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

Re: [Mpi-forum] Voting results
by Jeff Hammond 30 May '12

30 May '12

On Wed, May 30, 2012 at 7:57 PM, George Bosilca <bosilca(a)eecs.utk.edu> wrote: > In general, the term "simple majority" is biased by our origins. In US it simply means over 1/2 of the valid votes cast (so everything except the absentees). Everywhere else, and this includes some other english speaking countries, a simple majority is the greatest number of parts, in other words the choice that got more votes than any other. Indeed, there are only two countries - none outside of North American - that use the first definition [http://en.wikipedia.org/wiki/Simple_majority] > I can hardly believe I'm the only one with this understanding of the simple majority term. Moreover, I remember live discussion where we clearly stated that there __are__ differences between the "abstain" and the "no"vote, and that one must carefully choose her vote. While these discussions were not in the context of first of second voting, they were at least in the context of readings (which is similarly important as it decide the original fate of a ticket). Maintaining several possible interpretations, and changing the counting before a remote vote doesn't make the process more transparent. I agree completely. If abstain=no then there is no point in having both. > Another valid issue raised during this meeting was about the blessing from the user community. Interestingly enough for a standard targeted toward a user community, such an issue was never brought forward. I think users should be an important factor in our decisions, as important as their contributions to the ongoing discussions. And yes the forum should grant them a significant period of time between the moment when we vote a ticket and the moment when this ticket gets into the standard. Our users will clearly take advantage of such a grace period to thoughtfully evaluate the proposal and raise any issues they might have encountered. We should have done this from day one! No users asked for C++ bindings to be deprecated and a number protested quite vigorously. Many users have asked for active messages and the Forum is has made no progress on this. Even something as simple as MPI_Iwin_create, which was requested by multiple users and had a very valid use case, was shot down in the RMA working group because it was too hard for Brian to implement. The notion that users are a high priority to the Forum is rather farcical as far as I'm concerned. I wish this was not the case. Jeff > On May 31, 2012, at 06:10 , Bronis R. de Supinski wrote: > >> >> All: >> >> Hmm. Quite the controversy. However, the rules as enforced >> in Japan are consistent with my understanding of what they >> have always been. More importantly, they are consistent >> with the wording in the bylaws. Here is what Jeff quoted: >> >> ï¿½ ï¿½a simple majority is defined as a simple majority >> ï¿½ ï¿½of those present and eligible to vote. >> >> Those who abstained were present and eligible to vote. >> They did not vote yes. The effect is that they voted >> "no" by this definition. If they did not want their >> vote effectively to be "no" then they should have left >> the room. I recall several instances in which someone >> was out of the room (perhaps even momentarily for a bio >> break) and Jeff recorded their vote as "not present". >> See the definition above -- they then do not count as >> present so they do not figure into the required "yes" count. >> >> As I stated, my understanding of the rules is consistent >> with the interpretation used in Japan. I would object to >> any other interpretation since the by-laws are actually >> clear on this point. I agree that the by-laws should be >> clear in general; while I think they are clear, I would >> not object to a clarifying statement being added to the >> effect that "abstentions are effectively negative votes." >> I think we have many other issues that should be made >> concrete in the by-laws and this is the least important. >> What is required to pass a first reading is probably the >> most obvious issue. >> >> Bronis >> >> >> >> >> On Wed, 30 May 2012, Fab Tillier wrote: >> >>> Jeff Squyres wrote on Wed, 30 May 2012 at 12:12:26 >>> >>>> On May 30, 2012, at 2:57 PM, Jeff Hammond wrote: >>>> >>>>>> The fact that some votes were still recorded as 'abstain' is an indication >>>>>> that this bylaw change was half baked. >>>>> >>>>> Especially when the meeting is attended by so few people due to the >>>>> location. ï¿½It seems like a weasel tactic to pick a remote location to >>>>> change the by-laws with a single vote. >>>> >>>> >>>> To be clear, the process document states: >>>> >>>> ï¿½ ï¿½For the purposes of voting, a simple majority is defined as a simple >>>> ï¿½ ï¿½majority of those present and eligible to vote. >>>> In the context of the document, the phrase "simple majority" is used to >>>> describe what is needed for ballots to pass; this sentence is attempting to >>>> define that phrase. ï¿½So even though the above sentence looks like a circular >>>> definition, I think it's really an open-ended definition (e.g., a google search >>>> for "simple majority definition" turns up both definitions). >>>> >>>> I was not there and don't know *exactly* what happened, so I'll refrain from >>>> commenting further. >>> >>> If the bylaws are vague, we should clarify them. ï¿½We should not however reinterpret them at each meeting, and should all agree on a proper interpretation and stick to it, such that ambiguity is removed going forward. ï¿½Allowing our bylaws to be vague enough to afford a re-interpretation at each meeting does nobody any good. >>> >>> -Fab >>> >>> >>> >>> >>> _______________________________________________ >>> 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 -- Jeff Hammond Argonne Leadership Computing Facility University of Chicago Computation Institute jhammond(a)alcf.anl.gov / (630) 252-5381 http://www.linkedin.com/in/jeffhammond https://wiki.alcf.anl.gov/parts/index.php/User:Jhammond

