June 2015 - mpiwg-p2p - mailman3.cels.anl.gov

Re: [mpiwg-p2p] Minor correction in MPI_WAIT
by William Gropp 29 Jun '15

29 Jun '15

That ticket should never have been left marked as complete (that might be my fault for not checking the status after comment 18). The change never made it into the errata for 30, and commit 1822 was made into mpi-3.0-minutiae/chap-pt2pt/pt2pt.tex , which doesn’t count as getting it into the standard. Bill William Gropp Director, Parallel Computing Institute Thomas M. Siebel Chair in Computer Science Chief Scientist, NCSA University of Illinois Urbana-Champaign On Jun 29, 2015, at 3:20 PM, Thakur, Rajeev <thakur(a)mcs.anl.gov> wrote: > OK, I found the reference. Comment 18 in this ticket seems to indicate that the bug was fixed in the source, but probably not in the right place. > > https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/321#comment:18 > > Rajeev > > On Jun 29, 2015, at 7:47 AM, William Gropp <wgropp(a)illinois.edu> > wrote: > >> I added this as a pending errata to the errata-31.tex file - This may help ensure we catch this for the update. >> >> Bill >> >> William Gropp >> Director, Parallel Computing Institute >> Thomas M. Siebel Chair in Computer Science >> Chief Scientist, NCSA >> University of Illinois Urbana-Champaign >> >> >> >> >> >> On Jun 29, 2015, at 2:28 PM, Thakur, Rajeev <thakur(a)mcs.anl.gov> wrote: >> >>> This typo was introduced in MPI 3.0. I seem to remember that someone had reported it earlier, but I can't find the reference. >>> >>> Rajeev >>> >>> On Jun 26, 2015, at 1:54 PM, "Balaji, Pavan" <balaji(a)anl.gov> wrote: >>> >>>> >>>> FYI. >>>> >>>> >>>> >>>> >>>> >>>> On 6/26/15, 1:46 PM, "Balaji, Pavan" <balaji(a)anl.gov> wrote: >>>> >>>>> >>>>> MPI-3.1 standard, page 53, lines 24--25: >>>>> >>>>> "Any other type of request is and the request handle is set to MPI_REQUEST_NULL. MPI_WAIT is a non-local operation." >>>>> >>>>> >>>>> Should be: >>>>> >>>>> >>>>> "Any other type of request is deallocated and the request handle is set to MPI_REQUEST_NULL. MPI_WAIT is a non-local operation." >>>>> >>>>> -- Pavan >>>>> >>>> _______________________________________________ >>>> mpiwg-p2p mailing list >>>> mpiwg-p2p(a)lists.mpi-forum.org >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p >>> >>> _______________________________________________ >>> mpiwg-p2p mailing list >>> mpiwg-p2p(a)lists.mpi-forum.org >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p >> >> _______________________________________________ >> mpiwg-p2p mailing list >> mpiwg-p2p(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p > > _______________________________________________ > mpiwg-p2p mailing list > mpiwg-p2p(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p

1 0

Re: [mpiwg-p2p] Minor correction in MPI_WAIT
by Thakur, Rajeev 29 Jun '15

29 Jun '15

OK, I found the reference. Comment 18 in this ticket seems to indicate that the bug was fixed in the source, but probably not in the right place. https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/321#comment:18 Rajeev On Jun 29, 2015, at 7:47 AM, William Gropp <wgropp(a)illinois.edu> wrote: > I added this as a pending errata to the errata-31.tex file - This may help ensure we catch this for the update. > > Bill > > William Gropp > Director, Parallel Computing Institute > Thomas M. Siebel Chair in Computer Science > Chief Scientist, NCSA > University of Illinois Urbana-Champaign > > > > > > On Jun 29, 2015, at 2:28 PM, Thakur, Rajeev <thakur(a)mcs.anl.gov> wrote: > >> This typo was introduced in MPI 3.0. I seem to remember that someone had reported it earlier, but I can't find the reference. >> >> Rajeev >> >> On Jun 26, 2015, at 1:54 PM, "Balaji, Pavan" <balaji(a)anl.gov> wrote: >> >>> >>> FYI. >>> >>> >>> >>> >>> >>> On 6/26/15, 1:46 PM, "Balaji, Pavan" <balaji(a)anl.gov> wrote: >>> >>>> >>>> MPI-3.1 standard, page 53, lines 24--25: >>>> >>>> "Any other type of request is and the request handle is set to MPI_REQUEST_NULL. MPI_WAIT is a non-local operation." >>>> >>>> >>>> Should be: >>>> >>>> >>>> "Any other type of request is deallocated and the request handle is set to MPI_REQUEST_NULL. MPI_WAIT is a non-local operation." >>>> >>>> -- Pavan >>>> >>> _______________________________________________ >>> mpiwg-p2p mailing list >>> mpiwg-p2p(a)lists.mpi-forum.org >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p >> >> _______________________________________________ >> mpiwg-p2p mailing list >> mpiwg-p2p(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p > > _______________________________________________ > mpiwg-p2p mailing list > mpiwg-p2p(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p

