May 2015 - mpi-forum - mailman3.cels.anl.gov

Dates for the MPI Forum Meeting in September 2015 / Bordeaux
by Martin Schulz 20 May '15

20 May '15

Hi all, The EuroMPI conference organizers for this year¹s event are planning some slight changes to the schedule and are likely to shorten the conference by a half day. Due to this, the plan is to shift the MPI Forum dates slightly to make logistics easier the forum would be held all day Thursday and Friday (only two full days). Based on the schedules at the recent EuroMPI based forum meetings, this seems sufficient (especially since I don¹t expect many plenaries due to us having hopefully just finished MPI 3.1) and this would avoid the weekend. If this causes problems for anyone or if people have concerns about this plan, please let me know. Thanks! Martin

1 0

Re: [Mpi-forum] Last call for readings and/or vote requests for the June 2015 MPI meeting
by Rolf Rabenseifner 19 May '15

19 May '15

> I do not find the latest version of the MPI-3.1 draft on the web page. My apologies! I got it, all data is there on the bottom of the agenda page: http://meetings.mpi-forum.org/secretary/2015/06/agenda.php For the chapter authors: ------------------------ I attached an idx file, which contains all location of the terms in your chapter (just check for the page numbers that are within your chapter). Explanation: - Terms within normal text: - TERM ...... without "|uu": cursiv in text & in general index - TERMnoindex without "|uu": cursiv in text & NOT in general index - TERM ...... with "|uu": BOLD in text & UNDERLINED in general index - TERMnoindex with "|uu": BOLD in text & NOT in general index - Terms with section titles: - TERM ...... with "|bold": part of a section title & BOLD number in the general index Note that TERMnoindex is mainly used - if there is already a "|bold" on the same page - if the text of the TERM for the index differs from the higlighted text in the section, e.g., \indexentry{TERMnoindex:Create|uu}{10} \indexentry{TERM:create -- in function names|hyperpage}{10} Note that the intro for the general index on page 816 says: "This index lists mainly terms of the MPI specification. The underlined page numbers refer to the definitions or parts of the definition of the terms. Bold face numbers mark section titles." BOLD in text is used for definitions, cursiv for other locations. Best regards Rolf ----- Original Message ----- > From: "Rolf Rabenseifner" <rabenseifner(a)hlrs.de> > To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> > Sent: Monday, May 18, 2015 8:08:12 PM > Subject: Re: [Mpi-forum] Last call for readings and/or vote requests for the June 2015 MPI meeting > Hi Martin & Bill, hi all, > > I do not find the latest version of the MPI-3.1 draft > on the web page. > > Based on the formatting with \mpiterm, I added a "general index" > at the beginning of all index sections. > > All was done before the 4 weeks deadline, i.e., until May 4. > > On the svn, I find the latest mpi30-31diff.pdf, > but nowhere the mpi31-draft-rev2104.pdf. > > Where is it? > I know, we do not store a draft pdf on the web without a password. > Did I missed an email during my vacation - I'm just back. > > Must I do something for the general index. > > Best regards > Rolf > > ----- Original Message ----- >> From: "schulzm" <schulzm(a)llnl.gov> >> To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> >> Sent: Tuesday, May 19, 2015 2:17:03 AM >> Subject: [Mpi-forum] Last call for readings and/or vote requests for the June >> 2015 MPI meeting > >> Hi all, >> >> We are now only two weeks away from the MPI forum face to face meeting in >> Chicago. Please announce any planned readings or vote requests by today. >> >> Thanks! >> >> Martin >> >> >> _______________________________________________ >> 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) -- 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

Last call for readings and/or vote requests for the June 2015 MPI meeting
by Martin Schulz 18 May '15

18 May '15

Hi all, We are now only two weeks away from the MPI forum face to face meeting in Chicago. Please announce any planned readings or vote requests by today. Thanks! Martin

1 0

Re: [Mpi-forum] Last call for readings and/or vote requests for the June 2015 MPI meeting
by Rolf Rabenseifner 18 May '15