1 0

Re: [Mpi-forum] Voting results
by George Bosilca 30 May '12

30 May '12

In general, the term "simple majority" is biased by our origins. In US it simply means over 1/2 of the valid votes cast (so everything except the absentees). Everywhere else, and this includes some other english speaking countries, a simple majority is the greatest number of parts, in other words the choice that got more votes than any other. I can hardly believe I'm the only one with this understanding of the simple majority term. Moreover, I remember live discussion where we clearly stated that there __are__ differences between the "abstain" and the "no"vote, and that one must carefully choose her vote. While these discussions were not in the context of first of second voting, they were at least in the context of readings (which is similarly important as it decide the original fate of a ticket). Maintaining several possible interpretations, and changing the counting before a remote vote doesn't make the process more transparent. Another valid issue raised during this meeting was about the blessing from the user community. Interestingly enough for a standard targeted toward a user community, such an issue was never brought forward. I think users should be an important factor in our decisions, as important as their contributions to the ongoing discussions. And yes the forum should grant them a significant period of time between the moment when we vote a ticket and the moment when this ticket gets into the standard. Our users will clearly take advantage of such a grace period to thoughtfully evaluate the proposal and raise any issues they might have encountered. We should have done this from day one! george. On May 31, 2012, at 06:10 , Bronis R. de Supinski wrote: > > All: > > Hmm. Quite the controversy. However, the rules as enforced > in Japan are consistent with my understanding of what they > have always been. More importantly, they are consistent > with the wording in the bylaws. Here is what Jeff quoted: > > a simple majority is defined as a simple majority > of those present and eligible to vote. > > Those who abstained were present and eligible to vote. > They did not vote yes. The effect is that they voted > "no" by this definition. If they did not want their > vote effectively to be "no" then they should have left > the room. I recall several instances in which someone > was out of the room (perhaps even momentarily for a bio > break) and Jeff recorded their vote as "not present". > See the definition above -- they then do not count as > present so they do not figure into the required "yes" count. > > As I stated, my understanding of the rules is consistent > with the interpretation used in Japan. I would object to > any other interpretation since the by-laws are actually > clear on this point. I agree that the by-laws should be > clear in general; while I think they are clear, I would > not object to a clarifying statement being added to the > effect that "abstentions are effectively negative votes." > I think we have many other issues that should be made > concrete in the by-laws and this is the least important. > What is required to pass a first reading is probably the > most obvious issue. > > Bronis > > > > > On Wed, 30 May 2012, Fab Tillier wrote: > >> Jeff Squyres wrote on Wed, 30 May 2012 at 12:12:26 >> >>> On May 30, 2012, at 2:57 PM, Jeff Hammond wrote: >>> >>>>> The fact that some votes were still recorded as 'abstain' is an indication >>>>> that this bylaw change was half baked. >>>> >>>> Especially when the meeting is attended by so few people due to the >>>> location. It seems like a weasel tactic to pick a remote location to >>>> change the by-laws with a single vote. >>> >>> >>> To be clear, the process document states: >>> >>> For the purposes of voting, a simple majority is defined as a simple >>> majority of those present and eligible to vote. >>> In the context of the document, the phrase "simple majority" is used to >>> describe what is needed for ballots to pass; this sentence is attempting to >>> define that phrase. So even though the above sentence looks like a circular >>> definition, I think it's really an open-ended definition (e.g., a google search >>> for "simple majority definition" turns up both definitions). >>> >>> I was not there and don't know *exactly* what happened, so I'll refrain from >>> commenting further. >> >> If the bylaws are vague, we should clarify them. We should not however reinterpret them at each meeting, and should all agree on a proper interpretation and stick to it, such that ambiguity is removed going forward. Allowing our bylaws to be vague enough to afford a re-interpretation at each meeting does nobody any good. >> >> -Fab >> >> >> >> >> _______________________________________________ >> 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

