April 2008 - Mpi-21 - mailman3.cels.anl.gov

Re: [Mpi-21] AUTHORS: new consistent version and write-token to you
by unknown＠example.com 16 Apr '08

16 Apr '08

Hi Rajeev, Thanks for all of your feedback. I've made every change but two, which I think we should delay to a decision for the Forum (or at least wait to get more votes from reviewers). Rolf, Can you add the following items to some ballot to be considered by the Forum? * pg 134, ln 32-33: What does matching arguments mean? Does it mean same arguments? That's not true for the root argument of rooted collectives on the group containing the root. Maybe just delete the sentence. * pg 163, Sec 5.9.3: Delete the sentence "This is an extension to MPI-1.2, ..." since we are all one now. I've incorporated the rest of the edits Rajeev suggested and committed the new chapter. -Adam Rajeev Thakur wrote: >Adam, > You have done a wonderful job of fixing the merged collectives chapter, >which was a nontrivial task (and hence I didn't volunteer for it :-)). I >only have some minor comments listed below. > >Rajeev > > >* pg 133, ln 27-28: change to the plural -- "patterns of" and "do not fit" > > Done. >* pg 134, ln 32-33: What does matching arguments mean? Does it mean same >arguments? That's not true for the root argument of rooted collectives on >the group containing the root. Maybe just delete the sentence. > > Yes, I agree, but it's not clear to me how to clarify this point either. There's probably a clean way to explain the intent here, but I don't want to make that decision. I'll leave this in to get more input from the Forum. >* pg 136, ln 22-23: move "for intracommunicators" to the beginning of the >sentence > > Done. >* pg 140, ln 28: change "are using" to "use" > > Done. >* pg 149, ln 45: change "are using" to "use" > > Done. >* pg 154, ln 33: change "are using" to "use" > > Done. >* pg 162, ln 10: change "are using" to "use" > > Done. >* pg 163, Sec 5.9.3: Delete the sentence "This is an extension to MPI-1.2, >..." since we are all one now. Merge the next paragraph on heterogeneous >into this one, and fix spelling of "charater" to character. > > I fixed the "charater" mis-spelling and merged the paragraphs. I left the statement about MPI-1.2 for now, although I moved it so that, after merging the other paragraph, the resulting paragraph reads a little more smoothly. This statement seems to highlight an important, but subtle change. I want to get feedback from the Forum on this item. >* pg 165, ln 10: change "are using" to "use" > > Done. >* pg 167, ln 38-39: Question - Do we want to also list MPI_Accumulate with a >forward reference to it? > >* pg 170, ln 3: change "is using" to "uses" > > Done. >* pg 171, ln 38: should be "In such a case" or better "In this case" > > Done. >* pg 171, ln 43: change "is using" to "uses" and "an intracommunicators" to >the singular > > Done. >* pg 174, ln 11: change "illegal" to "invalid" > > Done. >* pg 174, ln 12-20: Delete the entire rationale. It was included in MPI-1 to >justify not having an exclusive scan. Doesn't apply any more. You deleted >only the last sentence, which was explicit, but the whole rationale should >go. > > Done. >* pg 175, ln 13-20: By the same token, delete the entire rationale here. It >was used to justify adding an exclusive scan to MPI-2. Even the "no in-place >version" is covered above in lin 5. Maybe you can change that line 5 to >include the justification as follows: "No in-place version is supported >because it is not clear what this means for the process with rank zero". >Note that it should be "process with rank zero" rather than "process for >rank zero" as in the current text. > > I kept the rationale, because I think it provides some useful information here, but I changed it to the following: The exclusive scan is more general than the inclusive scan. Any inclusive scan operation can be achieved by using the exclusive scan and then locally combining the local contribution. Note that for non-invertable operations such as MPI_MAX, the exclusive scan cannot be computed with the inclusive scan. No