- mpiwg-ft - mailman3.cels.anl.gov

FTWG Call Cancelled Today
by Wesley Bland 02 Sep '19

02 Sep '19

Hi all, Due to the US Holiday, most of us have the day off today. So we will be cancel today’s call. For those of you coming to Zürich, see you there! Thanks, Wesley

1 0

FTWG Con Call Today
by Wesley Bland 19 Aug '19

19 Aug '19

The Fault Tolerance Working Group’s weekly con call is today at 12:00 PM Eastern. Today's agenda: * Look at MPI_INFO_ENV proposal before Thursday deadline * Discuss the virtual meeting presentation for Wednesday on course grained recovery If there's something else that people would like to discuss, please just send an email to the WG so we can get it on the agenda. Thanks, Wesley ......................................................................................................................................... Join from PC, Mac, Linux, iOS or Android: https://tennessee.zoom.us/j/632356722?pwd=lI4_169CGcewIumekTziMw Password: mpiforum Or iPhone one-tap (US Toll): +16468769923,632356722# or +16699006833,632356722# Or Telephone: Dial: +1 646 876 9923 (US Toll) +1 669 900 6833 (US Toll) Meeting ID: 632 356 722 International numbers available: https://zoom.us/u/6uINe Or an H.323/SIP room system: H.323: 162.255.37.11 (US West) or 162.255.36.11 (US East) Meeting ID: 632 356 722 Password: 364216 SIP: 632356722(a)zoomcrc.com Password: 364216 .........................................................................................................................................

1 0

FTWG Con Call - 2019-08-05
by Wesley Bland 05 Aug '19

05 Aug '19

The Fault Tolerance Working Group’s weekly con call is today at 12:00 PM Eastern. Today's agenda: * Using MPI_INFO_ENV before MPI_INIT (https://github.com/mpi-forum/mpi-standard/pull/126 <https://github.com/mpi-forum/mpi-standard/pull/126>) * Updates from fine and course-grained recovery work. If there's something else that people would like to discuss, please just send an email to the WG so we can get it on the agenda. Thanks, Wesley ......................................................................................................................................... Join from PC, Mac, Linux, iOS or Android: https://tennessee.zoom.us/j/632356722?pwd=lI4_169CGcewIumekTziMw Password: mpiforum Or iPhone one-tap (US Toll): +16468769923,632356722# or +16699006833,632356722# Or Telephone: Dial: +1 646 876 9923 (US Toll) +1 669 900 6833 (US Toll) Meeting ID: 632 356 722 International numbers available: https://zoom.us/u/6uINe Or an H.323/SIP room system: H.323: 162.255.37.11 (US West) or 162.255.36.11 (US East) Meeting ID: 632 356 722 Password: 364216 SIP: 632356722(a)zoomcrc.com Password: 364216 .........................................................................................................................................

1 0

Re: [mpiwg-ft] [Mpi-forum] Planning for the September Forum Meeting
by Wesley Bland 29 Jul '19

29 Jul '19

Hey Martin, The FTWG will have three things: 2nd Votes: Define error/failure behavior in MPI_INIT/FINALIZE Issue - https://github.com/mpi-forum/mpi-issues/issues/102 <https://github.com/mpi-forum/mpi-issues/issues/102> PR - https://github.com/mpi-forum/mpi-standard/pull/50 <https://github.com/mpi-forum/mpi-standard/pull/50> Clarify what MPI_SUCCESS guarantees, and what is left undefined when MPI_ERR_XXX happens Issue - https://github.com/mpi-forum/mpi-issues/issues/119 <https://github.com/mpi-forum/mpi-issues/issues/119> PR - https://github.com/mpi-forum/mpi-standard/pull/95 <https://github.com/mpi-forum/mpi-standard/pull/95> New Reading: MPI_GET_INFO_ENV - A new function to construct MPI_INFO_ENV from argv/argc when needed. The text for this one isn’t written yet since we just decided we needed it on the call earlier today, but I’ll work on it and get something ready for the meeting. Should be short. 🤞 Thanks, Wesley > On Jul 23, 2019, at 9:00 AM, Martin Schulz via mpi-forum <mpi-forum(a)lists.mpi-forum.org> wrote: > > Hi all, > > I would like to get an early(er) start on planning the forum meeting in September to make sure we allocate enough time to readings, which are necessary/intended for MPI 4.0. A skeleton agenda is now on the web, but I still have to fill in the details. For this I would like to > > A) ask all WG chairs to let me know if they would like WG time at the Zürich meeting and how long > B) ask all WG chairs and/or other ticket proposers to send me any intentions for readings and votes. > > The latter doesn’t have to be final, yet, and the text version to be read/voted on must still be distributed at most 2 weeks before the meeting, but it would help lay out the schedule. > > Thanks! > > Martin > > PS: As a note to all, please keep in mind that all our schedules for MPI 4.0 are built on the idea of reaching quorum in Zürich - hence, I would strongly encourage everyone to attend (or send a representative). Thanks! > > — > Prof. Dr. Martin Schulz, Chair of Computer Architecture and Parallel Systems > Department of Informatics, TU-Munich, Boltzmannstraße 3, D-85748 Garching > Member of the Board of Directors at the Leibniz Supercomputing Centre (LRZ) > Email: schulzm(a)in.tum.de <mailto:[email protected]> > > > > > > > > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpi-forum