1 0

Re: [Mpi-forum] Voting results
by Martin Schulz 30 May '12

30 May '12

On May 31, 2012, at 6:23 AM, Bronis R. de Supinski wrote: > > Because "sum>0" is a poor measure for something that > is ready for standardization. ISO standards require > unanimity. While that requirement is perhaps overly > strict, it is much better than "more in favor than > those who already recognize problems in the current form". I agree with Bronis - the simple majority rule is not the best idea for creating standards, in particular since we are anyway always saying that we are consensus driven. I personally think a 2/3 or 3/4 majority (with either counting abstains as no or at least having a quorum for the number of yes votes) would not be unreasonable and follow the spirit of the MPI forum better. To also second something else: I fully agree that we need a better documentation about the process and that this is not restricted to the votes. In addition to the requirements for readings, that Bronis mentioned, it would also be good to define what we consider ticket 0 and what not (we have gotten very open on this the closer the deadline came, which in my opinion is exactly the wrong thing to do), what are the time requirements to make documents available, when and what can we reread within one meeting, the exact role and duties of the chapter committees, when and how many reviews/reviewers are needed for tickets, etc. Martin > > On Wed, 30 May 2012, Jeff Hammond wrote: > >> If abstain=no then why bother having the category at all? Are we >> trying to implement abstain="polite no" and no="cruel no"? >> >> Why don't we vote with +1, 0 and -1, where sum>0 passes? >> >> Jeff >> >> On Wed, May 30, 2012 at 4:10 PM, Bronis R. de Supinski <bronis(a)llnl.gov> wrote: >>> >>> All: >>> >>> Hmm. Quite the controversy. However, the rules as enforced >>> in Japan are consistent with my understanding of what they >>> have always been. More importantly, they are consistent >>> with the wording in the bylaws. Here is what Jeff quoted: >>> >>> >>> a simple majority is defined as a simple majority >>> of those present and eligible to vote. >>> >>> Those who abstained were present and eligible to vote. >>> They did not vote yes. The effect is that they voted >>> "no" by this definition. If they did not want their >>> vote effectively to be "no" then they should have left >>> the room. I recall several instances in which someone >>> was out of the room (perhaps even momentarily for a bio >>> break) and Jeff recorded their vote as "not present". >>> See the definition above -- they then do not count as >>> present so they do not figure into the required "yes" count. >>> >>> As I stated, my understanding of the rules is consistent >>> with the interpretation used in Japan. I would object to >>> any other interpretation since the by-laws are actually >>> clear on this point. I agree that the by-laws should be >>> clear in general; while I think they are clear, I would >>> not object to a clarifying statement being added to the >>> effect that "abstentions are effectively negative votes." >>> I think we have many other issues that should be made >>> concrete in the by-laws and this is the least important. >>> What is required to pass a first reading is probably the >>> most obvious issue. >>> >>> Bronis >>> >>> >>> >>> >>> >>> On Wed, 30 May 2012, Fab Tillier wrote: >>> >>>> Jeff Squyres wrote on Wed, 30 May 2012 at 12:12:26 >>>> >>>>> On May 30, 2012, at 2:57 PM, Jeff Hammond wrote: >>>>> >>>>>>> The fact that some votes were still recorded as 'abstain' is an >>>>>>> indication >>>>>>> that this bylaw change was half baked. >>>>>> >>>>>> >>>>>> Especially when the meeting is attended by so few people due to the >>>>>> location. It seems like a weasel tactic to pick a remote location to >>>>>> change the by-laws with a single vote. >>>>> >>>>> >>>>> >>>>> To be clear, the process document states: >>>>> >>>>> For the purposes of voting, a simple majority is defined as a simple >>>>> majority of those present and eligible to vote. >>>>> In the context of the document, the phrase "simple majority" is used to >>>>> describe what is needed for ballots to pass; this sentence is attempting >>>>> to >>>>> define that phrase. So even though the above sentence looks like a >>>>> circular >>>>> definition, I think it's really an open-ended definition (e.g., a google >>>>> search >>>>> for "simple majority definition" turns up both definitions). >>>>> >>>>> I was not there and don't know *exactly* what happened, so I'll refrain >>>>> from >>>>> commenting further. >>>> >>>> >>>> If the bylaws are vague, we should clarify them. We should not however >>>> reinterpret them at each meeting, and should all agree on a proper >>>> interpretation and stick to it, such that ambiguity is removed going >>>> forward. Allowing our bylaws to be vague enough to afford a >>>> re-interpretation at each meeting does nobody any good. >>>> >>>> -Fab >>>> >>>> >>>> >>>> >>>> _______________________________________________ >>>> 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 Hammond >> Argonne Leadership Computing Facility >> University of Chicago Computation Institute >> jhammond(a)alcf.anl.gov / (630) 252-5381 >> http://www.linkedin.com/in/jeffhammond >> https://wiki.alcf.anl.gov/parts/index.php/User:Jhammond > _______________________________________________ > 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