in-place version is specified for MPI_EXSCAN because it is not clear what this means for the process with rank zero. >* pg 175, ln 25: change "is using" to "uses" > > Done. >Rest looks great! > > > > > > > >>-----Original Message----- >>From: mpi-21-bounces_at_[hidden] >>[mailto:mpi-21-bounces_at_[hidden]] On Behalf Of Rolf >>Rabenseifner >>Sent: Saturday, April 12, 2008 10:15 AM >>To: MPI 2.1 Mailing List >>Subject: Re: [Mpi-21] AUTHORS: new consistent version and >>write-token to you >> >>Hi Adam, >> >>this is really a good merge that you have done. >>Good ideas and very carefully implemented. >> >>The reviewers can see (nearly) all of your modifications in red. >>I expect that your reviewersand the Forum will enjoy your >>modifications. >> >>I have restored two minor mods you have done: >>(Line numbers refer to the new file you'll get with svn update) >> >>coll.tex, line 412 - I restored the "intracommunicator" >> because the categories describe the behavior only on >> intracommunicators, not on all, e.g., ALLGATHER has >> one result on intra and two on inter and the wording is >> "the result". >> >>coll.tex, line 491 - restored the word "also". >> These routines are valid for intra and inter. >> Writing "apply to inter" may be interpreted that they are >> only valid for inter. >> Writing "also apply to inter" should solve this problem. >> I also colored this sentence red. >> >>I hop, this isokay with you. >> >>Best regards >>Rolf >> >> >>On Fri, 11 Apr 2008 15:22:10 -0700 >> Adam Moody <moody20_at_[hidden]> wrote: >> >> >>>Hi Rolf, >>>I have checked in my changes for the Collectives chapter. >>> >>> >>It's ready to >> >> >>>be reviewed by the collective chapter reviewers. >>>-Adam >>> >>>Rolf Rabenseifner wrote: >>> >>> >>> >>>>>>write-token again to the chapter authors << >>>>>>- but only until Saturday 8:00 European GMT << >>>>>> >>>>>> >>>>>> >>>>>> >>>>Hi all authors and reviewers, >>>> >>>>Please read this mail carfully, if you want to make any further >>>>modification of the text. >>>> >>>>You must now call >>>> svn update >>>>to get the newest release!! >>>>We have a again a consistent version. Nearly all of >>>> >>>> >>>> >>>> >>http://www.hlrs.de/mpi/mpi21/doc/MPI-2.1draft-2008-02-23-review.txt >> >> >>>>is now implemented in the tex-source. >>>>Only a few items are still open, see statistics at the beginning. >>>> >>>>You can view the current version (it is svn revision 92) at >>>> >>>> http://www.hlrs.de/mpi/mpi21/doc/mpi-report.pdf >>>> >>>>It is only today and Friday remaining for finishing MPI-2.1. >>>>I give back the write-token to the authors. >>>> >>>>On Saturday morning European time, I take the write-token >>>>automatically on Saturday morning to build final pdf for the >>>>reviews. >>>> >>>>As I can see, many chapter authors made fine cleaning to the text. >>>>All larger changes should be defered to MPI-2.2! >>>>Especially the merge of MPI Environment (inquiry.tex), >>>> >>>> >>Misc (misc-2.tex) >> >> >>>>and External Interfaces (ei-2.tex) >>>>to one chapter "MPI Environment and Miscellany" (with 45 pages) >>>>may be defered also to MPI-2.2. >>>> >>>>Up to now, I was able to check most changes and they look fine. >>>> >>>>Best regards >>>>Rolf >>>> >>>> >>>> >>>>Dr. Rolf Rabenseifner . . . . . . . . . .. email >>>> >>>> >>rabenseifner_at_[hidden] >> >> >>>>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-21 mailing list >>>>mpi-21_at_[hidden] >>>>http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-21 >>>> >>>> >>>> >>>> >>>> >>>_______________________________________________ >>>mpi-21 mailing list >>>mpi-21_at_[hidden] >>>http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-21 >>> >>> >> >>Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner_at_[hidden] >>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-21 mailing list >>mpi-21_at_[hidden] >>http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-21 >> >> >> >> > >_______________________________________________ >mpi-21 mailing list >mpi-21_at_[hidden] >http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-21 > > >

