February 2016 - mpi-forum - mailman3.cels.anl.gov

Fwd: Re: [Mpi-comments] Bug in example 7.9 about neighbourhood collectives
by Daniel Holmes 12 Feb '16

12 Feb '16

Hi all, I have created an issue (#34) to correct a bug in example 7.9 and created a pull request (#16) for that change. The issue can be found here: https://github.com/mpi-forum/mpi-issues/issues/34 The pull request can be found here: https://github.com/mpi-forum/mpi-standard/pull/16 The diff can be found here: https://github.com/mpi-forum/mpi-standard/pull/16/files I will bring this forward for a reading/vote/ticket-0-adjudication at the next meeting. Cheers, Dan. -------- Forwarded Message -------- Subject: Re: [Mpi-comments] Bug in example 7.9 about neighbourhood collectives Date: Tue, 26 Jan 2016 08:41:11 -0600 From: William Gropp <wgropp(a)illinois.edu> To: dholmes(a)epcc.ed.ac.uk, Comments on the MPI Standard <mpi-comments(a)lists.mpi-forum.org> Thanks, Dan. Can you open an issue for this? Bill William Gropp Director, Parallel Computing Institute Thomas M. Siebel Chair in Computer Science Chief Scientist, NCSA University of Illinois Urbana-Champaign On Jan 26, 2016, at 7:59 AM, Daniel Holmes <dholmes(a)epcc.ed.ac.uk> wrote: > In Figure 7.2 (part of example 7.9, in section 7.6, page 330, line 18), the call to "MPI_DIMS_CREATE" is erroneous because the first parameter is given as "comm" rather than "comm_size". To pass "comm_size", the example would need an extra variable declaration and an extra line of code (between lines 17 and 18) that calls the "MPI_COMM_SIZE" function. > > Cheers, > Dan. > > -- > Dan Holmes > Applications Consultant in HPC Research > EPCC, The University of Edinburgh > James Clerk Maxwell Building > The Kings Buildings > Peter Guthrie Tait Road > Edinburgh > EH9 3FD > 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. > > _______________________________________________ > mpi-comments mailing list > mpi-comments(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-comments -- Dan Holmes Applications Consultant in HPC Research EPCC, The University of Edinburgh James Clerk Maxwell Building The Kings Buildings Peter Guthrie Tait Road Edinburgh EH9 3FD 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

Reading for deprecation of send cancel (issue #26)
by Daniel Holmes 12 Feb '16

12 Feb '16

Hi all, The point-to-point working group will bring forward the send cancel deprecation issue for a reading in Chicago. The issue (#26) can be found here: https://github.com/mpi-forum/mpi-issues/issues/26 The pull request (#15) can be found here: https://github.com/mpi-forum/mpi-standard/pull/15 A color-coded diff is available from the commit in the point-to-point git repo: https://github.com/mpiwg-p2p/mpi-standard/commit/899223e207ef6aa589b5d2f319… or from the pull request in the mpi-forum git repo: https://github.com/mpi-forum/mpi-standard/pull/15/files Cheers, Dan. -- Dan Holmes Applications Consultant in HPC Research EPCC, The University of Edinburgh James Clerk Maxwell Building The Kings Buildings Peter Guthrie Tait Road Edinburgh EH9 3FD 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

Chicago MPI Forum meeting registration
by Richard Graham 12 Feb '16

12 Feb '16

The registration page is available at https://www.eventbrite.com/e/mpi-forum-tickets-21536496240 Rich

1 0

FTWG Readings in Chicago
by Bland, Wesley 11 Feb '16

11 Feb '16

Hi all, The fault tolerance working group will be presenting a few issues for reading and voting at the Chicago meeting in a few weeks. PDFs are attached to each of the issues below and colorized diffs can be found on each of the pull requests (let me know if you don’t have access). —— The first item will be an update of the reading done in San Jose about the scope of error handlers and MPI_ABORT. We’d like to read some minor updates based on the feedback in the room and do a no “no” vote to amend the previous reading. My understanding of the rules is that this is the correct way to move forward. mpi-forum/mpi-issues#1<https://github.com/mpi-forum/mpi-issues/issues/1> and mpi-forum/mpi-issues#3<https://github.com/mpi-forum/mpi-issues/issues/3> The commit with the updates since the last meeting is here: mpiwg-ft/mpi-standard@5a11f28<https://github.com/mpiwg-ft/mpi-standard/commit/5a11f28e669d0fe7393afe5afd0…> The full pull request is here: mpi-forum/mpi-standard#1<https://github.com/mpi-forum/mpi-standard/pull/1> —— The second item will be a first vote for mpi-forum/mpi-issues#7<https://github.com/mpi-forum/mpi-issues/issues/7> about removing some advice to implementors about MPI_COMM_FREE. —— The third item will be a reading of the issue on catastrophic and non-catastrophic errors discussed at the San Jose meeting. mpi-forum/mpi-issues#28<https://github.com/mpi-forum/mpi-issues/issues/28> The pull request is: mpi-forum/mpi-standard#9<https://github.com/mpi-forum/mpi-standard/pull/9> —— The final item will be a reading of ULFM. We’ll read the full text, but note that ULFM has been split into 3 issues so that we can avoid re-reading the entire text if there is a question about a subsection (such as RMA or I/O). mpi-forum/mpi-issues#20<https://github.com/mpi-forum/mpi-issues/issues/20> mpi-forum/mpi-issues#21<https://github.com/mpi-forum/mpi-issues/issues/21> mpi-forum/mpi-issues#22<https://github.com/mpi-forum/mpi-issues/issues/22> The full pull request is here: mpi-forum/mpi-standard#13<https://github.com/mpi-forum/mpi-standard/pull/13> Thanks, Wesley

1 0

Info Assertions Reading
by Jim Dinan 10 Feb '16

10 Feb '16

Hello, The point-to-point working group will bring forward an updated draft of the info assertions ticket for a reading in Chicago. The pull request is posted here: https://github.com/mpi-forum/mpi-standard/pull/14 And a color-coded PDF copy of the draft is available on the issue page, here: https://github.com/mpi-forum/mpi-issues/issues/11 Best, ~Jim.

1 0