1 0

Re: [mpiwg-p2p] Minor correction in MPI_WAIT
by William Gropp 29 Jun '15

29 Jun '15

I added this as a pending errata to the errata-31.tex file - This may help ensure we catch this for the update. Bill William Gropp Director, Parallel Computing Institute Thomas M. Siebel Chair in Computer Science Chief Scientist, NCSA University of Illinois Urbana-Champaign On Jun 29, 2015, at 2:28 PM, Thakur, Rajeev <thakur(a)mcs.anl.gov> wrote: > This typo was introduced in MPI 3.0. I seem to remember that someone had reported it earlier, but I can't find the reference. > > Rajeev > > On Jun 26, 2015, at 1:54 PM, "Balaji, Pavan" <balaji(a)anl.gov> wrote: > >> >> FYI. >> >> >> >> >> >> On 6/26/15, 1:46 PM, "Balaji, Pavan" <balaji(a)anl.gov> wrote: >> >>> >>> MPI-3.1 standard, page 53, lines 24--25: >>> >>> "Any other type of request is and the request handle is set to MPI_REQUEST_NULL. MPI_WAIT is a non-local operation." >>> >>> >>> Should be: >>> >>> >>> "Any other type of request is deallocated and the request handle is set to MPI_REQUEST_NULL. MPI_WAIT is a non-local operation." >>> >>> -- Pavan >>> >> _______________________________________________ >> mpiwg-p2p mailing list >> mpiwg-p2p(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p > > _______________________________________________ > mpiwg-p2p mailing list > mpiwg-p2p(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p

1 0

Re: [mpiwg-p2p] Minor correction in MPI_WAIT
by Thakur, Rajeev 29 Jun '15

29 Jun '15

This typo was introduced in MPI 3.0. I seem to remember that someone had reported it earlier, but I can't find the reference. Rajeev On Jun 26, 2015, at 1:54 PM, "Balaji, Pavan" <balaji(a)anl.gov> wrote: > > FYI. > > > > > > On 6/26/15, 1:46 PM, "Balaji, Pavan" <balaji(a)anl.gov> wrote: > >> >> MPI-3.1 standard, page 53, lines 24--25: >> >> "Any other type of request is and the request handle is set to MPI_REQUEST_NULL. MPI_WAIT is a non-local operation." >> >> >> Should be: >> >> >> "Any other type of request is deallocated and the request handle is set to MPI_REQUEST_NULL. MPI_WAIT is a non-local operation." >> >> -- Pavan >> > _______________________________________________ > mpiwg-p2p mailing list > mpiwg-p2p(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p

1 0

FW: Minor correction in MPI_WAIT
by Balaji, Pavan 26 Jun '15

26 Jun '15

FYI. On 6/26/15, 1:46 PM, "Balaji, Pavan" <balaji(a)anl.gov> wrote: > >MPI-3.1 standard, page 53, lines 24--25: > >"Any other type of request is and the request handle is set to MPI_REQUEST_NULL. MPI_WAIT is a non-local operation." > > >Should be: > > >"Any other type of request is deallocated and the request handle is set to MPI_REQUEST_NULL. MPI_WAIT is a non-local operation." > > -- Pavan >

1 0

Re: [mpiwg-p2p] Next meeting: 22nd June 2015 11am Central US
by Balaji, Pavan 22 Jun '15