1 0

Re: [Mpi-21] AUTHORS: new consistent version and write-token to you
by unknown＠example.com 16 Apr '08

16 Apr '08

Adam, You have done a wonderful job of fixing the merged collectives chapter, which was a nontrivial task (and hence I didn't volunteer for it :-)). I only have some minor comments listed below. Rajeev * pg 133, ln 27-28: change to the plural -- "patterns of" and "do not fit" * pg 134, ln 32-33: What does matching arguments mean? Does it mean same arguments? That's not true for the root argument of rooted collectives on the group containing the root. Maybe just delete the sentence. * pg 136, ln 22-23: move "for intracommunicators" to the beginning of the sentence * pg 140, ln 28: change "are using" to "use" * pg 149, ln 45: change "are using" to "use" * pg 154, ln 33: change "are using" to "use" * pg 162, ln 10: change "are using" to "use" * pg 163, Sec 5.9.3: Delete the sentence "This is an extension to MPI-1.2, ..." since we are all one now. Merge the next paragraph on heterogeneous into this one, and fix spelling of "charater" to character. * pg 165, ln 10: change "are using" to "use" * pg 167, ln 38-39: Question - Do we want to also list MPI_Accumulate with a forward reference to it? * pg 170, ln 3: change "is using" to "uses" * pg 171, ln 38: should be "In such a case" or better "In this case" * pg 171, ln 43: change "is using" to "uses" and "an intracommunicators" to the singular * pg 174, ln 11: change "illegal" to "invalid" * pg 174, ln 12-20: Delete the entire rationale. It was included in MPI-1 to justify not having an exclusive scan. Doesn't apply any more. You deleted only the last sentence, which was explicit, but the whole rationale should go. * pg 175, ln 13-20: By the same token, delete the entire rationale here. It was used to justify adding an exclusive scan to MPI-2. Even the "no in-place version" is covered above in lin 5. Maybe you can change that line 5 to include the justification as follows: "No in-place version is supported because it is not clear what this means for the process with rank zero". Note that it should be "process with rank zero" rather than "process for rank zero" as in the current text. * pg 175, ln 25: change "is using" to "uses" Rest looks great! > -----Original Message----- > From: mpi-21-bounces_at_[hidden] > [mailto:mpi-21-bounces_at_[hidden]] On Behalf Of Rolf > Rabenseifner > Sent: Saturday, April 12, 2008 10:15 AM > To: MPI 2.1 Mailing List > Subject: Re: [Mpi-21] AUTHORS: new consistent version and > write-token to you > > Hi Adam, > > this is really a good merge that you have done. > Good ideas and very carefully implemented. > > The reviewers can see (nearly) all of your modifications in red. > I expect that your reviewersand the Forum will enjoy your > modifications. > > I have restored two minor mods you have done: > (Line numbers refer to the new file you'll get with svn update) > > coll.tex, line 412 - I restored the "intracommunicator" > because the categories describe the behavior only on > intracommunicators, not on all, e.g., ALLGATHER has > one result on intra and two on inter and the wording is > "the result". > > coll.tex, line 491 - restored the word "also". > These routines are valid for intra and inter. > Writing "apply to inter" may be interpreted that they are > only valid for inter. > Writing "also apply to inter" should solve this problem. > I also colored this sentence red. > > I hop, this isokay with you. > > Best regards > Rolf > > > On Fri, 11 Apr 2008 15:22:10 -0700 > Adam Moody <moody20_at_[hidden]> wrote: > > Hi Rolf, > > I have checked in my changes for the Collectives chapter. > It's ready to > > be reviewed by the collective chapter reviewers. > > -Adam > > > > Rolf Rabenseifner wrote: > > > > >>>write-token again to the chapter authors << > > >>>- but only until Saturday 8:00 European GMT << > > >>> > > >>> > > > > > >Hi all authors and reviewers, > > > > > >Please read this mail carfully, if you want to make any further > > >modification of the text. > > > > > >You must now call > > > svn update > > >to get the newest release!! > > >We have a again a consistent version. Nearly all of > > > > > > > http://www.hlrs.de/mpi/mpi21/doc/MPI-2.1draft-2008-02-23-review.txt > > > > > >is now implemented in the tex-source. > > >Only a few items are still open, see statistics at the beginning. > > > > > >You can view the current version (it is svn revision 92) at > > > > > > http://www.hlrs.de/mpi/mpi21/doc/mpi-report.pdf > > > > > >It is only today and Friday remaining for finishing MPI-2.1. > > >I give back the write-token to the authors. > > > > > >On Saturday morning European time, I take the write-token > > >automatically on Saturday morning to build final pdf for the > > >reviews. > > > > > >As I can see, many chapter authors made fine cleaning to the text. > > >All larger changes should be defered to MPI-2.2! > > >Especially the merge of MPI Environment (inquiry.tex), > Misc (misc-2.tex) > > >and External Interfaces (ei-2.tex) > > >to one chapter "MPI Environment and Miscellany" (with 45 pages) > > >may be defered also to MPI-2.2. > > > > > >Up to now, I was able to check most changes and they look fine. > > > > > >Best regards > > >Rolf > > > > > > > > > > > >Dr. Rolf Rabenseifner . . . . . . . . . .. email > rabenseifner_at_[hidden] > > >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-21 mailing list > > >mpi-21_at_[hidden] > > >http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-21 > > > > > > > > > > > _______________________________________________ > > mpi-21 mailing list > > mpi-21_at_[hidden] > > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-21 > > > > Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner_at_[hidden] > 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-21 mailing list > mpi-21_at_[hidden] > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-21 > >

