September 2012 - mpi-forum - mailman3.cels.anl.gov

Re: [Mpi-forum] MPI Next leadership
by Anthony Skjellum 29 Sep '12

29 Sep '12

A motion to adjourn is nonetheless always in order... Anthony Skjellum, PhD 205-807-4968 On Sep 29, 2012, at 5:16 PM, Pavan Balaji <balaji(a)mcs.anl.gov> wrote: > > ----- Original Message ----- >> Simply adjourning might be less than ideal. The last time that >> happened, it took a decade to restart :-) However, six months or a >> year of hiatus might actually be a good thing. That wouldn't require >> people to stop working on ideas - just give time for the forum as a >> whole to catch their breath... > > While I would have agreed with this in general, we should remember that some working groups (such as Fault Tolerance) that could not make it into 3.0 have been counting on this upcoming release (MPI-3.1?) to get those features in. > > There are also some features that got discussed on the RMA mailing list for additions into MPI-3.1. > > We have similar plans in the hybrid working group as well (or a small subset of the very small working group). > > Also note that we are slowing down the pace of the Forum meetings to quarterly. That should help. Another option is to have just two meetings in 2013 if we want to "take it easy" for an year. That won't technically be a break, so workaholics can feel happy too. > > -- Pavan > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum

1 0

Re: [Mpi-forum] MPI Next leadership
by Pavan Balaji 29 Sep '12

29 Sep '12

----- Original Message ----- > Simply adjourning might be less than ideal. The last time that > happened, it took a decade to restart :-) However, six months or a > year of hiatus might actually be a good thing. That wouldn't require > people to stop working on ideas - just give time for the forum as a > whole to catch their breath... While I would have agreed with this in general, we should remember that some working groups (such as Fault Tolerance) that could not make it into 3.0 have been counting on this upcoming release (MPI-3.1?) to get those features in. There are also some features that got discussed on the RMA mailing list for additions into MPI-3.1. We have similar plans in the hybrid working group as well (or a small subset of the very small working group). Also note that we are slowing down the pace of the Forum meetings to quarterly. That should help. Another option is to have just two meetings in 2013 if we want to "take it easy" for an year. That won't technically be a break, so workaholics can feel happy too. -- Pavan

1 0

Re: [Mpi-forum] Meeting schedule for 2013
by Richard Graham 28 Sep '12

28 Sep '12

I will add to this - there is NO formal relationship between the meeting and the conference. Both are independent bodies, that make their own decisions. EuroMPI does attract many MPI implementers, and since it is the only conference that focuses on MPI, and there is concern about it's viability if the two are decoupled. As an MPI developer, this is of great concern to me if EuroMPI goes away. Others expressed similar concern, and since we did not have time to discuss this at the last forum meeting, this is coming up on the mailing list, as we are trying to finalize on the meeting dates for next year. So far, what I see from the mailing list and the in-person conversations earlier this week, is that most folks are in favor of keeping the connection. We can have a mail vote on this, or can postpone to the Dec meeting - no hurry on this one. Rich -----Original Message----- From: mpi-forum-bounces(a)lists.mpi-forum.org [mailto:[email protected]] On Behalf Of Aurï¿½lien Bouteiller Sent: Friday, September 28, 2012 10:12 AM To: Bronis R. de Supinski; Main MPI Forum mailing list Subject: Re: [Mpi-forum] Meeting schedule for 2013 Le 27 sept. 2012 ï¿½ 08:47, "Bronis R. de Supinski" <bronis(a)llnl.gov> a ï¿½crit : > > Rolf: > > Re: >>> even years: 3US+euroMPI >>> odd years: 2US,1Asia+EuroMPI. >> >> Yes, this was the last position at the Forum and also discussed at the EuroMPI business meeting. > > As I understand the situation, this option was not discussed at the > Forum. As stated previously, the matter was decided in Chicago and > reopening it without an announcement was inappropriate. > > In addition, what is discussed at the EuroMPI business meeting is not > relevant. It is not their decision. > Bronis, the matter has not been decided, as we are obviously still discussing it. Nothing was decided either in Vienna as it felt like there was no consensus still (straw votes where very divided, and many options came close to one another, we should be driven by consensus right?). No need to feel so strong on "rules" again, we are just expressing wishes so far. It so happens that many people are going to attend EuroMPI anyway, so collocating is a clear way of cutting cost for many attendees; however, it could be interesting to make so that we don't have to spend a WE in a foreign country when attending both the conference and the Forum, as this is expensive and annoying. Also, it is to be noted that european institutions participation is declining recently. Locating meetings in Asia was taken as a step to address the same issue when Asian participation was low. We might consider this in the equation, or risk that the forum becomes an US only body, and not an international body anymore. So the question is not in terms of us against them (the conference), it is in terms of what is convenient for the Forum members (many of whom do attend the conference, many of whom are not US based institutions) and makes sense for an -international- standardization body. Regards, Aurelien >> The MPI Forum was re-started by the activities at several EuroMPI conferences, the last one at Paris 2007. >> >> Therefore, there is a strong relationship and resposibility and to >> have the MPI Forum meeting in US and the EuroMPI in Europe both in >> fall, would not reflect this responsibility. >> >> As far as I see, the colleagues in Madrid would like to host the MPI >> Forum meeting together with the EuroMPI and therefore, I expect - >> based on the previous meeting rules - we will have >> >> MPI-next Forum meeting Sep. 11-13, 2013, (instead of Sep. 9-11 in San >> Jose), and EuroMPI, Sep. 15-18, 2013, both in Madrid, Spain >> >> This may change in later years if the EuroMPI may go to Asia, but >> this is another discussion within the EuroMPI organizers. > > What I hear is that holding the MPI Forum meeting separately from > EuroMPI would damage EuroMPI. Well, forcing the Forum to co-locate is > putting the cart before the horse. If EuroMPI cannot withstand a > separate meeting, as decided by the Forum, then EuroMPI needs to adapt > -- not the other way around. > > As was previously suggested, EuroMPI is free to rebrand as > International MPI and to co-locate with Forum meetings. > If the conference demonstrated that flexibility, I would be more > willing to consider their desires as an MPI Forum voting member. > > Bronis > > > > >> Best regards >> Rolf >> >> ----- Original Message ----- >>> From: "Aurï¿½lien Bouteiller" <bouteill(a)eecs.utk.edu> >>> To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> >>> Sent: Thursday, September 27, 2012 2:41:08 AM >>> Subject: Re: [Mpi-forum] Meeting schedule for 2013 The 3-1 with 1 >>> alternate between asia and europe collected the most vote. However >>> there was significant disappointment about not collocating a meeting >>> with EuroMPI, and after the straw vote closed, another interesting >>> possibility was envisioned: >>> even years: 3US+euroMPI >>> odd years: 2US,1Asia+EuroMPI. >>> This last proposal has not been straw voted, so it is not possible >>> to know if it has support, but it seemed from the reception it had >>> that the idea is not unbearable for most US institutions. >>> Aurelien >>> Le 26 sept. 2012 ï¿½ 15:17, Jeff Squyres <jsquyres(a)cisco.com> a ï¿½crit : >>> > I don't remember there being a majority for the 2-2 option here in >>> > Vienna -- am I remembering last week incorrectly? >>> > >>> > I definitely can't do 2 international meetings next week. >>> > >>> > >>> > >>> > >>> > On Sep 26, 2012, at 8:02 PM, Richard Graham wrote: >>> > >>> >> This was discussed and there was another vote at the meeting in >>> >> Vienna. This note is just a follow up to the discussions at that >>> >> meeting. >>> >> >>> >> Rich >>> >> >>> >> -----Original Message----- >>> >> From: mpi-forum-bounces(a)lists.mpi-forum.org >>> >> [mailto:[email protected]] On Behalf Of >>> >> Schulz, Martin >>> >> Sent: Wednesday, September 26, 2012 7:36 PM >>> >> To: Bronis R. de Supinski; Main MPI Forum mailing list >>> >> Subject: Re: [Mpi-forum] Meeting schedule for 2013 >>> >> >>> >> Hi Rich, >>> >> >>> >> I have the same recollection as Bronis and Mohamad. I thought >>> >> there was a strong majority (at least at the Chicago meeting) for >>> >> a 3-1 solution with alternating non-US locations (one year Asia, >>> >> one year Europe, but not both in one year - as it also shows on >>> >> the current tentative meeting calendar for next year). I know >>> >> this is not fair to everyone, but the group (again, at least in >>> >> Chicago, which was pretty complete, though, since we were >>> >> wrapping up 3.0) felt that this would maximize active participation. >>> >> >>> >> Martin >>> >> >>> >> >>> >> On Sep 26, 2012, at 9:39 AM, Bronis R. de Supinski wrote: >>> >> >>> >>> >>> >>> Rich: >>> >>> >>> >>> I would have to disagree. At best, you are splitting hairs to >>> >>> say this option was not voted on. In the "alternate" >>> >>> years, it is simply the 2-2 option that was voted on and did not >>> >>> come anywhere near the 3-1 option in terms of votes. >>> >>> >>> >>> The arguments against 2-2 (even in alternating years) still apply. >>> >>> It >>> >>> significantly lowers attendance by active participants. >>> >>> >>> >>> Bronis >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> On Wed, 26 Sep 2012, Richard Graham wrote: >>> >>> >>> >>>> Such decisions are not made lightly - I assure you this. This >>> >>>> was discussed at the last meeting, and this was an option that >>> >>>> was not voted on. However, this was discussed again at the end >>> >>>> of the meeting, and some at EuroMPI. >>> >>>> >>> >>>> Rich >>> >>>> >>> >>>> -----Original Message----- >>> >>>> From: mpi-forum-bounces(a)lists.mpi-forum.org >>> >>>> [mailto:[email protected]] On Behalf Of >>> >>>> Mohamad Chaarawi >>> >>>> Sent: Wednesday, September 26, 2012 6:30 PM >>> >>>> To: mpi-forum(a)lists.mpi-forum.org >>> >>>> Subject: Re: [Mpi-forum] Meeting schedule for 2013 >>> >>>> >>> >>>> Hi Rich, >>> >>>> >>> >>>> On 9/26/2012 10:30 AM, Richard Graham wrote: >>> >>>>> Therefore, I propose to keep the European meeting of the Forum >>> >>>>> in conjunction with EuroMPI, and every other year also have a >>> >>>>> meeting in Japan. With this in mind, here is the proposed >>> >>>>> schedule for >>> >>>>> 2012 >>> >>>> >>> >>>> The majority at the last meeting in Chicago were against having >>> >>>> 2 meetings internationally. We had several votes on different >>> >>>> formations (4-0, 3-1, and 2-2), and 3-1 was the majority vote. >>> >>>> >>> >>>> It would have been fair to have a similar vote in Europe and >>> >>>> see how the participants vote.. >>> >>>> Did the attendance at the Vienna meeting vote on a different >>> >>>> schedule, or is your change coming from just some concerns >>> >>>> (which are also valid) from some people? >>> >>>> >>> >>>> Can we have a formal organizational Vote by eligible >>> >>>> organizations on this instead of just scheduling meetings based >>> >>>> on suggestions that vary from meeting to meeting? >>> >>>> This vote can be done over the phone. >>> >>>> >>> >>>> Thanks, >>> >>>> Mohamad >>> >>>> _______________________________________________ >>> >>>> mpi-forum mailing list >>> >>>> mpi-forum(a)lists.mpi-forum.org >>> >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>> >>>> >>> >>>> _______________________________________________ >>> >>>> mpi-forum mailing list >>> >>>> mpi-forum(a)lists.mpi-forum.org >>> >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>> >>>> >>> >>> _______________________________________________ >>> >>> mpi-forum mailing list >>> >>> mpi-forum(a)lists.mpi-forum.org >>> >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>> >> >>> >> _________________________________________________________________ >>> >> _______ Martin Schulz, schulzm(a)llnl.gov, >>> >> http://people.llnl.gov/schulzm CASC @ Lawrence Livermore National >>> >> Laboratory, Livermore, USA >>> >> >>> >> >>> >> >>> >> >>> >> _______________________________________________ >>> >> mpi-forum mailing list >>> >> mpi-forum(a)lists.mpi-forum.org >>> >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>> >> >>> >> _______________________________________________ >>> >> mpi-forum mailing list >>> >> mpi-forum(a)lists.mpi-forum.org >>> >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>> > >>> > >>> > -- >>> > Jeff Squyres >>> > jsquyres(a)cisco.com >>> > For corporate legal information go to: >>> > http://www.cisco.com/web/about/doing_business/legal/cri/ >>> > >>> > >>> > _______________________________________________ >>> > mpi-forum mailing list >>> > mpi-forum(a)lists.mpi-forum.org >>> > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>> -- >>> * Dr. Aurï¿½lien Bouteiller >>> * Researcher at Innovative Computing Laboratory >>> * University of Tennessee >>> * 1122 Volunteer Boulevard, suite 309b >>> * Knoxville, TN 37996 >>> * 865 974 9375 >>> _______________________________________________ >>> 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: Allmandring 30) >> >> _______________________________________________ >> mpi-forum mailing list >> mpi-forum(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum -- * Dr. Aurï¿½lien Bouteiller * Researcher at Innovative Computing Laboratory * University of Tennessee * 1122 Volunteer Boulevard, suite 309b * Knoxville, TN 37996 * 865 974 9375 _______________________________________________ mpi-forum mailing list mpi-forum(a)lists.mpi-forum.org http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum

1 0

Re: [Mpi-forum] Meeting schedule for 2013
by Aurélien Bouteiller 28 Sep '12

28 Sep '12

Le 27 sept. 2012 ï¿½ 08:47, "Bronis R. de Supinski" <bronis(a)llnl.gov> a ï¿½crit : > > Rolf: > > Re: >>> even years: 3US+euroMPI >>> odd years: 2US,1Asia+EuroMPI. >> >> Yes, this was the last position at the Forum and also discussed at the EuroMPI business meeting. > > As I understand the situation, this option was not discussed > at the Forum. As stated previously, the matter was decided in > Chicago and reopening it without an announcement was inappropriate. > > In addition, what is discussed at the EuroMPI business meeting > is not relevant. It is not their decision. > Bronis, the matter has not been decided, as we are obviously still discussing it. Nothing was decided either in Vienna as it felt like there was no consensus still (straw votes where very divided, and many options came close to one another, we should be driven by consensus right?). No need to feel so strong on "rules" again, we are just expressing wishes so far. It so happens that many people are going to attend EuroMPI anyway, so collocating is a clear way of cutting cost for many attendees; however, it could be interesting to make so that we don't have to spend a WE in a foreign country when attending both the conference and the Forum, as this is expensive and annoying. Also, it is to be noted that european institutions participation is declining recently. Locating meetings in Asia was taken as a step to address the same issue when Asian participation was low. We might consider this in the equation, or risk that the forum becomes an US only body, and not an international body anymore. So the question is not in terms of us against them (the conference), it is in terms of what is convenient for the Forum members (many of whom do attend the conference, many of whom are not US based institutions) and makes sense for an -international- standardization body. Regards, Aurelien >> The MPI Forum was re-started by the activities at several EuroMPI conferences, the last one at Paris 2007. >> >> Therefore, there is a strong relationship and resposibility >> and to have the MPI Forum meeting in US and the EuroMPI in Europe >> both in fall, would not reflect this responsibility. >> >> As far as I see, the colleagues in Madrid would like to host the MPI Forum meeting together with the EuroMPI and therefore, I expect - based on the previous meeting rules - we will have >> >> MPI-next Forum meeting Sep. 11-13, 2013, >> (instead of Sep. 9-11 in San Jose), and >> EuroMPI, Sep. 15-18, 2013, both in Madrid, Spain >> >> This may change in later years if the EuroMPI may go to Asia, >> but this is another discussion within the EuroMPI organizers. > > What I hear is that holding the MPI Forum meeting > separately from EuroMPI would damage EuroMPI. Well, > forcing the Forum to co-locate is putting the cart > before the horse. If EuroMPI cannot withstand a > separate meeting, as decided by the Forum, then > EuroMPI needs to adapt -- not the other way around. > > As was previously suggested, EuroMPI is free to rebrand > as International MPI and to co-locate with Forum meetings. > If the conference demonstrated that flexibility, I would > be more willing to consider their desires as an MPI > Forum voting member. > > Bronis > > > > >> Best regards >> Rolf >> >> ----- Original Message ----- >>> From: "Aurï¿½lien Bouteiller" <bouteill(a)eecs.utk.edu> >>> To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> >>> Sent: Thursday, September 27, 2012 2:41:08 AM >>> Subject: Re: [Mpi-forum] Meeting schedule for 2013 >>> The 3-1 with 1 alternate between asia and europe collected the most >>> vote. However there was significant disappointment about not >>> collocating a meeting with EuroMPI, and after the straw vote closed, >>> another interesting possibility was envisioned: >>> even years: 3US+euroMPI >>> odd years: 2US,1Asia+EuroMPI. >>> This last proposal has not been straw voted, so it is not possible to >>> know if it has support, but it seemed from the reception it had that >>> the idea is not unbearable for most US institutions. >>> Aurelien >>> Le 26 sept. 2012 ï¿½ 15:17, Jeff Squyres <jsquyres(a)cisco.com> a ï¿½crit : >>> > I don't remember there being a majority for the 2-2 option here in >>> > Vienna -- am I remembering last week incorrectly? >>> > >>> > I definitely can't do 2 international meetings next week. >>> > >>> > >>> > >>> > >>> > On Sep 26, 2012, at 8:02 PM, Richard Graham wrote: >>> > >>> >> This was discussed and there was another vote at the meeting in >>> >> Vienna. This note is just a follow up to the discussions at that >>> >> meeting. >>> >> >>> >> Rich >>> >> >>> >> -----Original Message----- >>> >> From: mpi-forum-bounces(a)lists.mpi-forum.org >>> >> [mailto:[email protected]] On Behalf Of Schulz, >>> >> Martin >>> >> Sent: Wednesday, September 26, 2012 7:36 PM >>> >> To: Bronis R. de Supinski; Main MPI Forum mailing list >>> >> Subject: Re: [Mpi-forum] Meeting schedule for 2013 >>> >> >>> >> Hi Rich, >>> >> >>> >> I have the same recollection as Bronis and Mohamad. I thought there >>> >> was a strong majority (at least at the Chicago meeting) for a 3-1 >>> >> solution with alternating non-US locations (one year Asia, one year >>> >> Europe, but not both in one year - as it also shows on the current >>> >> tentative meeting calendar for next year). I know this is not fair >>> >> to everyone, but the group (again, at least in Chicago, which was >>> >> pretty complete, though, since we were wrapping up 3.0) felt that >>> >> this would maximize active participation. >>> >> >>> >> Martin >>> >> >>> >> >>> >> On Sep 26, 2012, at 9:39 AM, Bronis R. de Supinski wrote: >>> >> >>> >>> >>> >>> Rich: >>> >>> >>> >>> I would have to disagree. At best, you are splitting hairs to say >>> >>> this >>> >>> option was not voted on. In the "alternate" >>> >>> years, it is simply the 2-2 option that was voted on and did not >>> >>> come >>> >>> anywhere near the 3-1 option in terms of votes. >>> >>> >>> >>> The arguments against 2-2 (even in alternating years) still apply. >>> >>> It >>> >>> significantly lowers attendance by active participants. >>> >>> >>> >>> Bronis >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> On Wed, 26 Sep 2012, Richard Graham wrote: >>> >>> >>> >>>> Such decisions are not made lightly - I assure you this. This was >>> >>>> discussed at the last meeting, and this was an option that was >>> >>>> not voted on. However, this was discussed again at the end of the >>> >>>> meeting, and some at EuroMPI. >>> >>>> >>> >>>> Rich >>> >>>> >>> >>>> -----Original Message----- >>> >>>> From: mpi-forum-bounces(a)lists.mpi-forum.org >>> >>>> [mailto:[email protected]] On Behalf Of >>> >>>> Mohamad >>> >>>> Chaarawi >>> >>>> Sent: Wednesday, September 26, 2012 6:30 PM >>> >>>> To: mpi-forum(a)lists.mpi-forum.org >>> >>>> Subject: Re: [Mpi-forum] Meeting schedule for 2013 >>> >>>> >>> >>>> Hi Rich, >>> >>>> >>> >>>> On 9/26/2012 10:30 AM, Richard Graham wrote: >>> >>>>> Therefore, I propose to keep the European meeting of the Forum >>> >>>>> in >>> >>>>> conjunction with EuroMPI, and every other year also have a >>> >>>>> meeting >>> >>>>> in Japan. With this in mind, here is the proposed schedule for >>> >>>>> 2012 >>> >>>> >>> >>>> The majority at the last meeting in Chicago were against having 2 >>> >>>> meetings internationally. We had several votes on different >>> >>>> formations (4-0, 3-1, and 2-2), and 3-1 was the majority vote. >>> >>>> >>> >>>> It would have been fair to have a similar vote in Europe and see >>> >>>> how the participants vote.. >>> >>>> Did the attendance at the Vienna meeting vote on a different >>> >>>> schedule, or is your change coming from just some concerns (which >>> >>>> are also valid) from some people? >>> >>>> >>> >>>> Can we have a formal organizational Vote by eligible >>> >>>> organizations on this instead of just scheduling meetings based >>> >>>> on suggestions that vary from meeting to meeting? >>> >>>> This vote can be done over the phone. >>> >>>> >>> >>>> Thanks, >>> >>>> Mohamad >>> >>>> _______________________________________________ >>> >>>> mpi-forum mailing list >>> >>>> mpi-forum(a)lists.mpi-forum.org >>> >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>> >>>> >>> >>>> _______________________________________________ >>> >>>> mpi-forum mailing list >>> >>>> mpi-forum(a)lists.mpi-forum.org >>> >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>> >>>> >>> >>> _______________________________________________ >>> >>> mpi-forum mailing list >>> >>> mpi-forum(a)lists.mpi-forum.org >>> >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>> >> >>> >> ________________________________________________________________________ >>> >> Martin Schulz, schulzm(a)llnl.gov, http://people.llnl.gov/schulzm >>> >> CASC @ Lawrence Livermore National Laboratory, Livermore, USA >>> >> >>> >> >>> >> >>> >> >>> >> _______________________________________________ >>> >> mpi-forum mailing list >>> >> mpi-forum(a)lists.mpi-forum.org >>> >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>> >> >>> >> _______________________________________________ >>> >> mpi-forum mailing list >>> >> mpi-forum(a)lists.mpi-forum.org >>> >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>> > >>> > >>> > -- >>> > Jeff Squyres >>> > jsquyres(a)cisco.com >>> > For corporate legal information go to: >>> > http://www.cisco.com/web/about/doing_business/legal/cri/ >>> > >>> > >>> > _______________________________________________ >>> > mpi-forum mailing list >>> > mpi-forum(a)lists.mpi-forum.org >>> > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>> -- >>> * Dr. Aurï¿½lien Bouteiller >>> * Researcher at Innovative Computing Laboratory >>> * University of Tennessee >>> * 1122 Volunteer Boulevard, suite 309b >>> * Knoxville, TN 37996 >>> * 865 974 9375 >>> _______________________________________________ >>> 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: Allmandring 30) >> >> _______________________________________________ >> mpi-forum mailing list >> mpi-forum(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum -- * 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: [Mpi-forum] MPI Next leadership
by Underwood, Keith D 27 Sep '12