1 0

Re: [Mpi-forum] [EXTERNAL] Re: ticket 273
by Mohamad Chaarawi 30 May '12

30 May '12

Hi Brian, On 5/30/2012 5:22 PM, Barrett, Brian W wrote: > Mohamad - > > I can't speak for others. I can say that there was a general feeling from those who voted no or abstain that the need for a non-blocking file I/O collective, when split collectives are already unable to provide any asynchrony, is a problem for implementors. I agree with you that an implementation is not easy for those routines.. But I don't see that it should be a reason for voting a ticket down, or worse saying that it should be postponed, because the implementation challenge will still be there for 3.1 > The problems with file position and ordering is still a bit unsettling to some. Is that something new for any of the file routines (blocking and non blocking)? The consistency semantics are defined as clear as any of the other file routines. > > Also, I know that for myself, I haven't voiced objections in the last couple of meetings because my objections had been made previously and not well addressed, so I was going to vote no. This happens in any voting body. > > It was poor scheduling that the last first vote occurred on a foreign trip. I tried very hard to get someone to speak up about each ticket ahead of time, as I can't speak for many of the tickets (and it would not be fair for me to give initial background on a ticket I was going to vote "no" for). This was obviously problematic for some tickets. The discussion on the ifile ticket largely covered the points I raised in the initial paragraph. Admittedly I might have done something to try and get people to speak up about this ticket a couple of weeks before the meeting, but my problem with this, by looking at my slides for all the presentations I made for this ticket where I record all the straw votes, there were two straw votes at 2 separate meetings, a 17-0-1 and another similar straw votes on this ticket. The ticket that got a lot of objections (many from you) was the nonblocking file manipulation routines ticket, which was withdrawn. The only problem that was voiced for this ticket was the issue with shared file pointers. This was addressed at the last meeting, and everybody seemed content with the draft read. I don't want to sound stubborn here, but between the voting rules change, and the lack of feedback given from people for this ticket, I'm a bit frustrated. Mohamad > > Unlike some of the other tickets that were voted down, the ifile ticket was one of two (FT being the other) that people were more receptive to seeing in MPI-3.1. > > Brian > > -- > Brian W. Barrett > Scalable System Software Group > Sandia National Laboratories > ________________________________________ > From: mpi-forum-bounces(a)lists.mpi-forum.org [mpi-forum-bounces(a)lists.mpi-forum.org] on behalf of Mohamad Chaarawi [chaarawi(a)hdfgroup.org] > Sent: Wednesday, May 30, 2012 2:24 PM > To: mpi-forum(a)lists.mpi-forum.org > Subject: [EXTERNAL] Re: [Mpi-forum] ticket 273 > > Hi Rich, > > On 5/30/2012 3:03 PM, Richard Graham wrote: >> Mohamad, >> Two items that I wrote down based on comments: >> - One view was expressed that adding a nonblocking interface, will not change the fact that implementation will still be blocking > The implementation that UH provides builds on top of libnbc which > implements the non-blocking collective communication operations.. They > use non blocking point-to-point operations and asynchronous > aio_read/aio_write functions. So what part of that is blocking? > > btw, why was this concern not mentioned during the formal reading or on > the I/O mailing list? > >> - The ticket was rushed, and not ready > This ticket was going back and forth for a little less than a year.. I > don't any reason why it is considered rushed.. Again most of the I/O > chapter committee reviewed and approved it.. > Other than final merging issues what was wrong with ticket? and again > why were they not brought up till the Japan vote, where we were not > there to address them? > > Thanks, > Mohamad > >> Rich >> >> -----Original Message----- >> From: mpi-forum-bounces(a)lists.mpi-forum.org [mailto:[email protected]] On Behalf Of Mohamad Chaarawi >> Sent: Thursday, May 31, 2012 4:37 AM >> To: Main MPI Forum mailing list >> Subject: [Mpi-forum] ticket 273 >> >> Hi All, >> >> I'm still trying to figure out why ticket 273 was voted down. At the last meeting, I would say there was a consensus that the non blocking collective I/O ticket (273) was ready to be voted in. Nobody expressed an objection after me removing the shared file pointer nbc routines and re-reading the draft. All the I/O chapter committee reviewed the draft and commented on the ticket that it looks good. >> >> At the Japan meeting, we get 8 yes , 2 nos, and 6 abstains, which under the new rules marks the ticket as failed. >> None from our organization could attend the meeting to really understand what the 2 no votes were for, and why 6 organizations abstained. >> If the abstains are because people don't care or don't understand what is being voted on or missed the vote (which are the reasons why I vote abstain), then the new voting rule doesn't really make sense. If the abstains where for something else, then the votes should be NO. So could someone mention why this was voted down? >> >> 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 > > > > _______________________________________________ > 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] [EXTERNAL] Re: Voting results
by William Gropp 30 May '12