1 0

Re: [mpiwg-ft] Readings / Votes / WG Time in Zurich
by Wesley Bland 29 Jul '19

29 Jul '19

In the Sessions / FT / Tools special call we had this morning to discuss what to do with info objects before MPI_INIT, I volunteered to put together some text for creating an MPI_INFO_ENV object from a new function that takes argv and argc as a parameter (needed when using Fortran, passing NULL to MPI_INIT, or when using Sessions). I’ll add that to this list and send it to Martin. > On Jul 29, 2019, at 11:01 AM, Wesley Bland <work(a)wesbland.com> wrote: > > Hi all, > > I believe we’re not planning to have any new readings led by us in Zurich, is that correct? For our WG related items, I plan to submit: > > 2nd Votes: > > Define error/failure behavior in MPI_INIT/FINALIZE > Issue - https://github.com/mpi-forum/mpi-issues/issues/102 > PR - https://github.com/mpi-forum/mpi-standard/pull/50 > > Clarify what MPI_SUCCESS guarantees, and what is left undefined when MPI_ERR_XXX happens > Issue - https://github.com/mpi-forum/mpi-issues/issues/119 > PR - https://github.com/mpi-forum/mpi-standard/pull/95 > > I don’t think we’re going to have enough attendees to request WG time if I’m correct. > > Thanks, > Wesley

1 0

Readings / Votes / WG Time in Zurich
by Wesley Bland 29 Jul '19

29 Jul '19

Hi all, I believe we’re not planning to have any new readings led by us in Zurich, is that correct? For our WG related items, I plan to submit: 2nd Votes: Define error/failure behavior in MPI_INIT/FINALIZE Issue - https://github.com/mpi-forum/mpi-issues/issues/102 PR - https://github.com/mpi-forum/mpi-standard/pull/50 Clarify what MPI_SUCCESS guarantees, and what is left undefined when MPI_ERR_XXX happens Issue - https://github.com/mpi-forum/mpi-issues/issues/119 PR - https://github.com/mpi-forum/mpi-standard/pull/95 I don’t think we’re going to have enough attendees to request WG time if I’m correct. Thanks, Wesley

1 0

Re: [mpiwg-ft] OpenMP Cleanup
by Laguna Peralta, Ignacio 08 Jul '19

08 Jul '19

Hi Wesley, Thank you - this is very useful. I'll follow up. Ignacio On 7/8/19 11:07 AM, Wesley Bland wrote: > Hey Ignacio, > > I talked to our OpenMP team about the discussion we had on the call this morning. The concern (for those not on the call) was that if an application is using MPI + OpenMP, you could end up inside a parallel region when you need to “reinit” and jump back to some other point in the code, which would be illegal in the case of OpenMP. > > One of the solutions you introduced was an API to specify when to handle the errors. OpenMP has a similar thing in the form of "#pragma omp cancel” and “#pragma omp cancellation point” which might work with that. It might also be possible to construct something more general with cancellation + “omp_pause_resource_all” where you would have to wrap all MPI functions with a cancellation point. > > I’m not sure if this would work or not, but it’s something to explore. > > Thanks, > Wesley >

1 0

OpenMP Cleanup
by Wesley Bland 08 Jul '19

08 Jul '19

Hey Ignacio, I talked to our OpenMP team about the discussion we had on the call this morning. The concern (for those not on the call) was that if an application is using MPI + OpenMP, you could end up inside a parallel region when you need to “reinit” and jump back to some other point in the code, which would be illegal in the case of OpenMP. One of the solutions you introduced was an API to specify when to handle the errors. OpenMP has a similar thing in the form of "#pragma omp cancel” and “#pragma omp cancellation point” which might work with that. It might also be possible to construct something more general with cancellation + “omp_pause_resource_all” where you would have to wrap all MPI functions with a cancellation point. I’m not sure if this would work or not, but it’s something to explore. Thanks, Wesley

