Dec '14

05 Dec '14

Agenda items for today: 1. If no further reviews of the MPI-3 Fortran errata, I'll just commit: http://lists.mpi-forum.org/mpiwg-fortran/2014/05/1495.php 2. Deprecate mpif.h (and not mpi module) https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/408 3. Deprecate MPI_SIZEOF (there is no current proposal/ticket about this) 4. Fortran cross-reference correction (ticket 0 / errata) https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/440 5. Fortran MPI_BOTTOM not for initialization or assignment (ticket 0 / errata) https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/443 6. Fortran cross-reference update (ticket-0/errata) https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/446 7. MPI 3.0 Errata: MPI RMA synchronization for MPI shared memory https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/456 -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0

Re: [MPIWG Fortran] Webex meeting
by Jeff Hammond 04 Dec '14

04 Dec '14

Hi Bill64, You should read the MPI-3 RMA chapter, focusing on references to MPI_WIN_ALLOCATE_SHARED. Standardizing interoperability with Fortran images is pretty far out of scope. You can refer to past discussion of interoperability with UPC (it might have all been face-to-face, unfortunately) and the endpoints proposal for further information. When the Fortran community cares enough about coarrays to implement them in an scalable* way on machines other than the high-end Cray systems**, it may become more important for the MPI Forum to worry about coarray interoperability. * Meaning that access to images on other nodes achieves a reasonable fraction of Send-Recv bandwidth. ** Has the situation changed regarding Cray Fortran not being available on the Cray CS series? Best, Jeff On Thu, Dec 4, 2014 at 6:07 AM, Bill Long <longb(a)cray.com> wrote: > Perhaps I’m confused by terminology here, since everywhere else “shared memory” means SMP memory accessible locally through OpenMP or threaded language constructs, whereas memory on different ranks is “distributed memory”, accessed by an SPMD programming model. Since we’re talking about MPI, I’m assuming the interest is accessing memory on different ranks. This is really an issue of integrating MPI with Fortran, which is woefully out of date in the current MPI spec. Fortran has “images” which are basically like ranks (separate instances of the program execution, with the capability of referencing memory that is part of a different image). Fortran already has extensive facilities for synchronizing memory references and definitions on remote images, and rules for standard conforming programs that ensure safe accesses. The main difference with the MPI model is the MPI concept of a “window”, which seems unnecessarily complicated to a Fortran programmer. But I suspect it is somewhat similar to a coarray, in that the processor has to set up mechanisms to easily access coarrays on different images. This is handled automatically for Fortran images by the language - no manual intervention by the user. However, Fortran programmers also have ways to access remote memory that is not a coarray, so the analogy is not exact. The real question seems to be whether SYNC MEMORY statements, for example, force flushes of MPI RMA buffers, or whether a SYNC ALL statement has the effect of an MPI barrier. Since the MPI spec seems to lack any direction on how the library interacts with these parts of Fortran, we (Cray) tell customers to use MPI and regular Fortran remote accesses in separate “phases” of the program, where an MPI phase ends with an MPI barrier, and a Fortran phase ends with a SYNC ALL statement or the end of the program. That way any runtime-dependent clean-up is assured for both runtimes. Of course, one solution would be to implement the entire MPI library in coarray-C++, at which point there is only one runtime, eliminating the need for separate phases. But that seems like a pretty big change (although one with potential performance advantages). I think the real 800-pund gorilla in the room is the lack of a complete specification of how MPI and Fortran interact. That spec needs to be part of the MPI document. > > Cheers, > Bill (the Fortran one) > > > On Dec 4, 2014, at 7:12 AM, William Gropp <wgropp(a)illinois.edu> wrote: > >> This is not my position. My position remains that details of access/update to shared memory, including memory synchronization, *must* be handled correctly within the language. C11/C++11 appears to do this, thereby letting MPI+C or MPI+C++ programmers write standard conforming code with shared memory. Fortran is a little behind here, so users will have to depend on extensions (as they do for *other* parts of the MPI-Fortran interface). Including some memory fences in some of the MPI routines might help some Fortran programmers, but many of the issues with respect to correct, standard conforming programming with Fortran would remain. >> >> Bill >> >> On Dec 4, 2014, at 3:45 AM, Rolf Rabenseifner <rabenseifner(a)hlrs.de> wrote: >> >>> because Bill Gropp would like to remove any Fortran Support >>> by substituting memory barriers (that are currently automatically >>> included into shared memory synchronizations, see MPI-3.0 410:15-19) >>> by C99 memory fences. >> >> _______________________________________________ >> mpiwg-fortran mailing list >> mpiwg-fortran(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran > > Bill Long longb(a)cray.com > Fortran Technical Suport & voice: 651-605-9024 > Bioinformatics Software Development fax: 651-605-9142 > Cray Inc./ Cray Plaza, Suite 210/ 380 Jackson St./ St. Paul, MN 55101 > > > _______________________________________________ > mpiwg-fortran mailing list > mpiwg-fortran(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran -- Jeff Hammond jeff.science(a)gmail.com http://jeffhammond.github.io/