1 0

Re: [Mpi-21] Relevance of C++ MPI binding of www.boost.org
by unknown＠example.com 16 Apr '08

16 Apr '08

On Apr 3, 2008, at 12:09 PM, Rolf Rabenseifner wrote: > which relevance has the MPI C++ binding of Boost ( http://www.boost.org/ > )? > In the actual version 1.35 > http://www.boost.org/users/news/version_1_35_0 > one can find a "C++-friendly interface to the standard Message > Passing Interface". > > The comment from boost: > "Although there exist C++ bindings for MPI, they offer little > functionality over the C bindings." > > Is there any knowledge of whether in real C++ programming, whether > the official C++ MPI binding or the C++ boost MPI binding is > more often used in the C++ community? Boost's C++ bindings have only been in an official Boost release for a few weeks, so there's no way that the user base is even a fraction of the user base of the official C++ bindings. As such, these bindings would certainly not be ready for standardization. And, as Jeff said, the MPI forum has already rejected "richer" object- oriented C++ bindings in favor of what we have now. I don't see any indication that this is going to change, and I don't think it should. We should not spend our time discussing bindings (or improved bindings) for every language out there. Rather, we should spend our time making sure that such bindings *can* be built on top of MPI's C or Fortran bindings. MPI is good at message passing, Python is good at being Python. - Doug

1 0

Re: [Mpi-21] AUTHORS: new consistent version and write-token to you
by unknown＠example.com 14 Apr '08

14 Apr '08

Hi Rolf, Yes, those changes are fine by me. Thanks, -Adam Rolf Rabenseifner wrote: >Hi Adam, > >this is really a good merge that you have done. >Good ideas and very carefully implemented. > >The reviewers can see (nearly) all of your modifications in red. >I expect that your reviewersand the Forum will enjoy your modifications. > >I have restored two minor mods you have done: >(Line numbers refer to the new file you'll get with svn update) > >coll.tex, line 412 - I restored the "intracommunicator" > because the categories describe the behavior only on > intracommunicators, not on all, e.g., ALLGATHER has > one result on intra and two on inter and the wording is > "the result". > >coll.tex, line 491 - restored the word "also". > These routines are valid for intra and inter. > Writing "apply to inter" may be interpreted that they are > only valid for inter. > Writing "also apply to inter" should solve this problem. > I also colored this sentence red. > >I hop, this isokay with you. > >Best regards >Rolf > > >On Fri, 11 Apr 2008 15:22:10 -0700 > Adam Moody <moody20_at_[hidden]> wrote: > > >>Hi Rolf, >>I have checked in my changes for the Collectives chapter. It's ready to >>be reviewed by the collective chapter reviewers. >>-Adam >> >>Rolf Rabenseifner wrote: >> >> >> >>>>>write-token again to the chapter authors << >>>>>- but only until Saturday 8:00 European GMT << >>>>> >>>>> >>>>> >>>>> >>>Hi all authors and reviewers, >>> >>>Please read this mail carfully, if you want to make any further >>>modification of the text. >>> >>>You must now call >>> svn update >>>to get the newest release!! >>>We have a again a consistent version. Nearly all of >>> >>> http://www.hlrs.de/mpi/mpi21/doc/MPI-2.1draft-2008-02-23-review.txt >>> >>>is now implemented in the tex-source. >>>Only a few items are still open, see statistics at the beginning. >>> >>>You can view the current version (it is svn revision 92) at >>> >>> http://www.hlrs.de/mpi/mpi21/doc/mpi-report.pdf >>> >>>It is only today and Friday remaining for finishing MPI-2.1. >>>I give back the write-token to the authors. >>> >>>On Saturday morning European time, I take the write-token >>>automatically on Saturday morning to build final pdf for the >>>reviews. >>> >>>As I can see, many chapter authors made fine cleaning to the text. >>>All larger changes should be defered to MPI-2.2! >>>Especially the merge of MPI Environment (inquiry.tex), Misc (misc-2.tex) >>>and External Interfaces (ei-2.tex) >>>to one chapter "MPI Environment and Miscellany" (with 45 pages) >>>may be defered also to MPI-2.2. >>> >>>Up to now, I was able to check most changes and they look fine. >>> >>>Best regards >>>Rolf >>> >>> >>> >>>Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner_at_[hidden] >>>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-21 mailing list >>>mpi-21_at_[hidden] >>>http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-21 >>> >>> >>> >>> >>> >>_______________________________________________ >>mpi-21 mailing list >>mpi-21_at_[hidden] >>http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-21 >> >> > > > >Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner_at_[hidden] >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-21 mailing list >mpi-21_at_[hidden] >http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-21 > > >