30 May '12

The idea behind this is that institutions that have committed to understanding and participating in the entire effort have the perspective to cast an informed vote. The MPI Forum has, as has been alluded to, tried to reach a near consensus on issues, and views (correctly, in my opinion) a close vote as a failure that should be examined carefully. Yes, you do need travel funds. Yes, we should consider alternatives, but in my experience, the sort of in-depth discussion that is really needed is still hard to do except in person. Bill William Gropp Director, Parallel Computing Institute Deputy Director for Research Institute for Advanced Computing Applications and Technologies Paul and Cynthia Saylor Professor of Computer Science University of Illinois Urbana-Champaign On May 30, 2012, at 6:11 PM, Jeff Hammond wrote: >> One take away from being in the room on Wednesday is that there is a strong importance to having an advocate present for voting. > > This naturally prejudices in favor of institutions with substantial > travel funds. > > I still do not understand why voting remotely via video conference is > not an option. It's 2012. We might not have flying cars yet, but > we've got more than enough future tech to preclude the need for people > to fly around the world to participate in the MPI Forum. > > Jeff > > -- > Jeff Hammond > Argonne Leadership Computing Facility > University of Chicago Computation Institute > jhammond(a)alcf.anl.gov / (630) 252-5381 > http://www.linkedin.com/in/jeffhammond > https://wiki.alcf.anl.gov/parts/index.php/User:Jhammond > _______________________________________________ > 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] [EXTERNAL] Re: Voting results
by Jeff Hammond 30 May '12

30 May '12

> One take away from being in the room on Wednesday is that there is a strong importance to having an advocate present for voting. This naturally prejudices in favor of institutions with substantial travel funds. I still do not understand why voting remotely via video conference is not an option. It's 2012. We might not have flying cars yet, but we've got more than enough future tech to preclude the need for people to fly around the world to participate in the MPI Forum. Jeff -- Jeff Hammond Argonne Leadership Computing Facility University of Chicago Computation Institute jhammond(a)alcf.anl.gov / (630) 252-5381 http://www.linkedin.com/in/jeffhammond https://wiki.alcf.anl.gov/parts/index.php/User:Jhammond

1 0

Re: [Mpi-forum] [EXTERNAL] Re: Voting results
by Barrett, Brian W 30 May '12

30 May '12

Following up on Bronis's discussion; there have been no "close" votes for first or second votes for MPI-3 prior to the Japan meeting. The closest was the first vote for the new One-Sided chapter, at 13-0-5-1, with many of the abstains because we didn't have an implementation and it was really complicated. There was one close vote prior in MPI-2.2, a second vote which would not have passed under the rules used in Japan. Note, however, that MPI-2.2 operated under a different set of rules than MPI-3, so this is not necessarily applicable. One take away from being in the room on Wednesday is that there is a strong importance to having an advocate present for voting. The voting rules need to be clarified, no doubt in my mind, but there are a lot of organizations who give their voting members enough leeway that votes can change based on pre-vote discussion. Brian -- Brian W. Barrett Scalable System Software Group Sandia National Laboratories ________________________________________ From: mpi-forum-bounces(a)lists.mpi-forum.org [mpi-forum-bounces(a)lists.mpi-forum.org] on behalf of Bronis R. de Supinski [bronis(a)llnl.gov] Sent: Wednesday, May 30, 2012 3:20 PM To: Main MPI Forum mailing list Subject: [EXTERNAL] Re: [Mpi-forum] Voting results All: I also meant to address another question that arose. While I have missed some meetings, at no meeting at which I was present to a ticket pass under the apparent widespread misinterpretation of the rules. If one had, I would have objected -- first, because that is not what the by-laws say; and second, because adding something to the standard that cannot even get a simple majority of yes votes of THOSE PRESENT is an inherently stupid idea (adding it; not necessarily the proposal itself). I do not recall a particularly close vote when I was present although I might be mistaken. The only votes that I recall that were close were straw votes, which do not count. Many of those have been close. Bronis On Wed, 30 May 2012, Bronis R. de Supinski wrote: > > All: > > Hmm. Quite the controversy. However, the rules as enforced > in Japan are consistent with my understanding of what they > have always been. More importantly, they are consistent > with the wording in the bylaws. Here is what Jeff quoted: > > a simple majority is defined as a simple majority > of those present and eligible to vote. > > Those who abstained were present and eligible to vote. > They did not vote yes. The effect is that they voted > "no" by this definition. If they did not want their > vote effectively to be "no" then they should have left > the room. I recall several instances in which someone > was out of the room (perhaps even momentarily for a bio > break) and Jeff recorded their vote as "not present". > See the definition above -- they then do not count as > present so they do not figure into the required "yes" count. > > As I stated, my understanding of the rules is consistent > with the interpretation used in Japan. I would object to > any other interpretation since the by-laws are actually > clear on this point. I agree that the by-laws should be > clear in general; while I think they are clear, I would > not object to a clarifying statement being added to the > effect that "abstentions are effectively negative votes." > I think we have many other issues that should be made > concrete in the by-laws and this is the least important. > What is required to pass a first reading is probably the > most obvious issue. > > Bronis > > > > > On Wed, 30 May 2012, Fab Tillier wrote: > >> Jeff Squyres wrote on Wed, 30 May 2012 at 12:12:26 >> >>> On May 30, 2012, at 2:57 PM, Jeff Hammond wrote: >>> >>>>> The fact that some votes were still recorded as 'abstain' is an indication >>>>> that this bylaw change was half baked. >>>> >>>> Especially when the meeting is attended by so few people due to the >>>> location. It seems like a weasel tactic to pick a remote location to >>>> change the by-laws with a single vote. >>> >>> >>> To be clear, the process document states: >>> >>> For the purposes of voting, a simple majority is defined as a simple >>> majority of those present and eligible to vote. >>> In the context of the document, the phrase "simple majority" is used to >>> describe what is needed for ballots to pass; this sentence is attempting to >>> define that phrase. So even though the above sentence looks like a circular >>> definition, I think it's really an open-ended definition (e.g., a google search >>> for "simple majority definition" turns up both definitions). >>> >>> I was not there and don't know *exactly* what happened, so I'll refrain from >>> commenting further. >> >> If the bylaws are vague, we should clarify them. We should not however reinterpret them at each meeting, and should all agree on a proper interpretation and stick to it, such that ambiguity is removed going forward. Allowing our bylaws to be vague enough to afford a re-interpretation at each meeting does nobody any good. >> >> -Fab >> >> >> >> >> _______________________________________________ >> 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