22 Jun '15

I'll be at the SC PC meeting. Can someone cover send cancel? -- Pavan Sent from my iPhone > On Jun 22, 2015, at 10:17 AM, Jeff Hammond <jeff.science(a)gmail.com> wrote: > > There will be no update on ARECV+FSEND. No progress has been made since April. > > Jeff > >> On Mon, Jun 22, 2015 at 3:15 AM, Daniel Holmes <dholmes(a)epcc.ed.ac.uk> wrote: >> Hi all, >> >> The next point-to-point teleconference will be today on Monday 22th June >> 2015 at 11am Central US via webex. >> >> Connection details are on the point-to-point wiki page: >> https://svn.mpi-forum.org/trac/mpi-forum-web/wiki/PtpWikiPage >> >> Agenda: >> 1) Ticket #476: "Can INFO keys change the semantic behaviour of MPI?" issue >> - the communicator assertions ticket was presented at the June Forum meeting >> and we got some useful feedback that we should discuss. >> >> 2) Tickets #478, #479, #480: "everything you ever wanted to know about MPI >> send cancel, plus a whole lot more" [new!] >> >> 3) New communicator INFO keys - it looks like these will be allowed, so we >> should define a few and get a proposal ready for the next face-to-face >> meeting >> >> 4) Progress on Arecv and Fsend? >> >> 5) Other stuff? >> >> 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. >> >> _______________________________________________ >> mpiwg-p2p mailing list >> mpiwg-p2p(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p > > > > -- > Jeff Hammond > jeff.science(a)gmail.com > http://jeffhammond.github.io/ > _______________________________________________ > mpiwg-p2p mailing list > mpiwg-p2p(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p

1 0

Re: [mpiwg-p2p] Next meeting: 22nd June 2015 11am Central US
by Jeff Hammond 22 Jun '15

22 Jun '15

There will be no update on ARECV+FSEND. No progress has been made since April. Jeff On Mon, Jun 22, 2015 at 3:15 AM, Daniel Holmes <dholmes(a)epcc.ed.ac.uk> wrote: > Hi all, > > The next point-to-point teleconference will be today on Monday 22th June > 2015 at 11am Central US via webex. > > Connection details are on the point-to-point wiki page: > https://svn.mpi-forum.org/trac/mpi-forum-web/wiki/PtpWikiPage > > Agenda: > 1) Ticket #476: "Can INFO keys change the semantic behaviour of MPI?" issue > - the communicator assertions ticket was presented at the June Forum meeting > and we got some useful feedback that we should discuss. > > 2) Tickets #478, #479, #480: "everything you ever wanted to know about MPI > send cancel, plus a whole lot more" [new!] > > 3) New communicator INFO keys - it looks like these will be allowed, so we > should define a few and get a proposal ready for the next face-to-face > meeting > > 4) Progress on Arecv and Fsend? > > 5) Other stuff? > > 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. > > _______________________________________________ > mpiwg-p2p mailing list > mpiwg-p2p(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p -- Jeff Hammond jeff.science(a)gmail.com http://jeffhammond.github.io/

1 0

Next meeting: 22nd June 2015 11am Central US
by Daniel Holmes 22 Jun '15

22 Jun '15

Hi all, The next point-to-point teleconference will be today on Monday 22th June 2015 at 11am Central US via webex. Connection details are on the point-to-point wiki page: https://svn.mpi-forum.org/trac/mpi-forum-web/wiki/PtpWikiPage Agenda: 1) Ticket #476: "Can INFO keys change the semantic behaviour of MPI?" issue - the communicator assertions ticket was presented at the June Forum meeting and we got some useful feedback that we should discuss. 2) Tickets #478, #479, #480: "everything you ever wanted to know about MPI send cancel, plus a whole lot more" [new!] 3) New communicator INFO keys - it looks like these will be allowed, so we should define a few and get a proposal ready for the next face-to-face meeting 4) Progress on Arecv and Fsend? 5) Other stuff? 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

Re: [mpiwg-p2p] Next meeting: 8th June 2015 11am Central US
by Balaji, Pavan 08 Jun '15