18 May '15

Hi Martin & Bill, hi all, I do not find the latest version of the MPI-3.1 draft on the web page. Based on the formatting with \mpiterm, I added a "general index" at the beginning of all index sections. All was done before the 4 weeks deadline, i.e., until May 4. On the svn, I find the latest mpi30-31diff.pdf, but nowhere the mpi31-draft-rev2104.pdf. Where is it? I know, we do not store a draft pdf on the web without a password. Did I missed an email during my vacation - I'm just back. Must I do something for the general index. Best regards Rolf ----- Original Message ----- > From: "schulzm" <schulzm(a)llnl.gov> > To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> > Sent: Tuesday, May 19, 2015 2:17:03 AM > Subject: [Mpi-forum] Last call for readings and/or vote requests for the June 2015 MPI meeting > Hi all, > > We are now only two weeks away from the MPI forum face to face meeting in > Chicago. Please announce any planned readings or vote requests by today. > > Thanks! > > Martin > > > _______________________________________________ > 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

Ticket 466 Updated - Persistent Nonblocking Collective Communication Proposed for First Reading in Chicago
by Anthony Skjellum 18 May '15

18 May '15

Dear Forum, here is the gist of what we added to Ticket #466, which is proposed for first reading in Chicago: We have added two files - coll-18may15.pdf and topol-18may15.pdf, reflecting the changes needed to complete the proposal and address the major concerns from Jeff Hammond (thank you). Note that we have made additions * 5.1.3 - new section called "Persistent Collective Operations" ; p 74 in coll-18may15.pdf * We fixed text on page 57, lines 27-30, which addressed the issue with non-blocking collective requests being persistent. * we modified Topology Chapter (see topol-18may15.pdf) to contemplate persistent neighborhood communication; we added section 7.8 - Persistent Neighborhood Communication on Process Topologies NOTE: There is another change that we must make, but did not feel comfortable doing once we prototyped the LaTeX, namely, in Section 3.9 of the the current standard, Point-to-Point Chapter, Looking at p73, the current text says "A persistent communication request is created using one of the five following calls. These calls involve no communication." As part of this ticket, we also want to change these sentences to read as follows: "A persistent point-to-point communication request is created using of the five following calls (persistent collective communication requests are discussed in Section X.YZ). Persistent point-to-point communication initialization calls involve no communication. NOTE: This is a cross-reference to new logic, that makes no change in the standard with regard to the existing functionality described there, but specifically differentiates between extant point-to-point requests produced by point-to-point initialization functions and the new collective ones. It could be considered a ticket 0 change, but we didn't like how the LaTeX worked and needed not to check-in the chapter here. Thank you, Tony Skjellum, Puri Bangalore, and Dan Holmes Anthony Skjellum, PhD Professor of Computer Science and Software Engineering COLSA Professor of Cybersecurity and Information Assurance Director of the Auburn Cyber Research Center and Lead Cyber Scientist for Auburn Samuel Ginn College of Engineering Auburn University skjellum(a)auburn.edu or skjellum(a)gmail.com cell: +1-205-807-4968 ; office: +1-334-844-6360 CONFIDENTIALITY: This e-mail and any attachments are confidential and may be privileged. If you are not a named recipient, please notify the sender immediately and do not disclose the contents to another person, use it for any purpose or store or copy the information in any medium.

1 0

Re: [Mpi-forum] Ticket #324 Reading
by Bland, Wesley 18 May '15

18 May '15

Sorry about the bad link on the last email. Here’s the correct link: https://svn.mpi-forum.org/trac/mpi-forum-web/raw-attachment/ticket/324/mpi-… > On May 18, 2015, at 10:02 AM, Bland, Wesley <wesley.bland(a)intel.com> wrote: > > HI all, > > We’ve scheduled ticket #324 (Clarify MPI_ERRORS_ARE_FATAL scope of abort) for a reading at the June meeting in two weeks. It’s a relatively short ticket clarifying that MPI_ERRORS_ARE_FATAL is the same thing as calling MPI_ABORT on a communicator and that calling MPI_ABORT on a communicator should only abort that communicator. I’ll do a short plenary along with the reading to demonstrate our motivating use cases. If you’d like to read the text ahead of time, it hasn’t changed from the version that Josh presented a few years ago and can be found on the ticket as mpi-ticket-324-v2.pdf. The appropriate text is marked up with ticket324 on the PDF. > > https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/324 > https://svn.mpi-forum.org/trac/mpi-forum-web/raw-attachment/ticket/324/mpi-…<https://svn.mpi-forum.org/trac/mpi-forum-web/raw-attachment/ticket/324/mpi-…> > > Thanks, > Wesley > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum

1 0

Ticket #324 Reading
by Bland, Wesley 18 May '15

18 May '15

HI all, We’ve scheduled ticket #324 (Clarify MPI_ERRORS_ARE_FATAL scope of abort) for a reading at the June meeting in two weeks. It’s a relatively short ticket clarifying that MPI_ERRORS_ARE_FATAL is the same thing as calling MPI_ABORT on a communicator and that calling MPI_ABORT on a communicator should only abort that communicator. I’ll do a short plenary along with the reading to demonstrate our motivating use cases. If you’d like to read the text ahead of time, it hasn’t changed from the version that Josh presented a few years ago and can be found on the ticket as mpi-ticket-324-v2.pdf. The appropriate text is marked up with ticket324 on the PDF. https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/324 https://svn.mpi-forum.org/trac/mpi-forum-web/raw-attachment/ticket/324/mpi-…<https://svn.mpi-forum.org/trac/mpi-forum-web/raw-attachment/ticket/324/mpi-…> Thanks, Wesley

1 0

Re: [Mpi-forum] Logistics for the June 1-4 Forum meeting in Chicago
by Jeff Hammond 15 May '15

15 May '15

I just looked for hotels and at least Marriott is running 500+ a night at this point. I'm going to stay in Downers Grove and take the BNSF into the city, which brings the rate down to ~150, which is less than the govt rate for Chicago as far as I know. Anyways, I thought I would share in case anyone was having trouble finding a cost-effective solution. Jeff On Fri, Apr 24, 2015 at 7:43 AM, Richard Graham <richardg(a)mellanox.com> wrote: > The registration page for the meeting is up. It is at: > > https://www.eventbrite.com/e/mpi-forum-tickets-16703722283 > > > > To meet the request for a higher grade of food, the registration charge will > go up to $150/person. > > > > Annelise is working on trying to find a room block, but turns out there is a > big convention in town, and prices are through the roof – similar to last > year. We will see what can be done, but I am not very optimistic. > > > > Rich > > > > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum -- Jeff Hammond jeff.science(a)gmail.com http://jeffhammond.github.io/

1 0

June Meeting Agenda / Documents for MPI 3.1 and New Voting Rules
by Schulz Martin 12 May '15

12 May '15

Hi all, Just FYI all documents, including the final release candidate document for MPI 3.1, larger changes to be discussed, as well as the proposal for new voting rules, are posted together on the MPI Forum agenda page at: http://meetings.mpi-forum.org/secretary/2015/06/agenda.php Please take a look at them and help bring up any issues that you may see rather sooner than later. I also updated the agenda based on the input that I gave gotten if there are any more requests for time slots (WG or plenary), please let me know soon. Also, if any WG would like ticket or errata ballots, please announce those by next Monday. Note, that we have a quite different agenda this time to make sure we allow time for the various ballots (in multiple voting blocks). I may still adjust the WG and plenary times bit to accommodate the requests I get, but the general structure and voting times will stay as they are listed. Thanks! Martin PS: Please register the link is available at: http://meetings.mpi-forum.org/Jun_2015_logistics.php ________________________________________________________________________ Martin Schulz, schulzm(a)llnl.gov, http://scalability.llnl.gov/ CASC @ Lawrence Livermore National Laboratory, Livermore, USA

1 0

(Hopefully) Final copy of ratification procedures rules
by Jeff Squyres (jsquyres) 04 May '15

04 May '15

Based on some good feedback from Intel late last week, here's the final version that we hope to vote on at this upcoming meeting in Chicago (in 1 month!). Here's the changes compared to last week: - Item 9(c) on page 9 (section 2.2.4) is new. - Item 9(d) on page 10 (section 2.2.4) was modified. *** If you plan to vote "no" on this document at the upcoming meeting, Martin and I would greatly appreciate knowing why (ASAP). Thanks! -- Jeff Squyres jsquyres(a)cisco.com For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

1 0