1 0

FTWG Con Call Today
by Wesley Bland 08 Jul '19

08 Jul '19

The Fault Tolerance Working Group’s weekly con call is today at 12:00 PM Eastern. Today's agenda: * Progress on fine and course grained recovery models (Aurelien and Ignacio) If there's something else that people would like to discuss, please just send an email to the WG so we can get it on the agenda. Thanks, Wesley ......................................................................................................................................... Join from PC, Mac, Linux, iOS or Android: https://tennessee.zoom.us/j/632356722?pwd=lI4_169CGcewIumekTziMw Password: mpiforum Or iPhone one-tap (US Toll): +16468769923,632356722# or +16699006833,632356722# Or Telephone: Dial: +1 646 876 9923 (US Toll) +1 669 900 6833 (US Toll) Meeting ID: 632 356 722 International numbers available: https://zoom.us/u/6uINe Or an H.323/SIP room system: H.323: 162.255.37.11 (US West) or 162.255.36.11 (US East) Meeting ID: 632 356 722 Password: 364216 SIP: 632356722(a)zoomcrc.com Password: 364216 .........................................................................................................................................

1 0

Re: [mpiwg-ft] help with advice to implementors accompanying MPI_Abort
by HOLMES Daniel 26 Jun '19

26 Jun '19

Hi Aurélien, My concern with changing this text to refer to a different communicator was that the user might have to set an error handler on MPI_COMM_WORLD in addition to the one on that (sub)communicator - to avoid MPI just aborting all MPI processes in MCW at the first sign of trouble. The first sentence says “high quality implementation should” so maybe there is no reason to worry here. We will drop this item from our to-do list. Thanks! Cheers, Dan. — Dr Daniel Holmes PhD Applications Consultant in HPC Research d.holmes(a)epcc.ed.ac.uk<mailto:[email protected]> Phone: +44 (0) 131 651 3465 Mobile: +44 (0) 7940 524 088 Address: Room 2.09, Bayes Centre, 47 Potterrow, Central Area, Edinburgh, EH8 9BT — The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. — On 25 Jun 2019, at 19:33, Aurelien Bouteiller via mpiwg-ft <mpiwg-ft(a)lists.mpi-forum.org<mailto:[email protected]>> wrote: It does not appear to me that any change is necessary. The only part about MPI_COMM_WORLD is in an 'as an example' clause. That being said, substituting MPI_COMM_WORLD with some generic 'communicator comm' would also work. Aurelien On Tue, Jun 25, 2019 at 9:49 AM Pritchard Jr., Howard via mpiwg-ft <mpiwg-ft(a)lists.mpi-forum.org<mailto:[email protected]>> wrote: Hello MPI FTer’s, The Sessions WG could use some help/suggestions about how to adjust the following advice to implementors that accompanies the definition of MPI_Abort: \begin{implementors} After aborting a subset of processes, a high quality implementation should be able to provide error handling for communicators, windows, and files involving both aborted and non-aborted processes. As an example, if the user changes the error handler for \const{MPI\_COMM\_WORLD} to \const{MPI\_ERRORS\_RETURN} or a custom error handler, when a subset of \const{MPI\_COMM\_WORLD} is aborted, the remaining processes in \const{MPI\_COMM\_WORLD} should be able to continue communicating with each other and receive appropriate error codes when attempting communication with an aborted process. \end{implementors} We would like to generalize this advice to implementors to the case where MPI_COMM_WORLD isn’t a valid communicator, i.e. when an application is using the Sessions model. We think that there would need to be some reworking of the existing text to cover the sessions use case. Since the FT group has worked quite a bit on this text, we’d defer to your group for suggestions on how to generalize this text to cover the sessions use case. Thanks very much for any help, Howard -- Howard Pritchard HPC-ENV Los Alamos National Laboratory _______________________________________________ mpiwg-ft mailing list mpiwg-ft(a)lists.mpi-forum.org<mailto:[email protected]> https://lists.mpi-forum.org/mailman/listinfo/mpiwg-ft _______________________________________________ mpiwg-ft mailing list mpiwg-ft(a)lists.mpi-forum.org<mailto:[email protected]> https://lists.mpi-forum.org/mailman/listinfo/mpiwg-ft

1 0