1 0

Re: [Mpi-forum] [EXTERNAL] Re: ticket 273
by Barrett, Brian W 30 May '12

30 May '12

Mohamad - I can't speak for others. I can say that there was a general feeling from those who voted no or abstain that the need for a non-blocking file I/O collective, when split collectives are already unable to provide any asynchrony, is a problem for implementors. The problems with file position and ordering is still a bit unsettling to some. Also, I know that for myself, I haven't voiced objections in the last couple of meetings because my objections had been made previously and not well addressed, so I was going to vote no. This happens in any voting body. It was poor scheduling that the last first vote occurred on a foreign trip. I tried very hard to get someone to speak up about each ticket ahead of time, as I can't speak for many of the tickets (and it would not be fair for me to give initial background on a ticket I was going to vote "no" for). This was obviously problematic for some tickets. The discussion on the ifile ticket largely covered the points I raised in the initial paragraph. Unlike some of the other tickets that were voted down, the ifile ticket was one of two (FT being the other) that people were more receptive to seeing in MPI-3.1. Brian -- Brian W. Barrett Scalable System Software Group Sandia National Laboratories ________________________________________ From: mpi-forum-bounces(a)lists.mpi-forum.org [mpi-forum-bounces(a)lists.mpi-forum.org] on behalf of Mohamad Chaarawi [chaarawi(a)hdfgroup.org] Sent: Wednesday, May 30, 2012 2:24 PM To: mpi-forum(a)lists.mpi-forum.org Subject: [EXTERNAL] Re: [Mpi-forum] ticket 273 Hi Rich, On 5/30/2012 3:03 PM, Richard Graham wrote: > Mohamad, > Two items that I wrote down based on comments: > - One view was expressed that adding a nonblocking interface, will not change the fact that implementation will still be blocking The implementation that UH provides builds on top of libnbc which implements the non-blocking collective communication operations.. They use non blocking point-to-point operations and asynchronous aio_read/aio_write functions. So what part of that is blocking? btw, why was this concern not mentioned during the formal reading or on the I/O mailing list? > - The ticket was rushed, and not ready This ticket was going back and forth for a little less than a year.. I don't any reason why it is considered rushed.. Again most of the I/O chapter committee reviewed and approved it.. Other than final merging issues what was wrong with ticket? and again why were they not brought up till the Japan vote, where we were not there to address them? Thanks, Mohamad > > Rich > > -----Original Message----- > From: mpi-forum-bounces(a)lists.mpi-forum.org [mailto:[email protected]] On Behalf Of Mohamad Chaarawi > Sent: Thursday, May 31, 2012 4:37 AM > To: Main MPI Forum mailing list > Subject: [Mpi-forum] ticket 273 > > Hi All, > > I'm still trying to figure out why ticket 273 was voted down. At the last meeting, I would say there was a consensus that the non blocking collective I/O ticket (273) was ready to be voted in. Nobody expressed an objection after me removing the shared file pointer nbc routines and re-reading the draft. All the I/O chapter committee reviewed the draft and commented on the ticket that it looks good. > > At the Japan meeting, we get 8 yes , 2 nos, and 6 abstains, which under the new rules marks the ticket as failed. > None from our organization could attend the meeting to really understand what the 2 no votes were for, and why 6 organizations abstained. > If the abstains are because people don't care or don't understand what is being voted on or missed the vote (which are the reasons why I vote abstain), then the new voting rule doesn't really make sense. If the abstains where for something else, then the votes should be NO. So could someone mention why this was voted down? > > 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