1 0

Re: [MPIWG Fortran] Webex meeting
by Bill Long 04 Dec '14

04 Dec '14

Perhaps I’m confused by terminology here, since everywhere else “shared memory” means SMP memory accessible locally through OpenMP or threaded language constructs, whereas memory on different ranks is “distributed memory”, accessed by an SPMD programming model. Since we’re talking about MPI, I’m assuming the interest is accessing memory on different ranks. This is really an issue of integrating MPI with Fortran, which is woefully out of date in the current MPI spec. Fortran has “images” which are basically like ranks (separate instances of the program execution, with the capability of referencing memory that is part of a different image). Fortran already has extensive facilities for synchronizing memory references and definitions on remote images, and rules for standard conforming programs that ensure safe accesses. The main difference with the MPI model is the MPI concept of a “window”, which seems unnecessarily complicated to a Fortran programmer. But I suspect it is somewhat similar to a coarray, in that the processor has to set up mechanisms to easily access coarrays on different images. This is handled automatically for Fortran images by the language - no manual intervention by the user. However, Fortran programmers also have ways to access remote memory that is not a coarray, so the analogy is not exact. The real question seems to be whether SYNC MEMORY statements, for example, force flushes of MPI RMA buffers, or whether a SYNC ALL statement has the effect of an MPI barrier. Since the MPI spec seems to lack any direction on how the library interacts with these parts of Fortran, we (Cray) tell customers to use MPI and regular Fortran remote accesses in separate “phases” of the program, where an MPI phase ends with an MPI barrier, and a Fortran phase ends with a SYNC ALL statement or the end of the program. That way any runtime-dependent clean-up is assured for both runtimes. Of course, one solution would be to implement the entire MPI library in coarray-C++, at which point there is only one runtime, eliminating the need for separate phases. But that seems like a pretty big change (although one with potential performance advantages). I think the real 800-pund gorilla in the room is the lack of a complete specification of how MPI and Fortran interact. That spec needs to be part of the MPI document. Cheers, Bill (the Fortran one) On Dec 4, 2014, at 7:12 AM, William Gropp <wgropp(a)illinois.edu> wrote: > This is not my position. My position remains that details of access/update to shared memory, including memory synchronization, *must* be handled correctly within the language. C11/C++11 appears to do this, thereby letting MPI+C or MPI+C++ programmers write standard conforming code with shared memory. Fortran is a little behind here, so users will have to depend on extensions (as they do for *other* parts of the MPI-Fortran interface). Including some memory fences in some of the MPI routines might help some Fortran programmers, but many of the issues with respect to correct, standard conforming programming with Fortran would remain. > > Bill > > On Dec 4, 2014, at 3:45 AM, Rolf Rabenseifner <rabenseifner(a)hlrs.de> wrote: > >> because Bill Gropp would like to remove any Fortran Support >> by substituting memory barriers (that are currently automatically >> included into shared memory synchronizations, see MPI-3.0 410:15-19) >> by C99 memory fences. > > _______________________________________________ > mpiwg-fortran mailing list > mpiwg-fortran(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran Bill Long longb(a)cray.com Fortran Technical Suport & voice: 651-605-9024 Bioinformatics Software Development fax: 651-605-9142 Cray Inc./ Cray Plaza, Suite 210/ 380 Jackson St./ St. Paul, MN 55101

1 0

Re: [MPIWG Fortran] Webex meeting
by William Gropp 04 Dec '14

04 Dec '14

This is not my position. My position remains that details of access/update to shared memory, including memory synchronization, *must* be handled correctly within the language. C11/C++11 appears to do this, thereby letting MPI+C or MPI+C++ programmers write standard conforming code with shared memory. Fortran is a little behind here, so users will have to depend on extensions (as they do for *other* parts of the MPI-Fortran interface). Including some memory fences in some of the MPI routines might help some Fortran programmers, but many of the issues with respect to correct, standard conforming programming with Fortran would remain. Bill On Dec 4, 2014, at 3:45 AM, Rolf Rabenseifner <rabenseifner(a)hlrs.de> wrote: > because Bill Gropp would like to remove any Fortran Support > by substituting memory barriers (that are currently automatically > included into shared memory synchronizations, see MPI-3.0 410:15-19) > by C99 memory fences.