27 Sep '12

Simply adjourning might be less than ideal. The last time that happened, it took a decade to restart :-) However, six months or a year of hiatus might actually be a good thing. That wouldn't require people to stop working on ideas - just give time for the forum as a whole to catch their breath... Keith > -----Original Message----- > From: mpi-forum-bounces(a)lists.mpi-forum.org [mailto:mpi-forum- > bounces(a)lists.mpi-forum.org] On Behalf Of Anthony Skjellum > Sent: Thursday, September 27, 2012 8:18 AM > To: Main MPI Forum mailing list > Subject: Re: [Mpi-forum] MPI Next leadership > > Besides leadership, we should ask "why" MPI Next "now" vs "later" vs. > "ever." > > According to Robert's Rules of Order, a motion to adjourn is always in order. > Why are we assuming that MPI goes on as before, with multiyear planning? > We should first entertain, second, and either accept or defeat a motion to > adjourn, and those voting against adjourning should provide rationale for > continuing "now." > A second vote could be to "wait X months/years" and then automatically > return. > > For instance, why not give the OpenMPI and MPICH development teams a > chance to finish their work, distribute releases, write up findings, discover > errata, tune, let people get trained, let applications be ported and use those > new features, achieve buy-in and acceptance, and have actual users give > feedback, so we come tothe "Next" with the experience of results? > > In the meantime, research on the areas that are still controversial could > continue without the need to "be standardized" at the same time as they are > being figured out. > > > > > ----- Original Message ----- > From: "Rolf Rabenseifner" <rabenseifner(a)hlrs.de> > To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> > Sent: Thursday, September 27, 2012 5:17:15 AM > Subject: Re: [Mpi-forum] MPI Next leadership > > As MPI-2.1 chair, I must clearly say that I could not have been the MPI-2.1 > chair if there would not have been Rich as the convener of the meetings. > And I'm thankful that Rich has done it the whole time together with the local > organizers, mainly Jeff, the team of Microsoft and all European organizers at > the EuroMPI conferences. > > As far as I understand, Rich would continue to be our convener. > > About MPI-next chair: > Although the chair should clearly be neutral, I would prefer to have a MPI- > next chair from a public research institution. > Of course he/she must have the support from her/his institution to travel to > the meetings. > I see also that he/she need not only to focus on the new topics, but also on > the quality of the standard, i.e., after the time and content is right, to > organizing also the finalization of such an new version of MPI. > > Rolf > > ----- Original Message ----- > > From: "Rajeev Thakur" <thakur(a)mcs.anl.gov> > > To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> > > Sent: Thursday, September 27, 2012 10:51:45 AM > > Subject: Re: [Mpi-forum] MPI Next leadership It has also been > > suggested that the Forum take a break for a bit before resuming again, > > but I don't know if that was discussed at the last meeting. > > > > Rajeev > > > > > > > > On Sep 27, 2012, at 1:09 AM, Schulz, Martin wrote: > > > > > On Sep 26, 2012, at 9:46 PM, Bronis R. de Supinski wrote: > > > > > >> > > >> Jeff: > > >> > > >> My intent is not to "neuter" the chair. Instead, I am only stating > > >> that the chair should clearly state when (s)he is voicing > > >> personal/institutional opinions as opposed to acting as leader > > >> (sometimes the roles conflict and also sometimes the leader does > > >> have to enforce the clear general opinion). > > > > > > I think you are saying the same thing. This is, of course, aside the > > > fact that you wouldn't find anybody who is active in the forum and > > > who doesn't have a personal agenda. We all have one and that's why > > > we are engaged. The people with a double role just have to keep it > > > separate - and that's not only true for the chair, but (perhaps to a > > > lesser extent) also for the other roles that Jeff pointed out. > > > > > > I also agree with Jeff - if we are setting up a new structure, let's > > > look at all roles and define them better. > > > > > >> As to logistics, yes, I see your point but I still think that > > >> combining them would generally simplify things. I also think it is > > >> fine for the chair to delegate aspects that the convener handles > > >> but they should still be responsible for them (we all blame the > > >> chair if they do not go right so they should also have the > > >> authority since they de facto have responsibility). > > > > > > The convener role will always have to be split up, in particular if > > > meetings get hosted at individual organizations. A certain part of > > > the work (e.g., wireless, badging, room reservation, ...) always > > > fell on people from the hosting organization and that won't change. > > > > > > The more critical question, IMHO, is the financial part that Jeff > > > brought up. We have organizations that have more restrictions than > > > others and we have to take that into account. Further, if the person > > > responsible for financial transactions changes, we have to transfer > > > money - the same if the person responsible changes organization (as > > > we can see at Rich's example, which ended up being non trivial or > > > even impossible, if I understood the situation right). The easiest > > > thing would be to have a constant entity who manages all financial > > > aspects of the forum along with a separate role of a treasurer. > > > > > > Martin > > > > > > > > >> > > >> Bronis > > >> > > >> > > >> On Wed, 26 Sep 2012, Jeff Squyres wrote: > > >> > > >>> This is exactly why we should have the discussion: "chair" and > > >>> "convener" have not necessarily been the same person over the last > > >>> few years. Rich has been the convener since 2.1, but has only been > > >>> the chair for 3.0. > > >>> > > >>> My $0.02 is that I don't think they need to be the same person. > > >>> The chair is the [neutral] technical leader (and therefore sets > > >>> the meeting agendas), but the convener is someone with an > > >>> organization who can make meeting room reservations, accept > > >>> registration and payments from meeting attendees, etc. That being > > >>> said, as long as the chair remains neutral in setting the agenda > > >>> and guiding the process, the chair should be allowed to have their > > >>> own technical opinions, just like anyone else. Otherwise, you're > > >>> just neutering the chair, and their incentive to participate is > > >>> quite limited. > > >>> > > >>> In short: I see the convener as a largely logistical role, but the > > >>> chair as more of a technical role. > > >>> > > >>> > > >>> > > >>> On Sep 27, 2012, at 6:33 AM, Bronis R. de Supinski wrote: > > >>> > > >>>> > > >>>> Personally, I think "chair" should also be "meeting convener". > > >>>> The chair should be largely administrative/neutral/meeting > > >>>> organizer and runner. Convening meetings is a natural part of > > >>>> that activity. The chair should not be promoting a particular > > >>>> technical agenda (at least not as "chair"). > > >>>> > > >>>> > > >>>> On Wed, 26 Sep 2012, Jeff Squyres wrote: > > >>>> > > >>>>> As part of this discussion, it might be worthwhile to also bring > > >>>>> up the auxiliary roles -- there's a bunch of things that happen > > >>>>> behind the scenes to make the Forum function: > > >>>>> > > >>>>> - secretary > > >>>>> - meeting convener > > >>>>> - web maintainers (main site, meetings site, lists site) > > >>>>> - svn / trac maintainer > > >>>>> > > >>>>> I mention this because, although I'm happy to continue in the > > >>>>> role as secretary, a) if there's someone else burning to do that > > >>>>> role, it would be fair to discuss it, and b) sometimes there has > > >>>>> been confusion about exactly what the secretary is supposed to > > >>>>> do. Maybe it would be useful to delineate the responsibilities > > >>>>> of the auxiliary roles. Just like we discovered that we all > > >>>>> didn't have the same conceptions of how the secretary applies > > >>>>> the voting rules, perhaps there are other unwritten > > >>>>> inconsistencies, too. > > >>>>> > > >>>>> > > >>>>> > > >>>>> On Sep 26, 2012, at 10:27 PM, Brightwell, Ronald wrote: > > >>>>> > > >>>>>> > > >>>>>> Greetings: > > >>>>>> > > >>>>>> I wasn't able to attend the Forum meeting last week, so I don't > > >>>>>> know if this came up or not, but given that it's election > > >>>>>> season in the US and MPI 3.0 has been officially accepted, it > > >>>>>> seems like a good time to discuss leadership of the Forum going > > >>>>>> forward. I know several people who have been involved from a > > >>>>>> leadership standpoint are in different situations now then they > > >>>>>> were when the 3.0 process started, and their priorities and > > >>>>>> commitments are different. I think it might be good to have an > > >>>>>> open discussion about making changes to some of the leadership > > >>>>>> roles in an effort to make sure that we start this next phase > > >>>>>> in the best way possible. > > >>>>>> > > >>>>>> -Ron > > >>>>>> > > >>>>>> > > >>>>>> > > >>>>>> _______________________________________________ > > >>>>>> mpi-forum mailing list > > >>>>>> mpi-forum(a)lists.mpi-forum.org > > >>>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > >>>>> > > >>>>> > > >>>>> -- > > >>>>> Jeff Squyres > > >>>>> jsquyres(a)cisco.com > > >>>>> For corporate legal information go to: > > >>>>> http://www.cisco.com/web/about/doing_business/legal/cri/ > > >>>>> > > >>>>> > > >>>>> _______________________________________________ > > >>>>> mpi-forum mailing list > > >>>>> mpi-forum(a)lists.mpi-forum.org > > >>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > >>>>> > > >>>> _______________________________________________ > > >>>> mpi-forum mailing list > > >>>> mpi-forum(a)lists.mpi-forum.org > > >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > >>> > > >>> > > >>> -- > > >>> Jeff Squyres > > >>> jsquyres(a)cisco.com > > >>> For corporate legal information go to: > > >>> http://www.cisco.com/web/about/doing_business/legal/cri/ > > >>> > > >>> > > >> _______________________________________________ > > >> mpi-forum mailing list > > >> mpi-forum(a)lists.mpi-forum.org > > >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > > > > > > __________________________________________________________ > __________ > > > ____ Martin Schulz, schulzm(a)llnl.gov, http://people.llnl.gov/schulzm > > > CASC @ Lawrence Livermore National Laboratory, Livermore, USA > > > > > > > > > > > > > > > _______________________________________________ > > > mpi-forum mailing list > > > mpi-forum(a)lists.mpi-forum.org > > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > > > > > _______________________________________________ > > 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: Allmandring 30) > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > -- > Anthony Skjellum, PhD > Professor and Chair > Dept. of Computer and Information Sciences Director, UAB Center for > Information Assurance and Joint Forensics Research ("The Center") > University of Alabama at Birmingham > +1-(205)934-8657; FAX: +1- (205)934-5473 > > ___________________________________________ > 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. > > Please consider the environment before printing this e-mail > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum

1 0

Re: [Mpi-forum] Meeting schedule for 2013
by Bronis R. de Supinski 27 Sep '12

27 Sep '12

Rolf: Re: >> even years: 3US+euroMPI >> odd years: 2US,1Asia+EuroMPI. > > Yes, this was the last position at the Forum and also discussed > at the EuroMPI business meeting. As I understand the situation, this option was not discussed at the Forum. As stated previously, the matter was decided in Chicago and reopening it without an announcement was inappropriate. In addition, what is discussed at the EuroMPI business meeting is not relevant. It is not their decision. > The MPI Forum was re-started by the activities at several EuroMPI > conferences, the last one at Paris 2007. > > Therefore, there is a strong relationship and resposibility > and to have the MPI Forum meeting in US and the EuroMPI in Europe > both in fall, would not reflect this responsibility. > > As far as I see, the colleagues in Madrid would like to host the > MPI Forum meeting together with the EuroMPI and therefore, > I expect - based on the previous meeting rules - we will have > > MPI-next Forum meeting Sep. 11-13, 2013, > (instead of Sep. 9-11 in San Jose), and > EuroMPI, Sep. 15-18, 2013, both in Madrid, Spain > > This may change in later years if the EuroMPI may go to Asia, > but this is another discussion within the EuroMPI organizers. What I hear is that holding the MPI Forum meeting separately from EuroMPI would damage EuroMPI. Well, forcing the Forum to co-locate is putting the cart before the horse. If EuroMPI cannot withstand a separate meeting, as decided by the Forum, then EuroMPI needs to adapt -- not the other way around. As was previously suggested, EuroMPI is free to rebrand as International MPI and to co-locate with Forum meetings. If the conference demonstrated that flexibility, I would be more willing to consider their desires as an MPI Forum voting member. Bronis > Best regards > Rolf > > > ----- Original Message ----- >> From: "Aurélien Bouteiller" <bouteill(a)eecs.utk.edu> >> To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> >> Sent: Thursday, September 27, 2012 2:41:08 AM >> Subject: Re: [Mpi-forum] Meeting schedule for 2013 >> The 3-1 with 1 alternate between asia and europe collected the most >> vote. However there was significant disappointment about not >> collocating a meeting with EuroMPI, and after the straw vote closed, >> another interesting possibility was envisioned: >> >> even years: 3US+euroMPI >> odd years: 2US,1Asia+EuroMPI. >> >> This last proposal has not been straw voted, so it is not possible to >> know if it has support, but it seemed from the reception it had that >> the idea is not unbearable for most US institutions. >> >> Aurelien >> >> Le 26 sept. 2012 à 15:17, Jeff Squyres <jsquyres(a)cisco.com> a écrit : >> >> > I don't remember there being a majority for the 2-2 option here in >> > Vienna -- am I remembering last week incorrectly? >> > >> > I definitely can't do 2 international meetings next week. >> > >> > >> > >> > >> > On Sep 26, 2012, at 8:02 PM, Richard Graham wrote: >> > >> >> This was discussed and there was another vote at the meeting in >> >> Vienna. This note is just a follow up to the discussions at that >> >> meeting. >> >> >> >> Rich >> >> >> >> -----Original Message----- >> >> From: mpi-forum-bounces(a)lists.mpi-forum.org >> >> [mailto:[email protected]] On Behalf Of Schulz, >> >> Martin >> >> Sent: Wednesday, September 26, 2012 7:36 PM >> >> To: Bronis R. de Supinski; Main MPI Forum mailing list >> >> Subject: Re: [Mpi-forum] Meeting schedule for 2013 >> >> >> >> Hi Rich, >> >> >> >> I have the same recollection as Bronis and Mohamad. I thought there >> >> was a strong majority (at least at the Chicago meeting) for a 3-1 >> >> solution with alternating non-US locations (one year Asia, one year >> >> Europe, but not both in one year - as it also shows on the current >> >> tentative meeting calendar for next year). I know this is not fair >> >> to everyone, but the group (again, at least in Chicago, which was >> >> pretty complete, though, since we were wrapping up 3.0) felt that >> >> this would maximize active participation. >> >> >> >> Martin >> >> >> >> >> >> On Sep 26, 2012, at 9:39 AM, Bronis R. de Supinski wrote: >> >> >> >>> >> >>> Rich: >> >>> >> >>> I would have to disagree. At best, you are splitting hairs to say >> >>> this >> >>> option was not voted on. In the "alternate" >> >>> years, it is simply the 2-2 option that was voted on and did not >> >>> come >> >>> anywhere near the 3-1 option in terms of votes. >> >>> >> >>> The arguments against 2-2 (even in alternating years) still apply. >> >>> It >> >>> significantly lowers attendance by active participants. >> >>> >> >>> Bronis >> >>> >> >>> >> >>> >> >>> >> >>> On Wed, 26 Sep 2012, Richard Graham wrote: >> >>> >> >>>> Such decisions are not made lightly - I assure you this. This was >> >>>> discussed at the last meeting, and this was an option that was >> >>>> not voted on. However, this was discussed again at the end of the >> >>>> meeting, and some at EuroMPI. >> >>>> >> >>>> Rich >> >>>> >> >>>> -----Original Message----- >> >>>> From: mpi-forum-bounces(a)lists.mpi-forum.org >> >>>> [mailto:[email protected]] On Behalf Of >> >>>> Mohamad >> >>>> Chaarawi >> >>>> Sent: Wednesday, September 26, 2012 6:30 PM >> >>>> To: mpi-forum(a)lists.mpi-forum.org >> >>>> Subject: Re: [Mpi-forum] Meeting schedule for 2013 >> >>>> >> >>>> Hi Rich, >> >>>> >> >>>> On 9/26/2012 10:30 AM, Richard Graham wrote: >> >>>>> Therefore, I propose to keep the European meeting of the Forum >> >>>>> in >> >>>>> conjunction with EuroMPI, and every other year also have a >> >>>>> meeting >> >>>>> in Japan. With this in mind, here is the proposed schedule for >> >>>>> 2012 >> >>>> >> >>>> The majority at the last meeting in Chicago were against having 2 >> >>>> meetings internationally. We had several votes on different >> >>>> formations (4-0, 3-1, and 2-2), and 3-1 was the majority vote. >> >>>> >> >>>> It would have been fair to have a similar vote in Europe and see >> >>>> how the participants vote.. >> >>>> Did the attendance at the Vienna meeting vote on a different >> >>>> schedule, or is your change coming from just some concerns (which >> >>>> are also valid) from some people? >> >>>> >> >>>> Can we have a formal organizational Vote by eligible >> >>>> organizations on this instead of just scheduling meetings based >> >>>> on suggestions that vary from meeting to meeting? >> >>>> This vote can be done over the phone. >> >>>> >> >>>> Thanks, >> >>>> Mohamad >> >>>> _______________________________________________ >> >>>> mpi-forum mailing list >> >>>> mpi-forum(a)lists.mpi-forum.org >> >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >> >>>> >> >>>> _______________________________________________ >> >>>> mpi-forum mailing list >> >>>> mpi-forum(a)lists.mpi-forum.org >> >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >> >>>> >> >>> _______________________________________________ >> >>> mpi-forum mailing list >> >>> mpi-forum(a)lists.mpi-forum.org >> >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >> >> >> >> ________________________________________________________________________ >> >> Martin Schulz, schulzm(a)llnl.gov, http://people.llnl.gov/schulzm >> >> CASC @ Lawrence Livermore National Laboratory, Livermore, USA >> >> >> >> >> >> >> >> >> >> _______________________________________________ >> >> mpi-forum mailing list >> >> mpi-forum(a)lists.mpi-forum.org >> >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >> >> >> >> _______________________________________________ >> >> mpi-forum mailing list >> >> mpi-forum(a)lists.mpi-forum.org >> >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >> > >> > >> > -- >> > Jeff Squyres >> > jsquyres(a)cisco.com >> > For corporate legal information go to: >> > http://www.cisco.com/web/about/doing_business/legal/cri/ >> > >> > >> > _______________________________________________ >> > mpi-forum mailing list >> > mpi-forum(a)lists.mpi-forum.org >> > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >> >> -- >> * Dr. Aurélien Bouteiller >> * Researcher at Innovative Computing Laboratory >> * University of Tennessee >> * 1122 Volunteer Boulevard, suite 309b >> * Knoxville, TN 37996 >> * 865 974 9375 >> >> >> >> >> >> >> >> >> _______________________________________________ >> 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: Allmandring 30) > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum

1 0

Re: [Mpi-forum] MPI Next leadership
by Anthony Skjellum 27 Sep '12

27 Sep '12

Besides leadership, we should ask "why" MPI Next "now" vs "later" vs. "ever." According to Robert's Rules of Order, a motion to adjourn is always in order. Why are we assuming that MPI goes on as before, with multiyear planning? We should first entertain, second, and either accept or defeat a motion to adjourn, and those voting against adjourning should provide rationale for continuing "now." A second vote could be to "wait X months/years" and then automatically return. For instance, why not give the OpenMPI and MPICH development teams a chance to finish their work, distribute releases, write up findings, discover errata, tune, let people get trained, let applications be ported and use those new features, achieve buy-in and acceptance, and have actual users give feedback, so we come tothe "Next" with the experience of results? In the meantime, research on the areas that are still controversial could continue without the need to "be standardized" at the same time as they are being figured out. ----- Original Message ----- From: "Rolf Rabenseifner" <rabenseifner(a)hlrs.de> To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> Sent: Thursday, September 27, 2012 5:17:15 AM Subject: Re: [Mpi-forum] MPI Next leadership As MPI-2.1 chair, I must clearly say that I could not have been the MPI-2.1 chair if there would not have been Rich as the convener of the meetings. And I'm thankful that Rich has done it the whole time together with the local organizers, mainly Jeff, the team of Microsoft and all European organizers at the EuroMPI conferences. As far as I understand, Rich would continue to be our convener. About MPI-next chair: Although the chair should clearly be neutral, I would prefer to have a MPI-next chair from a public research institution. Of course he/she must have the support from her/his institution to travel to the meetings. I see also that he/she need not only to focus on the new topics, but also on the quality of the standard, i.e., after the time and content is right, to organizing also the finalization of such an new version of MPI. Rolf ----- Original Message ----- > From: "Rajeev Thakur" <thakur(a)mcs.anl.gov> > To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> > Sent: Thursday, September 27, 2012 10:51:45 AM > Subject: Re: [Mpi-forum] MPI Next leadership > It has also been suggested that the Forum take a break for a bit > before resuming again, but I don't know if that was discussed at the > last meeting. > > Rajeev > > > > On Sep 27, 2012, at 1:09 AM, Schulz, Martin wrote: > > > On Sep 26, 2012, at 9:46 PM, Bronis R. de Supinski wrote: > > > >> > >> Jeff: > >> > >> My intent is not to "neuter" the chair. Instead, I am only > >> stating that the chair should clearly state when (s)he is > >> voicing personal/institutional opinions as opposed to acting > >> as leader (sometimes the roles conflict and also sometimes > >> the leader does have to enforce the clear general opinion). > > > > I think you are saying the same thing. This is, of course, aside the > > fact that you wouldn't find anybody who is active in the forum and > > who doesn't have a personal agenda. We all have one and that's why > > we are engaged. The people with a double role just have to keep it > > separate - and that's not only true for the chair, but (perhaps to a > > lesser extent) also for the other roles that Jeff pointed out. > > > > I also agree with Jeff - if we are setting up a new structure, let's > > look at all roles and define them better. > > > >> As to logistics, yes, I see your point but I still think > >> that combining them would generally simplify things. I > >> also think it is fine for the chair to delegate aspects > >> that the convener handles but they should still be > >> responsible for them (we all blame the chair if they do > >> not go right so they should also have the authority since > >> they de facto have responsibility). > > > > The convener role will always have to be split up, in particular if > > meetings get hosted at individual organizations. A certain part of > > the work (e.g., wireless, badging, room reservation, ...) always > > fell on people from the hosting organization and that won't change. > > > > The more critical question, IMHO, is the financial part that Jeff > > brought up. We have organizations that have more restrictions than > > others and we have to take that into account. Further, if the person > > responsible for financial transactions changes, we have to transfer > > money - the same if the person responsible changes organization (as > > we can see at Rich's example, which ended up being non trivial or > > even impossible, if I understood the situation right). The easiest > > thing would be to have a constant entity who manages all financial > > aspects of the forum along with a separate role of a treasurer. > > > > Martin > > > > > >> > >> Bronis > >> > >> > >> On Wed, 26 Sep 2012, Jeff Squyres wrote: > >> > >>> This is exactly why we should have the discussion: "chair" and > >>> "convener" have not necessarily been the same person over the last > >>> few years. Rich has been the convener since 2.1, but has only been > >>> the chair for 3.0. > >>> > >>> My $0.02 is that I don't think they need to be the same person. > >>> The chair is the [neutral] technical leader (and therefore sets > >>> the meeting agendas), but the convener is someone with an > >>> organization who can make meeting room reservations, accept > >>> registration and payments from meeting attendees, etc. That being > >>> said, as long as the chair remains neutral in setting the agenda > >>> and guiding the process, the chair should be allowed to have their > >>> own technical opinions, just like anyone else. Otherwise, you're > >>> just neutering the chair, and their incentive to participate is > >>> quite limited. > >>> > >>> In short: I see the convener as a largely logistical role, but the > >>> chair as more of a technical role. > >>> > >>> > >>> > >>> On Sep 27, 2012, at 6:33 AM, Bronis R. de Supinski wrote: > >>> > >>>> > >>>> Personally, I think "chair" should also be "meeting convener". > >>>> The chair should be largely administrative/neutral/meeting > >>>> organizer and runner. Convening meetings is a natural part > >>>> of that activity. The chair should not be promoting a > >>>> particular technical agenda (at least not as "chair"). > >>>> > >>>> > >>>> On Wed, 26 Sep 2012, Jeff Squyres wrote: > >>>> > >>>>> As part of this discussion, it might be worthwhile to also bring > >>>>> up the auxiliary roles -- there's a bunch of things that happen > >>>>> behind the scenes to make the Forum function: > >>>>> > >>>>> - secretary > >>>>> - meeting convener > >>>>> - web maintainers (main site, meetings site, lists site) > >>>>> - svn / trac maintainer > >>>>> > >>>>> I mention this because, although I'm happy to continue in the > >>>>> role as secretary, a) if there's someone else burning to do that > >>>>> role, it would be fair to discuss it, and b) sometimes there has > >>>>> been confusion about exactly what the secretary is supposed to > >>>>> do. Maybe it would be useful to delineate the responsibilities > >>>>> of the auxiliary roles. Just like we discovered that we all > >>>>> didn't have the same conceptions of how the secretary applies > >>>>> the voting rules, perhaps there are other unwritten > >>>>> inconsistencies, too. > >>>>> > >>>>> > >>>>> > >>>>> On Sep 26, 2012, at 10:27 PM, Brightwell, Ronald wrote: > >>>>> > >>>>>> > >>>>>> Greetings: > >>>>>> > >>>>>> I wasn't able to attend the Forum meeting last week, so I don't > >>>>>> know if this came up or not, but given that it's election > >>>>>> season in the US and MPI 3.0 has been officially accepted, it > >>>>>> seems like a good time to discuss leadership of the Forum going > >>>>>> forward. I know several people who have been involved from a > >>>>>> leadership standpoint are in different situations now then they > >>>>>> were when the 3.0 process started, and their priorities and > >>>>>> commitments are different. I think it might be good to have an > >>>>>> open discussion about making changes to some of the leadership > >>>>>> roles in an effort to make sure that we start this next phase > >>>>>> in the best way possible. > >>>>>> > >>>>>> -Ron > >>>>>> > >>>>>> > >>>>>> > >>>>>> _______________________________________________ > >>>>>> mpi-forum mailing list > >>>>>> mpi-forum(a)lists.mpi-forum.org > >>>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > >>>>> > >>>>> > >>>>> -- > >>>>> Jeff Squyres > >>>>> jsquyres(a)cisco.com > >>>>> For corporate legal information go to: > >>>>> http://www.cisco.com/web/about/doing_business/legal/cri/ > >>>>> > >>>>> > >>>>> _______________________________________________ > >>>>> mpi-forum mailing list > >>>>> mpi-forum(a)lists.mpi-forum.org > >>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > >>>>> > >>>> _______________________________________________ > >>>> mpi-forum mailing list > >>>> mpi-forum(a)lists.mpi-forum.org > >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > >>> > >>> > >>> -- > >>> Jeff Squyres > >>> jsquyres(a)cisco.com > >>> For corporate legal information go to: > >>> http://www.cisco.com/web/about/doing_business/legal/cri/ > >>> > >>> > >> _______________________________________________ > >> mpi-forum mailing list > >> mpi-forum(a)lists.mpi-forum.org > >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > > > ________________________________________________________________________ > > Martin Schulz, schulzm(a)llnl.gov, http://people.llnl.gov/schulzm > > CASC @ Lawrence Livermore National Laboratory, Livermore, USA > > > > > > > > > > _______________________________________________ > > mpi-forum mailing list > > mpi-forum(a)lists.mpi-forum.org > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > > _______________________________________________ > 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: Allmandring 30) _______________________________________________ mpi-forum mailing list mpi-forum(a)lists.mpi-forum.org http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum -- Anthony Skjellum, PhD Professor and Chair Dept. of Computer and Information Sciences Director, UAB Center for Information Assurance and Joint Forensics Research ("The Center") University of Alabama at Birmingham +1-(205)934-8657; FAX: +1- (205)934-5473 ___________________________________________ 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. Please consider the environment before printing this e-mail