1 0

Re: [Mpi-forum] Voting results
by Fab Tillier 30 May '12

30 May '12

If someone else is afforded the privilege of voting "not present", I'd like that right too (who are you to judge whether I deserve that right?). In fact, I was under the impression that "abstain" did exactly this... -Fab Bronis R. de Supinski wrote on Wed, 30 May 2012 at 14:25:31 > > I disagree. They are able to enter the room; they can leave it. > More importantly, if we had anyone who had such impairments, I > am certain that we would allow them to voice themselves as "not present". > > On Wed, 30 May 2012, Jeff Hammond wrote: > >> I would like it noted that the following is extremely unfriendly to >> anyone with mobility problems and probably not ADA-compliant. >> >> "If they did not want their vote effectively to be "no" then they >> should have left the room." >> >> Jeff >> >> On Wed, May 30, 2012 at 4:10 PM, Bronis R. de Supinski <bronis(a)llnl.gov> > wrote: >>> >>> All: >>> >>> Hmm. Quite the controversy. However, the rules as enforced >>> in Japan are consistent with my understanding of what they >>> have always been. More importantly, they are consistent >>> with the wording in the bylaws. Here is what Jeff quoted: >>> >>> >>> ï¿½ ï¿½a simple majority is defined as a simple majority >>> ï¿½ ï¿½of those present and eligible to vote. >>> >>> Those who abstained were present and eligible to vote. >>> They did not vote yes. The effect is that they voted >>> "no" by this definition. If they did not want their >>> vote effectively to be "no" then they should have left >>> the room. I recall several instances in which someone >>> was out of the room (perhaps even momentarily for a bio >>> break) and Jeff recorded their vote as "not present". >>> See the definition above -- they then do not count as >>> present so they do not figure into the required "yes" count. >>> >>> As I stated, my understanding of the rules is consistent >>> with the interpretation used in Japan. I would object to >>> any other interpretation since the by-laws are actually >>> clear on this point. I agree that the by-laws should be >>> clear in general; while I think they are clear, I would >>> not object to a clarifying statement being added to the >>> effect that "abstentions are effectively negative votes." >>> I think we have many other issues that should be made >>> concrete in the by-laws and this is the least important. >>> What is required to pass a first reading is probably the >>> most obvious issue. >>> >>> Bronis >>> >>> >>> >>> >>> >>> On Wed, 30 May 2012, Fab Tillier wrote: >>> >>>> Jeff Squyres wrote on Wed, 30 May 2012 at 12:12:26 >>>> >>>>> On May 30, 2012, at 2:57 PM, Jeff Hammond wrote: >>>>> >>>>>>> The fact that some votes were still recorded as 'abstain' is an >>>>>>> indication >>>>>>> that this bylaw change was half baked. >>>>>> >>>>>> >>>>>> Especially when the meeting is attended by so few people due to the >>>>>> location. ï¿½It seems like a weasel tactic to pick a remote location to >>>>>> change the by-laws with a single vote. >>>>> >>>>> >>>>> >>>>> To be clear, the process document states: >>>>> >>>>> ï¿½ ï¿½For the purposes of voting, a simple majority is defined as a simple >>>>> ï¿½ ï¿½majority of those present and eligible to vote. >>>>> In the context of the document, the phrase "simple majority" is used to >>>>> describe what is needed for ballots to pass; this sentence is attempting >>>>> to >>>>> define that phrase. ï¿½So even though the above sentence looks like a >>>>> circular >>>>> definition, I think it's really an open-ended definition (e.g., a google >>>>> search >>>>> for "simple majority definition" turns up both definitions). >>>>> >>>>> I was not there and don't know *exactly* what happened, so I'll refrain >>>>> from >>>>> commenting further. >>>> >>>> >>>> If the bylaws are vague, we should clarify them. ï¿½We should not however >>>> reinterpret them at each meeting, and should all agree on a proper >>>> interpretation and stick to it, such that ambiguity is removed going >>>> forward. ï¿½Allowing our bylaws to be vague enough to afford a >>>> re-interpretation at each meeting does nobody any good. >>>> >>>> -Fab >>>> >>>> >>>> >>>> >>>> _______________________________________________ >>>> 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 Hammond >> Argonne Leadership Computing Facility >> University of Chicago Computation Institute >> jhammond(a)alcf.anl.gov / (630) 252-5381 >> http://www.linkedin.com/in/jeffhammond >> https://wiki.alcf.anl.gov/parts/index.php/User:Jhammond >>