1 0

Re: [MPIWG Fortran] Webex meeting
by Rolf Rabenseifner 04 Dec '14

04 Dec '14

Hello all, this is only a reminder that we have a MPI-Fortran-WG webex telcon tomorrow > Friday, December 5, 2014 > 11am US Eastern time Agenda may include #440, #443, #446 but also #456 because Bill Gropp would like to remove any Fortran Support by substituting memory barriers (that are currently automatically included into shared memory synchronizations, see MPI-3.0 410:15-19) by C99 memory fences. This is related #456. Best regards Rolf ----- Original Message ----- > From: "Jeff Squyres (jsquyres)" <jsquyres(a)cisco.com> > To: "MPI Fortran WG" <mpiwg-fortran(a)lists.mpi-forum.org> > Sent: Wednesday, November 12, 2014 12:09:51 AM > Subject: Re: [MPIWG Fortran] Webex meeting > > Sigh. Sorry. > > Friday, December 5, 2014 > 11am US Eastern time > > > On Nov 11, 2014, at 5:31 PM, Bill Long <longb(a)cray.com> wrote: > > > Did you mean Dec 5? Dec 4 (Thursday) was not an option in the doodle. Only > > Monday, Wednesday, and Friday showed up. > > > > Cheers, > > Bill > > > > On Nov 11, 2014, at 4:17 PM, Jeff Squyres (jsquyres) <jsquyres(a)cisco.com> > > wrote: > > > >> Thanks for answering the Doodle. We're setup for 4 Dec, 11am US Eastern > >> (which is 1 business day before the Forum :-) ). > >> > >> ------------------------------------------------------- > >> To join the online meeting (Now from mobile devices!) > >> ------------------------------------------------------- > >> 1. Go to > >> https://cisco.webex.com/ciscosales/j.php?MTID=m81dbf893282eecc809b5fe13aeb1… > >> 2. Enter your name and email address. > >> 3. Enter the meeting password: fortran > >> 4. Click "Join Now". > >> > >> See you then. > >> > >> -- > >> Jeff Squyres > >> jsquyres(a)cisco.com > >> For corporate legal information go to: > >> http://www.cisco.com/web/about/doing_business/legal/cri/ > >> > >> _______________________________________________ > >> mpiwg-fortran mailing list > >> mpiwg-fortran(a)lists.mpi-forum.org > >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran > > > > Bill Long > > longb(a)cray.com > > Fortran Technical Suport & voice: > > 651-605-9024 > > Bioinformatics Software Development fax: 651-605-9142 > > Cray Inc./ Cray Plaza, Suite 210/ 380 Jackson St./ St. Paul, MN 55101 > > > > > > _______________________________________________ > > mpiwg-fortran mailing list > > mpiwg-fortran(a)lists.mpi-forum.org > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran > > > -- > Jeff Squyres > jsquyres(a)cisco.com > For corporate legal information go to: > http://www.cisco.com/web/about/doing_business/legal/cri/ > > _______________________________________________ > mpiwg-fortran mailing list > mpiwg-fortran(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran > -- 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: [MPIWG Fortran] Typos in MPI_WIN_*_FN
by Rolf Rabenseifner 18 Nov '14

18 Nov '14

Dear Junchao Zhang, thank you for your feedback. The Problem was already detected and published in the errata from May 14, 2014, on page 21, lines 19+28: http://www.mpi-forum.org/docs/mpi-3.0/errata-30.pdf The solution is different to your proposal. It is TYPE(MPI_Win) :: oldwin because all callbacks have not any INTENT information. Best regards Rolf ----- Original Message ----- > From: "Junchao Zhang" <jczhang(a)mcs.anl.gov> > To: "MPI-WG Fortran working group" <mpiwg-fortran(a)lists.mpi-forum.org> > Sent: Monday, November 17, 2014 4:09:22 PM > Subject: [MPIWG Fortran] Typos in MPI_WIN_*_FN > > On p.730 of MPI-3.0, line 17 & 25, oldwin should be declared as > > TYPE(MPI_Win), INTENT(IN) :: oldwin > > --Junchao Zhang > > _______________________________________________ > mpiwg-fortran mailing list > mpiwg-fortran(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran -- 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

