June 2013 - mpiwg-ft - mailman3.cels.anl.gov

Re: [Mpi3-ft] Conference Call June 25
by Wesley Bland 26 Jun '13

26 Jun '13

Sorry about the late notice this time. We usually try to have all that information available earlier in the day or the day before. I'll make sure it's earlier next time. Wesley On Jun 26, 2013, at 10:48 AM, Jim Dinan <james.dinan(a)gmail.com> wrote: > Hi Wesley, > > Could you guys send out these reminders a little further in advance? It would also be helpful to include a meeting agenda. > > Thanks, > ~Jim. > > > On Tue, Jun 25, 2013 at 1:59 PM, Wesley Bland <wbland(a)mcs.anl.gov> wrote: > Sorry. I should have said, 2:00 Central, not Eastern. > > On Jun 25, 2013, at 1:58 PM, Wesley Bland <wbland(a)mcs.anl.gov> wrote: > > > Just a reminder that we're having our usual call today at 2:00. > > > > Call in information: > > > > Dial-in information: 218-339-4600 > > Code: 623998# > > > > Wesley > > _______________________________________________ > mpi3-ft mailing list > mpi3-ft(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft > > _______________________________________________ > mpi3-ft mailing list > mpi3-ft(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft

1 0

Re: [Mpi3-ft] Conference Call June 25
by Jim Dinan 26 Jun '13

26 Jun '13

Hi Wesley, Could you guys send out these reminders a little further in advance? It would also be helpful to include a meeting agenda. Thanks, ~Jim. On Tue, Jun 25, 2013 at 1:59 PM, Wesley Bland <wbland(a)mcs.anl.gov> wrote: > Sorry. I should have said, 2:00 Central, not Eastern. > > On Jun 25, 2013, at 1:58 PM, Wesley Bland <wbland(a)mcs.anl.gov> wrote: > > > Just a reminder that we're having our usual call today at 2:00. > > > > Call in information: > > > > Dial-in information: 218-339-4600 > > Code: 623998# > > > > Wesley > > _______________________________________________ > mpi3-ft mailing list > mpi3-ft(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft >

1 0

Re: [Mpi3-ft] Conference Call June 25
by Wesley Bland 25 Jun '13

25 Jun '13

Sorry. I should have said, 2:00 Central, not Eastern. On Jun 25, 2013, at 1:58 PM, Wesley Bland <wbland(a)mcs.anl.gov> wrote: > Just a reminder that we're having our usual call today at 2:00. > > Call in information: > > Dial-in information: 218-339-4600 > Code: 623998# > > Wesley

1 0

[Mpi3-ft] Conference Call June 25
by Wesley Bland 25 Jun '13

25 Jun '13

Just a reminder that we're having our usual call today at 2:00. Call in information: Dial-in information: 218-339-4600 Code: 623998# Wesley