1 0

Re: [Mpi-forum] Voting results
by Jeff Hammond 30 May '12

30 May '12

There are days where I have to ingest a substantial quantity of pills to be able to walk and this condition is aggravated by the amount that I have to walk and/or stand. I don't feel that I should have to reveal my particular physiological disposition to a collection of people some of whom I barely know in order to express that which should be plainly obvious from doing nothing ("null vote"), especially since physical pain level is correlated in some instances with depression, which make me even less inclined to communicate the details of my physiological disposition to a room full of strangers. Sufficient? Jeff On Wed, May 30, 2012 at 4:25 PM, Bronis R. de Supinski <bronis(a)llnl.gov> wrote: > > I disagree. They are able to enter the room; they can leave it. > More importantly, if we had anyone who had such impairments, I > am certain that we would allow them to voice themselves as "not present". > > > On Wed, 30 May 2012, Jeff Hammond wrote: > >> I would like it noted that the following is extremely unfriendly to >> anyone with mobility problems and probably not ADA-compliant. >> >> "If they did not want their vote effectively to be "no" then they >> should have left the room." >> >> Jeff >> >> On Wed, May 30, 2012 at 4:10 PM, Bronis R. de Supinski <bronis(a)llnl.gov> >> wrote: >>> >>> >>> All: >>> >>> Hmm. Quite the controversy. However, the rules as enforced >>> in Japan are consistent with my understanding of what they >>> have always been. More importantly, they are consistent >>> with the wording in the bylaws. Here is what Jeff quoted: >>> >>> >>> ï¿½ ï¿½a simple majority is defined as a simple majority >>> ï¿½ ï¿½of those present and eligible to vote. >>> >>> Those who abstained were present and eligible to vote. >>> They did not vote yes. The effect is that they voted >>> "no" by this definition. If they did not want their >>> vote effectively to be "no" then they should have left >>> the room. I recall several instances in which someone >>> was out of the room (perhaps even momentarily for a bio >>> break) and Jeff recorded their vote as "not present". >>> See the definition above -- they then do not count as >>> present so they do not figure into the required "yes" count. >>> >>> As I stated, my understanding of the rules is consistent >>> with the interpretation used in Japan. I would object to >>> any other interpretation since the by-laws are actually >>> clear on this point. I agree that the by-laws should be >>> clear in general; while I think they are clear, I would >>> not object to a clarifying statement being added to the >>> effect that "abstentions are effectively negative votes." >>> I think we have many other issues that should be made >>> concrete in the by-laws and this is the least important. >>> What is required to pass a first reading is probably the >>> most obvious issue. >>> >>> Bronis >>> >>> >>> >>> >>> >>> On Wed, 30 May 2012, Fab Tillier wrote: >>> >>>> Jeff Squyres wrote on Wed, 30 May 2012 at 12:12:26 >>>> >>>>> On May 30, 2012, at 2:57 PM, Jeff Hammond wrote: >>>>> >>>>>>> The fact that some votes were still recorded as 'abstain' is an >>>>>>> indication >>>>>>> that this bylaw change was half baked. >>>>>> >>>>>> >>>>>> >>>>>> Especially when the meeting is attended by so few people due to the >>>>>> location. ï¿½It seems like a weasel tactic to pick a remote location to >>>>>> change the by-laws with a single vote. >>>>> >>>>> >>>>> >>>>> >>>>> To be clear, the process document states: >>>>> >>>>> ï¿½ ï¿½For the purposes of voting, a simple majority is defined as a simple >>>>> ï¿½ ï¿½majority of those present and eligible to vote. >>>>> In the context of the document, the phrase "simple majority" is used to >>>>> describe what is needed for ballots to pass; this sentence is >>>>> attempting >>>>> to >>>>> define that phrase. ï¿½So even though the above sentence looks like a >>>>> circular >>>>> definition, I think it's really an open-ended definition (e.g., a >>>>> google >>>>> search >>>>> for "simple majority definition" turns up both definitions). >>>>> >>>>> I was not there and don't know *exactly* what happened, so I'll refrain >>>>> from >>>>> commenting further. >>>> >>>> >>>> >>>> If the bylaws are vague, we should clarify them. ï¿½We should not however >>>> reinterpret them at each meeting, and should all agree on a proper >>>> interpretation and stick to it, such that ambiguity is removed going >>>> forward. ï¿½Allowing our bylaws to be vague enough to afford a >>>> re-interpretation at each meeting does nobody any good. >>>> >>>> -Fab >>>> >>>> >>>> >>>> >>>> _______________________________________________ >>>> 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 Hammond >> Argonne Leadership Computing Facility >> University of Chicago Computation Institute >> jhammond(a)alcf.anl.gov / (630) 252-5381 >> http://www.linkedin.com/in/jeffhammond >> https://wiki.alcf.anl.gov/parts/index.php/User:Jhammond -- Jeff Hammond Argonne Leadership Computing Facility University of Chicago Computation Institute jhammond(a)alcf.anl.gov / (630) 252-5381 http://www.linkedin.com/in/jeffhammond https://wiki.alcf.anl.gov/parts/index.php/User:Jhammond

1 0