Typos in MPI_WIN_*_FN
by Junchao Zhang 17 Nov '14

17 Nov '14

On p.730 of MPI-3.0, line 17 & 25, oldwin should be declared as TYPE(MPI_Win), INTENT(IN) :: oldwin --Junchao Zhang

1 0

Re: [MPIWG Fortran] Webex meeting
by Jeff Squyres (jsquyres) 11 Nov '14

11 Nov '14

Sigh. Sorry. Friday, December 5, 2014 11am US Eastern time On Nov 11, 2014, at 5:31 PM, Bill Long <longb(a)cray.com> wrote: > Did you mean Dec 5? Dec 4 (Thursday) was not an option in the doodle. Only Monday, Wednesday, and Friday showed up. > > Cheers, > Bill > > On Nov 11, 2014, at 4:17 PM, Jeff Squyres (jsquyres) <jsquyres(a)cisco.com> wrote: > >> Thanks for answering the Doodle. We're setup for 4 Dec, 11am US Eastern (which is 1 business day before the Forum :-) ). >> >> ------------------------------------------------------- >> To join the online meeting (Now from mobile devices!) >> ------------------------------------------------------- >> 1. Go to https://cisco.webex.com/ciscosales/j.php?MTID=m81dbf893282eecc809b5fe13aeb1… >> 2. Enter your name and email address. >> 3. Enter the meeting password: fortran >> 4. Click "Join Now". >> >> See you then. >> >> -- >> Jeff Squyres >> jsquyres(a)cisco.com >> For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/ >> >> _______________________________________________ >> mpiwg-fortran mailing list >> mpiwg-fortran(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran > > Bill Long longb(a)cray.com > Fortran Technical Suport & voice: 651-605-9024 > Bioinformatics Software Development fax: 651-605-9142 > Cray Inc./ Cray Plaza, Suite 210/ 380 Jackson St./ St. Paul, MN 55101 > > > _______________________________________________ > mpiwg-fortran mailing list > mpiwg-fortran(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0

Re: [MPIWG Fortran] Webex meeting
by Bill Long 11 Nov '14

11 Nov '14

Did you mean Dec 5? Dec 4 (Thursday) was not an option in the doodle. Only Monday, Wednesday, and Friday showed up. Cheers, Bill On Nov 11, 2014, at 4:17 PM, Jeff Squyres (jsquyres) <jsquyres(a)cisco.com> wrote: > Thanks for answering the Doodle. We're setup for 4 Dec, 11am US Eastern (which is 1 business day before the Forum :-) ). > > ------------------------------------------------------- > To join the online meeting (Now from mobile devices!) > ------------------------------------------------------- > 1. Go to https://cisco.webex.com/ciscosales/j.php?MTID=m81dbf893282eecc809b5fe13aeb1… > 2. Enter your name and email address. > 3. Enter the meeting password: fortran > 4. Click "Join Now". > > See you then. > > -- > Jeff Squyres > jsquyres(a)cisco.com > For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/ > > _______________________________________________ > mpiwg-fortran mailing list > mpiwg-fortran(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran Bill Long longb(a)cray.com Fortran Technical Suport & voice: 651-605-9024 Bioinformatics Software Development fax: 651-605-9142 Cray Inc./ Cray Plaza, Suite 210/ 380 Jackson St./ St. Paul, MN 55101

1 0

Webex meeting
by Jeff Squyres (jsquyres) 11 Nov '14

11 Nov '14

Thanks for answering the Doodle. We're setup for 4 Dec, 11am US Eastern (which is 1 business day before the Forum :-) ). ------------------------------------------------------- To join the online meeting (Now from mobile devices!) ------------------------------------------------------- 1. Go to https://cisco.webex.com/ciscosales/j.php?MTID=m81dbf893282eecc809b5fe13aeb1… 2. Enter your name and email address. 3. Enter the meeting password: fortran 4. Click "Join Now". See you then. -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0