1 0

Re: [Mpi3-ft] Conference Call Jun, 11
by Wesley Bland 11 Jun '13

11 Jun '13

The notes from the call have been posted to the wiki page: https://svn.mpi-forum.org/trac/mpi-forum-web/wiki/2013-06-11 Wesley On Jun 11, 2013, at 1:56 PM, Wesley Bland <wbland(a)mcs.anl.gov> wrote: > In case you just subscribed to the list and would like to join, here's the call-in information: > > Dial-in information: 218-339-4600 > Code: 623998# > > On Jun 10, 2013, at 10:00 AM, Aurélien Bouteiller <bouteill(a)icl.utk.edu> wrote: > >> Dear WG members: >> >> According to our new schedule, we will have a confcall on TUESDAY JUNE 11, 15H00 EDT >> >> Agenda: >> * Debriefing from Forum >> * Consider relaxed semantics for I/O >> * Consider an additional fn to obtain the globally agreed on group of failed processes >> >> >> I will not be able to attend, but I think Wesley will. >> >> >> >> In any case, here are my personal notes of the meeting minutes. >> >> * I/O: >> * Mohamed expressed that he would find useful to be able to continue operating on a FILE after some error has been reported, at least to perform local, non-synchronizing operations. The issue is that we have defined the file pointer as "undefined" after a failure is reported, so it makes no sense to continue anything as it would result in incorrect behavior. >> >> Yet, it is difficult to offer better: with async operations, what should be the status of the file pointer when ops A B C are pending, Op C completes, op B fails, all during the same completion call ? We need to think and check with I/O People (And the I/O text to see what they define on backend errors). >> >> RMA: >> * No strong issue were found by the Forum, it seems that we are progressing on track for this one. >> >> Coll: >> * Thorsten has been very vocal, which spurred a lot of discussions. >> * One of the topic was "why do you need shrink, comm_create can do it too". I explained that shrink is FT and operates on Revoked comms, which comm_create is not (at least by spec, the typical implementation would at least operate correctly on comms with errors, w/o new failures). We do not want to overload non-FT functions with FT semantic. This is bad practice and will bite us back later if we do that >> * Yet, Jim made the comment that there is no clear way for a user to know what is the set of globally known failures, without revoking (or writing your own P2P FT all reduce on groups). We may want to consider this, at least to illustrate how one can code such a fn as a helper in the current context. >> >> Tickets: >> * No objections to proposed closure of tickets. >> >> >> >> >> Other progress: >> * Text is now rebased from MPI-3 document >> * Text repo is now read-public >> >> >> That's all folks! >> >> Aurelien >> >> -- >> * Dr. Aurélien Bouteiller >> * Researcher at Innovative Computing Laboratory >> * University of Tennessee >> * 1122 Volunteer Boulevard, suite 309b >> * Knoxville, TN 37996 >> * 865 974 9375 >> >> >> >> >> >> >> >> >> _______________________________________________ >> mpi3-ft mailing list >> mpi3-ft(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft >

1 0

Re: [Mpi3-ft] Conference Call Jun, 11
by Wesley Bland 11 Jun '13

11 Jun '13

In case you just subscribed to the list and would like to join, here's the call-in information: Dial-in information: 218-339-4600 Code: 623998# On Jun 10, 2013, at 10:00 AM, Aurï¿½lien Bouteiller <bouteill(a)icl.utk.edu> wrote: > Dear WG members: > > According to our new schedule, we will have a confcall on TUESDAY JUNE 11, 15H00 EDT > > Agenda: > * Debriefing from Forum > * Consider relaxed semantics for I/O > * Consider an additional fn to obtain the globally agreed on group of failed processes > > > I will not be able to attend, but I think Wesley will. > > > > In any case, here are my personal notes of the meeting minutes. > > * I/O: > * Mohamed expressed that he would find useful to be able to continue operating on a FILE after some error has been reported, at least to perform local, non-synchronizing operations. The issue is that we have defined the file pointer as "undefined" after a failure is reported, so it makes no sense to continue anything as it would result in incorrect behavior. > > Yet, it is difficult to offer better: with async operations, what should be the status of the file pointer when ops A B C are pending, Op C completes, op B fails, all during the same completion call ? We need to think and check with I/O People (And the I/O text to see what they define on backend errors). > > RMA: > * No strong issue were found by the Forum, it seems that we are progressing on track for this one. > > Coll: > * Thorsten has been very vocal, which spurred a lot of discussions. > * One of the topic was "why do you need shrink, comm_create can do it too". I explained that shrink is FT and operates on Revoked comms, which comm_create is not (at least by spec, the typical implementation would at least operate correctly on comms with errors, w/o new failures). We do not want to overload non-FT functions with FT semantic. This is bad practice and will bite us back later if we do that > * Yet, Jim made the comment that there is no clear way for a user to know what is the set of globally known failures, without revoking (or writing your own P2P FT all reduce on groups). We may want to consider this, at least to illustrate how one can code such a fn as a helper in the current context. > > Tickets: > * No objections to proposed closure of tickets. > > > > > Other progress: > * Text is now rebased from MPI-3 document > * Text repo is now read-public > > > That's all folks! > > Aurelien > > -- > * Dr. Aurï¿½lien Bouteiller > * Researcher at Innovative Computing Laboratory > * University of Tennessee > * 1122 Volunteer Boulevard, suite 309b > * Knoxville, TN 37996 > * 865 974 9375 > > > > > > > > > _______________________________________________ > mpi3-ft mailing list > mpi3-ft(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft

1 0

[Mpi3-ft] Conference Call Jun, 11
by Aurélien Bouteiller 10 Jun '13

10 Jun '13

Dear WG members: According to our new schedule, we will have a confcall on TUESDAY JUNE 11, 15H00 EDT Agenda: * Debriefing from Forum * Consider relaxed semantics for I/O * Consider an additional fn to obtain the globally agreed on group of failed processes I will not be able to attend, but I think Wesley will. In any case, here are my personal notes of the meeting minutes. * I/O: * Mohamed expressed that he would find useful to be able to continue operating on a FILE after some error has been reported, at least to perform local, non-synchronizing operations. The issue is that we have defined the file pointer as "undefined" after a failure is reported, so it makes no sense to continue anything as it would result in incorrect behavior. Yet, it is difficult to offer better: with async operations, what should be the status of the file pointer when ops A B C are pending, Op C completes, op B fails, all during the same completion call ? We need to think and check with I/O People (And the I/O text to see what they define on backend errors). RMA: * No strong issue were found by the Forum, it seems that we are progressing on track for this one. Coll: * Thorsten has been very vocal, which spurred a lot of discussions. * One of the topic was "why do you need shrink, comm_create can do it too". I explained that shrink is FT and operates on Revoked comms, which comm_create is not (at least by spec, the typical implementation would at least operate correctly on comms with errors, w/o new failures). We do not want to overload non-FT functions with FT semantic. This is bad practice and will bite us back later if we do that * Yet, Jim made the comment that there is no clear way for a user to know what is the set of globally known failures, without revoking (or writing your own P2P FT all reduce on groups). We may want to consider this, at least to illustrate how one can code such a fn as a helper in the current context. Tickets: * No objections to proposed closure of tickets. Other progress: * Text is now rebased from MPI-3 document * Text repo is now read-public That's all folks! Aurelien -- * Dr. Aurï¿½lien Bouteiller * Researcher at Innovative Computing Laboratory * University of Tennessee * 1122 Volunteer Boulevard, suite 309b * Knoxville, TN 37996 * 865 974 9375