08 Jun '15

Hi Dan, Unfortunately, I'm on travel and won't be able to attend this call. Can someone else help discuss these tickets? For #479 and #480, they are both added for completeness. It's not necessary that they both be voted in. The WG could, for example, recommend that we drop one and only pursue the other. #479 is more of a "best effort" proposal for send cancel (basically the same state as we are in today with MPI implementations). #480 simply drops that ability, so there's no more send cancel. Thanks, -- Pavan -- Pavan Balaji http://www.mcs.anl.gov/~balaji From: Daniel Holmes Organization: EPCC Reply-To: "dholmes(a)epcc.ed.ac.uk<mailto:[email protected]>" Date: Monday, June 8, 2015 at 4:13 AM To: MPI Point To Point WG, Pavan Balaji Subject: Next meeting: 8th June 2015 11am Central US Hi all, The next point-to-point teleconference will be today on Monday 8th June 2015 at 11am Central US via webex. Connection details are on the point-to-point wiki page: https://svn.mpi-forum.org/trac/mpi-forum-web/wiki/PtpWikiPage Agenda: 1) Ticket #476: "Can INFO keys change the semantic behaviour of MPI?" issue - the communicator assertions ticket was presented at the June Forum meeting and we got some useful feedback that we should discuss. 2) Tickets #478, #479, #480: "everything you ever wanted to know about MPI send cancel, plus a whole lot more" [new!] 3) New communicator INFO keys - it looks like these will be allowed, so we should define a few and get a proposal ready for the next face-to-face meeting 4) Progress on Arecv and Fsend or ReceiveReduce? 5) Other stuff? Cheers, Dan. -------- Forwarded Message -------- Subject: MPI Pt2pt WG tickets Date: Sun, 7 Jun 2015 21:04:50 +0000 From: Balaji, Pavan <balaji(a)anl.gov><mailto:[email protected]> To: Daniel Holmes <dholmes(a)epcc.ed.ac.uk><mailto:[email protected]> Hi Dan, I'd like to bring the following tickets forward to be vetted by the pt2pt WG before taking it to the Forum. Can you schedule some time for these in the WG telecon? When is the next telecon? https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/478 https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/479 https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/480 Thanks, -- Pavan -- Pavan Balaji http://www.mcs.anl.gov/~balaji -- 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<mailto:[email protected]> *Please consider the environment before printing this email.*

1 0

Re: [mpiwg-p2p] Next meeting: 8th June 2015 11am Central US
by Jeff Hammond 08 Jun '15

08 Jun '15

I'm very sorry to have to miss this call. At this point, please treat Recv_reduce as retired. We debated it within Intel and it was not viewed favorably enough to proceed. Arecv-Fsend is still active but there is nothing to report right now. It's possible that we will have something to share in Bordeaux, else I should have enough for a plenary in December. Jeff On Mon, Jun 8, 2015 at 2:13 AM, Daniel Holmes <dholmes(a)epcc.ed.ac.uk> wrote: > Hi all, > > The next point-to-point teleconference will be today on Monday 8th June > 2015 at 11am Central US via webex. > > Connection details are on the point-to-point wiki page: > https://svn.mpi-forum.org/trac/mpi-forum-web/wiki/PtpWikiPage > > Agenda: > 1) Ticket #476: "Can INFO keys change the semantic behaviour of MPI?" > issue - the communicator assertions ticket was presented at the June Forum > meeting and we got some useful feedback that we should discuss. > > 2) Tickets #478, #479, #480: "everything you ever wanted to know about MPI > send cancel, plus a whole lot more" [new!] > > 3) New communicator INFO keys - it looks like these will be allowed, so we > should define a few and get a proposal ready for the next face-to-face > meeting > > 4) Progress on Arecv and Fsend or ReceiveReduce? > > 5) Other stuff? > > Cheers, > Dan. > > -------- Forwarded Message -------- Subject: MPI Pt2pt WG tickets Date: Sun, > 7 Jun 2015 21:04:50 +0000 From: Balaji, Pavan <balaji(a)anl.gov> > <balaji(a)anl.gov> To: Daniel Holmes <dholmes(a)epcc.ed.ac.uk> > <dholmes(a)epcc.ed.ac.uk> > > Hi Dan, > > I'd like to bring the following tickets forward to be vetted by the > pt2pt WG before taking it to the Forum. Can you schedule some time for > these in the WG telecon? When is the next telecon? > > https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/478 > https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/479 > https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/480 > > Thanks, > > -- Pavan > > -- > Pavan Balaji > http://www.mcs.anl.gov/~balaji > > -- > 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. > > _______________________________________________ > mpiwg-p2p mailing list > mpiwg-p2p(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p > -- Jeff Hammond jeff.science(a)gmail.com http://jeffhammond.github.io/

1 0