1 0

Re: [Mpi-forum] Meeting schedule for 2013
by Rolf Rabenseifner 27 Sep '12

27 Sep '12

> even years: 3US+euroMPI > odd years: 2US,1Asia+EuroMPI. Yes, this was the last position at the Forum and also discussed at the EuroMPI business meeting. The MPI Forum was re-started by the activities at several EuroMPI conferences, the last one at Paris 2007. Therefore, there is a strong relationship and resposibility and to have the MPI Forum meeting in US and the EuroMPI in Europe both in fall, would not reflect this responsibility. As far as I see, the colleagues in Madrid would like to host the MPI Forum meeting together with the EuroMPI and therefore, I expect - based on the previous meeting rules - we will have MPI-next Forum meeting Sep. 11-13, 2013, (instead of Sep. 9-11 in San Jose), and EuroMPI, Sep. 15-18, 2013, both in Madrid, Spain This may change in later years if the EuroMPI may go to Asia, but this is another discussion within the EuroMPI organizers. Best regards Rolf ----- Original Message ----- > From: "Aurélien Bouteiller" <bouteill(a)eecs.utk.edu> > To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> > Sent: Thursday, September 27, 2012 2:41:08 AM > Subject: Re: [Mpi-forum] Meeting schedule for 2013 > The 3-1 with 1 alternate between asia and europe collected the most > vote. However there was significant disappointment about not > collocating a meeting with EuroMPI, and after the straw vote closed, > another interesting possibility was envisioned: > > even years: 3US+euroMPI > odd years: 2US,1Asia+EuroMPI. > > This last proposal has not been straw voted, so it is not possible to > know if it has support, but it seemed from the reception it had that > the idea is not unbearable for most US institutions. > > Aurelien > > Le 26 sept. 2012 à 15:17, Jeff Squyres <jsquyres(a)cisco.com> a écrit : > > > I don't remember there being a majority for the 2-2 option here in > > Vienna -- am I remembering last week incorrectly? > > > > I definitely can't do 2 international meetings next week. > > > > > > > > > > On Sep 26, 2012, at 8:02 PM, Richard Graham wrote: > > > >> This was discussed and there was another vote at the meeting in > >> Vienna. This note is just a follow up to the discussions at that > >> meeting. > >> > >> Rich > >> > >> -----Original Message----- > >> From: mpi-forum-bounces(a)lists.mpi-forum.org > >> [mailto:[email protected]] On Behalf Of Schulz, > >> Martin > >> Sent: Wednesday, September 26, 2012 7:36 PM > >> To: Bronis R. de Supinski; Main MPI Forum mailing list > >> Subject: Re: [Mpi-forum] Meeting schedule for 2013 > >> > >> Hi Rich, > >> > >> I have the same recollection as Bronis and Mohamad. I thought there > >> was a strong majority (at least at the Chicago meeting) for a 3-1 > >> solution with alternating non-US locations (one year Asia, one year > >> Europe, but not both in one year - as it also shows on the current > >> tentative meeting calendar for next year). I know this is not fair > >> to everyone, but the group (again, at least in Chicago, which was > >> pretty complete, though, since we were wrapping up 3.0) felt that > >> this would maximize active participation. > >> > >> Martin > >> > >> > >> On Sep 26, 2012, at 9:39 AM, Bronis R. de Supinski wrote: > >> > >>> > >>> Rich: > >>> > >>> I would have to disagree. At best, you are splitting hairs to say > >>> this > >>> option was not voted on. In the "alternate" > >>> years, it is simply the 2-2 option that was voted on and did not > >>> come > >>> anywhere near the 3-1 option in terms of votes. > >>> > >>> The arguments against 2-2 (even in alternating years) still apply. > >>> It > >>> significantly lowers attendance by active participants. > >>> > >>> Bronis > >>> > >>> > >>> > >>> > >>> On Wed, 26 Sep 2012, Richard Graham wrote: > >>> > >>>> Such decisions are not made lightly - I assure you this. This was > >>>> discussed at the last meeting, and this was an option that was > >>>> not voted on. However, this was discussed again at the end of the > >>>> meeting, and some at EuroMPI. > >>>> > >>>> Rich > >>>> > >>>> -----Original Message----- > >>>> From: mpi-forum-bounces(a)lists.mpi-forum.org > >>>> [mailto:[email protected]] On Behalf Of > >>>> Mohamad > >>>> Chaarawi > >>>> Sent: Wednesday, September 26, 2012 6:30 PM > >>>> To: mpi-forum(a)lists.mpi-forum.org > >>>> Subject: Re: [Mpi-forum] Meeting schedule for 2013 > >>>> > >>>> Hi Rich, > >>>> > >>>> On 9/26/2012 10:30 AM, Richard Graham wrote: > >>>>> Therefore, I propose to keep the European meeting of the Forum > >>>>> in > >>>>> conjunction with EuroMPI, and every other year also have a > >>>>> meeting > >>>>> in Japan. With this in mind, here is the proposed schedule for > >>>>> 2012 > >>>> > >>>> The majority at the last meeting in Chicago were against having 2 > >>>> meetings internationally. We had several votes on different > >>>> formations (4-0, 3-1, and 2-2), and 3-1 was the majority vote. > >>>> > >>>> It would have been fair to have a similar vote in Europe and see > >>>> how the participants vote.. > >>>> Did the attendance at the Vienna meeting vote on a different > >>>> schedule, or is your change coming from just some concerns (which > >>>> are also valid) from some people? > >>>> > >>>> Can we have a formal organizational Vote by eligible > >>>> organizations on this instead of just scheduling meetings based > >>>> on suggestions that vary from meeting to meeting? > >>>> This vote can be done over the phone. > >>>> > >>>> Thanks, > >>>> Mohamad > >>>> _______________________________________________ > >>>> mpi-forum mailing list > >>>> mpi-forum(a)lists.mpi-forum.org > >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > >>>> > >>>> _______________________________________________ > >>>> mpi-forum mailing list > >>>> mpi-forum(a)lists.mpi-forum.org > >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > >>>> > >>> _______________________________________________ > >>> mpi-forum mailing list > >>> mpi-forum(a)lists.mpi-forum.org > >>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > >> > >> ________________________________________________________________________ > >> Martin Schulz, schulzm(a)llnl.gov, http://people.llnl.gov/schulzm > >> CASC @ Lawrence Livermore National Laboratory, Livermore, USA > >> > >> > >> > >> > >> _______________________________________________ > >> mpi-forum mailing list > >> mpi-forum(a)lists.mpi-forum.org > >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > >> > >> _______________________________________________ > >> mpi-forum mailing list > >> mpi-forum(a)lists.mpi-forum.org > >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > > > > > -- > > Jeff Squyres > > jsquyres(a)cisco.com > > For corporate legal information go to: > > http://www.cisco.com/web/about/doing_business/legal/cri/ > > > > > > _______________________________________________ > > mpi-forum mailing list > > mpi-forum(a)lists.mpi-forum.org > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > -- > * Dr. Aurélien Bouteiller > * Researcher at Innovative Computing Laboratory > * University of Tennessee > * 1122 Volunteer Boulevard, suite 309b > * Knoxville, TN 37996 > * 865 974 9375 > > > > > > > > > _______________________________________________ > 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: Allmandring 30)

1 0

Re: [Mpi-forum] MPI Next leadership
by Rolf Rabenseifner 27 Sep '12

27 Sep '12