1 0

Re: [Mpi3-ft] Updated Document / Action Items
by Thomas Naughton 07 Jun '13

07 Jun '13

Hi George, Ok, thanks for pointer/URL. --tjn _________________________________________________________________________ Thomas Naughton naughtont(a)ornl.gov Research Associate (865) 576-4184 On Thu, 6 Jun 2013, George Bosilca wrote: > Up to now the document was managed in a private mercurial repository in bitbucket. Read access has been opened up, it is now available to everyone @ https://[email protected]/bosilca/mpi3ft > > George. > > > On Jun 6, 2013, at 20:29 , Thomas Naughton <naughtont(a)ornl.gov> wrote: > >> Hi, >> >> Does the LaTeX document live in an SVN? Can there just be a link to the >> source files and the latest generated PDF for the files? That forgoes the >> Trac/LaTeX sync issue. >> >> Sorry for my ignorance, I just don't know where the source files live for >> the specification. >> >> Thanks, >> --tjn >> >> _________________________________________________________________________ >> Thomas Naughton naughtont(a)ornl.gov >> Research Associate (865) 576-4184 >> >> >> On Thu, 6 Jun 2013, Wesley Bland wrote: >> >>> I agree wholeheartedly. I'll give it a few days so if someone wants to step up to save it, they can do so. >>> >>> Wesley >>> >>> On Jun 6, 2013, at 9:52 AM, Aurï¿½lien Bouteiller <bouteill(a)icl.utk.edu> wrote: >>> >>>> >>>> Le 6 juin 2013 ï¿½ 09:40, Wesley Bland <wbland(a)mcs.anl.gov> a ï¿½crit : >>>> >>>>> The updated ULFM proposal has been attached to the wiki page: >>>>> >>>>> https://svn.mpi-forum.org/trac/mpi-forum-web/wiki/User_Level_Failure_Mitiga… >>>>> >>>>> The actual text on that page is out of date. I doubt that it will ever actually stay in sync with the PDF version unless someone generates a tool to make that happen automagically. >>>>> >>>> I propose we remove this text from the webpage. The chapter is attached in PDF form there, it is better to have only an up to date text visible, older versions lying around are confusing. >>>> >>>>> This version has updates based on the discussion with the RMA and I/O working groups at the current forum meeting. The remaining action items are to bring in some of the shared file pointer I/O people and get their input (Dries, Rob L.) on the most recent version. I'll send out a Doodle poll next week to schedule a time for that call. >>>>> >>>>> The RMA sections is getting closer to finalized. The next action item on that front is to have a reference implementation. >>>>> >>>>> Thanks, >>>>> Wesley >>>>> _______________________________________________ >>>>> mpi3-ft mailing list >>>>> mpi3-ft(a)lists.mpi-forum.org >>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft >>>> >>>> -- >>>> * Dr. Aurï¿½lien Bouteiller >>>> * Researcher at Innovative Computing Laboratory >>>> * University of Tennessee >>>> * 1122 Volunteer Boulevard, suite 309b >>>> * Knoxville, TN 37996 >>>> * 865 974 9375 >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> _______________________________________________ >>>> mpi3-ft mailing list >>>> mpi3-ft(a)lists.mpi-forum.org >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft >>> >>> >>> _______________________________________________ >>> mpi3-ft mailing list >>> mpi3-ft(a)lists.mpi-forum.org >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft >> _______________________________________________ >> mpi3-ft mailing list >> mpi3-ft(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft > > > _______________________________________________ > mpi3-ft mailing list > mpi3-ft(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft >

1 0

Re: [Mpi3-ft] Updated Document / Action Items
by George Bosilca 06 Jun '13

06 Jun '13

Up to now the document was managed in a private mercurial repository in bitbucket. Read access has been opened up, it is now available to everyone @ https://[email protected]/bosilca/mpi3ft George. On Jun 6, 2013, at 20:29 , Thomas Naughton <naughtont(a)ornl.gov> wrote: > Hi, > > Does the LaTeX document live in an SVN? Can there just be a link to the > source files and the latest generated PDF for the files? That forgoes the > Trac/LaTeX sync issue. > > Sorry for my ignorance, I just don't know where the source files live for > the specification. > > Thanks, > --tjn > > _________________________________________________________________________ > Thomas Naughton naughtont(a)ornl.gov > Research Associate (865) 576-4184 > > > On Thu, 6 Jun 2013, Wesley Bland wrote: > >> I agree wholeheartedly. I'll give it a few days so if someone wants to step up to save it, they can do so. >> >> Wesley >> >> On Jun 6, 2013, at 9:52 AM, Aurï¿½lien Bouteiller <bouteill(a)icl.utk.edu> wrote: >> >>> >>> Le 6 juin 2013 ï¿½ 09:40, Wesley Bland <wbland(a)mcs.anl.gov> a ï¿½crit : >>> >>>> The updated ULFM proposal has been attached to the wiki page: >>>> >>>> https://svn.mpi-forum.org/trac/mpi-forum-web/wiki/User_Level_Failure_Mitiga… >>>> >>>> The actual text on that page is out of date. I doubt that it will ever actually stay in sync with the PDF version unless someone generates a tool to make that happen automagically. >>>> >>> I propose we remove this text from the webpage. The chapter is attached in PDF form there, it is better to have only an up to date text visible, older versions lying around are confusing. >>> >>>> This version has updates based on the discussion with the RMA and I/O working groups at the current forum meeting. The remaining action items are to bring in some of the shared file pointer I/O people and get their input (Dries, Rob L.) on the most recent version. I'll send out a Doodle poll next week to schedule a time for that call. >>>> >>>> The RMA sections is getting closer to finalized. The next action item on that front is to have a reference implementation. >>>> >>>> Thanks, >>>> Wesley >>>> _______________________________________________ >>>> mpi3-ft mailing list >>>> mpi3-ft(a)lists.mpi-forum.org >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft >>> >>> -- >>> * Dr. Aurï¿½lien Bouteiller >>> * Researcher at Innovative Computing Laboratory >>> * University of Tennessee >>> * 1122 Volunteer Boulevard, suite 309b >>> * Knoxville, TN 37996 >>> * 865 974 9375 >>> >>> >>> >>> >>> >>> >>> >>> >>> _______________________________________________ >>> mpi3-ft mailing list >>> mpi3-ft(a)lists.mpi-forum.org >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft >> >> >> _______________________________________________ >> mpi3-ft mailing list >> mpi3-ft(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft > _______________________________________________ > mpi3-ft mailing list > mpi3-ft(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft

1 0

Re: [Mpi3-ft] Updated Document / Action Items
by Thomas Naughton 06 Jun '13

06 Jun '13

Hi, Does the LaTeX document live in an SVN? Can there just be a link to the source files and the latest generated PDF for the files? That forgoes the Trac/LaTeX sync issue. Sorry for my ignorance, I just don't know where the source files live for the specification. Thanks, --tjn _________________________________________________________________________ Thomas Naughton naughtont(a)ornl.gov Research Associate (865) 576-4184 On Thu, 6 Jun 2013, Wesley Bland wrote: > I agree wholeheartedly. I'll give it a few days so if someone wants to step up to save it, they can do so. > > Wesley > > On Jun 6, 2013, at 9:52 AM, Aurï¿½lien Bouteiller <bouteill(a)icl.utk.edu> wrote: > >> >> Le 6 juin 2013 ï¿½ 09:40, Wesley Bland <wbland(a)mcs.anl.gov> a ï¿½crit : >> >>> The updated ULFM proposal has been attached to the wiki page: >>> >>> https://svn.mpi-forum.org/trac/mpi-forum-web/wiki/User_Level_Failure_Mitiga… >>> >>> The actual text on that page is out of date. I doubt that it will ever actually stay in sync with the PDF version unless someone generates a tool to make that happen automagically. >>> >> I propose we remove this text from the webpage. The chapter is attached in PDF form there, it is better to have only an up to date text visible, older versions lying around are confusing. >> >>> This version has updates based on the discussion with the RMA and I/O working groups at the current forum meeting. The remaining action items are to bring in some of the shared file pointer I/O people and get their input (Dries, Rob L.) on the most recent version. I'll send out a Doodle poll next week to schedule a time for that call. >>> >>> The RMA sections is getting closer to finalized. The next action item on that front is to have a reference implementation. >>> >>> Thanks, >>> Wesley >>> _______________________________________________ >>> mpi3-ft mailing list >>> mpi3-ft(a)lists.mpi-forum.org >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft >> >> -- >> * Dr. Aurï¿½lien Bouteiller >> * Researcher at Innovative Computing Laboratory >> * University of Tennessee >> * 1122 Volunteer Boulevard, suite 309b >> * Knoxville, TN 37996 >> * 865 974 9375 >> >> >> >> >> >> >> >> >> _______________________________________________ >> mpi3-ft mailing list >> mpi3-ft(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft > > > _______________________________________________ > mpi3-ft mailing list > mpi3-ft(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-ft >

1 0