As MPI-2.1 chair, I must clearly say that I could not have been the MPI-2.1 chair if there would not have been Rich as the convener of the meetings. And I'm thankful that Rich has done it the whole time together with the local organizers, mainly Jeff, the team of Microsoft and all European organizers at the EuroMPI conferences. As far as I understand, Rich would continue to be our convener. About MPI-next chair: Although the chair should clearly be neutral, I would prefer to have a MPI-next chair from a public research institution. Of course he/she must have the support from her/his institution to travel to the meetings. I see also that he/she need not only to focus on the new topics, but also on the quality of the standard, i.e., after the time and content is right, to organizing also the finalization of such an new version of MPI. Rolf ----- Original Message ----- > From: "Rajeev Thakur" <thakur(a)mcs.anl.gov> > To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> > Sent: Thursday, September 27, 2012 10:51:45 AM > Subject: Re: [Mpi-forum] MPI Next leadership > It has also been suggested that the Forum take a break for a bit > before resuming again, but I don't know if that was discussed at the > last meeting. > > Rajeev > > > > On Sep 27, 2012, at 1:09 AM, Schulz, Martin wrote: > > > On Sep 26, 2012, at 9:46 PM, Bronis R. de Supinski wrote: > > > >> > >> Jeff: > >> > >> My intent is not to "neuter" the chair. Instead, I am only > >> stating that the chair should clearly state when (s)he is > >> voicing personal/institutional opinions as opposed to acting > >> as leader (sometimes the roles conflict and also sometimes > >> the leader does have to enforce the clear general opinion). > > > > I think you are saying the same thing. This is, of course, aside the > > fact that you wouldn't find anybody who is active in the forum and > > who doesn't have a personal agenda. We all have one and that's why > > we are engaged. The people with a double role just have to keep it > > separate - and that's not only true for the chair, but (perhaps to a > > lesser extent) also for the other roles that Jeff pointed out. > > > > I also agree with Jeff - if we are setting up a new structure, let's > > look at all roles and define them better. > > > >> As to logistics, yes, I see your point but I still think > >> that combining them would generally simplify things. I > >> also think it is fine for the chair to delegate aspects > >> that the convener handles but they should still be > >> responsible for them (we all blame the chair if they do > >> not go right so they should also have the authority since > >> they de facto have responsibility). > > > > The convener role will always have to be split up, in particular if > > meetings get hosted at individual organizations. A certain part of > > the work (e.g., wireless, badging, room reservation, ...) always > > fell on people from the hosting organization and that won't change. > > > > The more critical question, IMHO, is the financial part that Jeff > > brought up. We have organizations that have more restrictions than > > others and we have to take that into account. Further, if the person > > responsible for financial transactions changes, we have to transfer > > money - the same if the person responsible changes organization (as > > we can see at Rich's example, which ended up being non trivial or > > even impossible, if I understood the situation right). The easiest > > thing would be to have a constant entity who manages all financial > > aspects of the forum along with a separate role of a treasurer. > > > > Martin > > > > > >> > >> Bronis > >> > >> > >> On Wed, 26 Sep 2012, Jeff Squyres wrote: > >> > >>> This is exactly why we should have the discussion: "chair" and > >>> "convener" have not necessarily been the same person over the last > >>> few years. Rich has been the convener since 2.1, but has only been > >>> the chair for 3.0. > >>> > >>> My $0.02 is that I don't think they need to be the same person. > >>> The chair is the [neutral] technical leader (and therefore sets > >>> the meeting agendas), but the convener is someone with an > >>> organization who can make meeting room reservations, accept > >>> registration and payments from meeting attendees, etc. That being > >>> said, as long as the chair remains neutral in setting the agenda > >>> and guiding the process, the chair should be allowed to have their > >>> own technical opinions, just like anyone else. Otherwise, you're > >>> just neutering the chair, and their incentive to participate is > >>> quite limited. > >>> > >>> In short: I see the convener as a largely logistical role, but the > >>> chair as more of a technical role. > >>> > >>> > >>> > >>> On Sep 27, 2012, at 6:33 AM, Bronis R. de Supinski wrote: > >>> > >>>> > >>>> Personally, I think "chair" should also be "meeting convener". > >>>> The chair should be largely administrative/neutral/meeting > >>>> organizer and runner. Convening meetings is a natural part > >>>> of that activity. The chair should not be promoting a > >>>> particular technical agenda (at least not as "chair"). > >>>> > >>>> > >>>> On Wed, 26 Sep 2012, Jeff Squyres wrote: > >>>> > >>>>> As part of this discussion, it might be worthwhile to also bring > >>>>> up the auxiliary roles -- there's a bunch of things that happen > >>>>> behind the scenes to make the Forum function: > >>>>> > >>>>> - secretary > >>>>> - meeting convener > >>>>> - web maintainers (main site, meetings site, lists site) > >>>>> - svn / trac maintainer > >>>>> > >>>>> I mention this because, although I'm happy to continue in the > >>>>> role as secretary, a) if there's someone else burning to do that > >>>>> role, it would be fair to discuss it, and b) sometimes there has > >>>>> been confusion about exactly what the secretary is supposed to > >>>>> do. Maybe it would be useful to delineate the responsibilities > >>>>> of the auxiliary roles. Just like we discovered that we all > >>>>> didn't have the same conceptions of how the secretary applies > >>>>> the voting rules, perhaps there are other unwritten > >>>>> inconsistencies, too. > >>>>> > >>>>> > >>>>> > >>>>> On Sep 26, 2012, at 10:27 PM, Brightwell, Ronald wrote: > >>>>> > >>>>>> > >>>>>> Greetings: > >>>>>> > >>>>>> I wasn't able to attend the Forum meeting last week, so I don't > >>>>>> know if this came up or not, but given that it's election > >>>>>> season in the US and MPI 3.0 has been officially accepted, it > >>>>>> seems like a good time to discuss leadership of the Forum going > >>>>>> forward. I know several people who have been involved from a > >>>>>> leadership standpoint are in different situations now then they > >>>>>> were when the 3.0 process started, and their priorities and > >>>>>> commitments are different. I think it might be good to have an > >>>>>> open discussion about making changes to some of the leadership > >>>>>> roles in an effort to make sure that we start this next phase > >>>>>> in the best way possible. > >>>>>> > >>>>>> -Ron > >>>>>> > >>>>>> > >>>>>> > >>>>>> _______________________________________________ > >>>>>> mpi-forum mailing list > >>>>>> mpi-forum(a)lists.mpi-forum.org > >>>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > >>>>> > >>>>> > >>>>> -- > >>>>> Jeff Squyres > >>>>> jsquyres(a)cisco.com > >>>>> For corporate legal information go to: > >>>>> http://www.cisco.com/web/about/doing_business/legal/cri/ > >>>>> > >>>>> > >>>>> _______________________________________________ > >>>>> mpi-forum mailing list > >>>>> mpi-forum(a)lists.mpi-forum.org > >>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > >>>>> > >>>> _______________________________________________ > >>>> mpi-forum mailing list > >>>> mpi-forum(a)lists.mpi-forum.org > >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > >>> > >>> > >>> -- > >>> Jeff Squyres > >>> jsquyres(a)cisco.com > >>> For corporate legal information go to: > >>> http://www.cisco.com/web/about/doing_business/legal/cri/ > >>> > >>> > >> _______________________________________________ > >> mpi-forum mailing list > >> mpi-forum(a)lists.mpi-forum.org > >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > > > ________________________________________________________________________ > > Martin Schulz, schulzm(a)llnl.gov, http://people.llnl.gov/schulzm > > CASC @ Lawrence Livermore National Laboratory, Livermore, USA > > > > > > > > > > _______________________________________________ > > mpi-forum mailing list > > mpi-forum(a)lists.mpi-forum.org > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > > _______________________________________________ > 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: Allmandring 30)

1 0

Re: [Mpi-forum] MPI Next leadership
by Rajeev Thakur 27 Sep '12

27 Sep '12

It has also been suggested that the Forum take a break for a bit before resuming again, but I don't know if that was discussed at the last meeting. Rajeev On Sep 27, 2012, at 1:09 AM, Schulz, Martin wrote: > On Sep 26, 2012, at 9:46 PM, Bronis R. de Supinski wrote: > >> >> Jeff: >> >> My intent is not to "neuter" the chair. Instead, I am only >> stating that the chair should clearly state when (s)he is >> voicing personal/institutional opinions as opposed to acting >> as leader (sometimes the roles conflict and also sometimes >> the leader does have to enforce the clear general opinion). > > I think you are saying the same thing. This is, of course, aside the fact that you wouldn't find anybody who is active in the forum and who doesn't have a personal agenda. We all have one and that's why we are engaged. The people with a double role just have to keep it separate - and that's not only true for the chair, but (perhaps to a lesser extent) also for the other roles that Jeff pointed out. > > I also agree with Jeff - if we are setting up a new structure, let's look at all roles and define them better. > >> As to logistics, yes, I see your point but I still think >> that combining them would generally simplify things. I >> also think it is fine for the chair to delegate aspects >> that the convener handles but they should still be >> responsible for them (we all blame the chair if they do >> not go right so they should also have the authority since >> they de facto have responsibility). > > The convener role will always have to be split up, in particular if meetings get hosted at individual organizations. A certain part of the work (e.g., wireless, badging, room reservation, ...) always fell on people from the hosting organization and that won't change. > > The more critical question, IMHO, is the financial part that Jeff brought up. We have organizations that have more restrictions than others and we have to take that into account. Further, if the person responsible for financial transactions changes, we have to transfer money - the same if the person responsible changes organization (as we can see at Rich's example, which ended up being non trivial or even impossible, if I understood the situation right). The easiest thing would be to have a constant entity who manages all financial aspects of the forum along with a separate role of a treasurer. > > Martin > > >> >> Bronis >> >> >> On Wed, 26 Sep 2012, Jeff Squyres wrote: >> >>> This is exactly why we should have the discussion: "chair" and "convener" have not necessarily been the same person over the last few years. Rich has been the convener since 2.1, but has only been the chair for 3.0. >>> >>> My $0.02 is that I don't think they need to be the same person. The chair is the [neutral] technical leader (and therefore sets the meeting agendas), but the convener is someone with an organization who can make meeting room reservations, accept registration and payments from meeting attendees, etc. That being said, as long as the chair remains neutral in setting the agenda and guiding the process, the chair should be allowed to have their own technical opinions, just like anyone else. Otherwise, you're just neutering the chair, and their incentive to participate is quite limited. >>> >>> In short: I see the convener as a largely logistical role, but the chair as more of a technical role. >>> >>> >>> >>> On Sep 27, 2012, at 6:33 AM, Bronis R. de Supinski wrote: >>> >>>> >>>> Personally, I think "chair" should also be "meeting convener". >>>> The chair should be largely administrative/neutral/meeting >>>> organizer and runner. Convening meetings is a natural part >>>> of that activity. The chair should not be promoting a >>>> particular technical agenda (at least not as "chair"). >>>> >>>> >>>> On Wed, 26 Sep 2012, Jeff Squyres wrote: >>>> >>>>> As part of this discussion, it might be worthwhile to also bring up the auxiliary roles -- there's a bunch of things that happen behind the scenes to make the Forum function: >>>>> >>>>> - secretary >>>>> - meeting convener >>>>> - web maintainers (main site, meetings site, lists site) >>>>> - svn / trac maintainer >>>>> >>>>> I mention this because, although I'm happy to continue in the role as secretary, a) if there's someone else burning to do that role, it would be fair to discuss it, and b) sometimes there has been confusion about exactly what the secretary is supposed to do. Maybe it would be useful to delineate the responsibilities of the auxiliary roles. Just like we discovered that we all didn't have the same conceptions of how the secretary applies the voting rules, perhaps there are other unwritten inconsistencies, too. >>>>> >>>>> >>>>> >>>>> On Sep 26, 2012, at 10:27 PM, Brightwell, Ronald wrote: >>>>> >>>>>> >>>>>> Greetings: >>>>>> >>>>>> I wasn't able to attend the Forum meeting last week, so I don't know if this came up or not, but given that it's election season in the US and MPI 3.0 has been officially accepted, it seems like a good time to discuss leadership of the Forum going forward. I know several people who have been involved from a leadership standpoint are in different situations now then they were when the 3.0 process started, and their priorities and commitments are different. I think it might be good to have an open discussion about making changes to some of the leadership roles in an effort to make sure that we start this next phase in the best way possible. >>>>>> >>>>>> -Ron >>>>>> >>>>>> >>>>>> >>>>>> _______________________________________________ >>>>>> mpi-forum mailing list >>>>>> mpi-forum(a)lists.mpi-forum.org >>>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>>>> >>>>> >>>>> -- >>>>> Jeff Squyres >>>>> jsquyres(a)cisco.com >>>>> For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/ >>>>> >>>>> >>>>> _______________________________________________ >>>>> mpi-forum mailing list >>>>> mpi-forum(a)lists.mpi-forum.org >>>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>>>> >>>> _______________________________________________ >>>> mpi-forum mailing list >>>> mpi-forum(a)lists.mpi-forum.org >>>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum >>> >>> >>> -- >>> Jeff Squyres >>> jsquyres(a)cisco.com >>> For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/ >>> >>> >> _______________________________________________ >> mpi-forum mailing list >> mpi-forum(a)lists.mpi-forum.org >> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum > > ________________________________________________________________________ > Martin Schulz, schulzm(a)llnl.gov, http://people.llnl.gov/schulzm > CASC @ Lawrence Livermore National Laboratory, Livermore, USA > > > > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum

1 0