1 0

Re: [Mpi-21] AUTHORS+REVIEWERS - Draft Apr. 12, 2008
by unknown＠example.com 14 Apr '08

14 Apr '08

All authors, Adam and Torsten, all authors have their write-token to their chapter. This week, they still should correct small things and should switch the __ in the form to DON2. On Friday 18, I'll take back the write-token witout any further announcment and will produce the final version of MPI-2.1 that will go into "official reading". I.e., until this Thursday, April 17, all **authors** should fix what they can and what reviewers report. I will not fix any further bug. The responsibility is at the chapters authors. Best regards Rolf On Sun, 13 Apr 2008 14:35:51 -0400 Torsten Hoefler <htor_at_[hidden]> wrote: > Hi Rolf, Adam, > I understand from your mail Rolf, that Adam lost his write token and > that only you can apply the changes/comments to the chapter (in my case > the collectives chapter). Thus, I'll send my review results to the list > instead the chapter author directly. My general comment: very well done > Adam :). > > I have some smaller editing remarks and also found some small errors and > clarification possibilities (see below). > > > ---------------------------------------------------------------------------------- > > 1234567890123456789012345678901234567890123456789012345678901234567890123456789012 > > All references are based on MPI-2.1 Draft Apr. 12, 2008 > > ww.a __ pnnn.ll text (with nnn=page number and ll=line number > > text continued > > ... > > ww.b __ pnnn.ll text ... and so on (total line length <= 82 characters.) > > ---------------------------------------------------------------------------------- > Ok, here my remarks, many of them (especially EDIT or CLAR) can probably > be ignored, but I thought I'd bring it up: > > general: the line numbers are completely out of sync ... it's hard to > classify a line that is between two line numbers. > > 5.a EDIT p129.19-41 "all group members" vs. "all members of a group" > should use the same phrasing (parallel structure) > 5.b NICE p129.41 Reference to 5.11 not 5.11.1 > 5.c EDIT p129.45 remove "," (comma) > 5.d EDIT p129.43 replace "the same" with "a" > 5.e EDIT p129.45 "the same group" in line 43 conflicts with "group or > groups" > 5.f CLAR p131.1 add "with exceptions stated in the following" to first > sentence > 5.g EDIT p131.4 no linebreak (Chapter~\ref{...}) > 5.h ERR p131.31 replace "will" with "might" > 5.i ERR p132. "special communicator must be created" this is wrong, > later (in the bcast section) it is stated differently > ... I would just remove this part of the advice or say > "might be created" (see Bcast advice) > 5.j CLAR p131.37 add "depending on the operation performed after last > sentence (right now it sounds like ths user is free to > choose which argument to replace) > 5.k EDIT p133.32 "a" -> "an"? > 5.l EDIT p136.22 start sentence with "If the communicator is an > intracommunicator" as for all the other operations, > remove "for intracommunicators" in line 23 > 5.m EDIT p136.24 replace "," by "and" > 5.n EDIT p136.26-31 this paragraph is redundant (it's all already stated > earlier, do we want this redundancy in the standard?) > 5.o EDIT p136.32 move sentence to end of the previous paragraph (line > 25) > 5.p EDIT p136.37 "data is braodcasted" (not braodcast) > 5.q CLAR p136.38-39 the text talks about send and receive buffer argument, > but bcast only has a single buffer argument ... > 5.r CLAR p136.38 "must be consistent with the ..." what does the word > consistent mean in this context? This seems undefined. > I think it means that the signature (size, count) is > the same. We should say this explicitely. The same > term "consistent" is used in all following operations > (I'm not going to add it again) > 5.s EDIT p138.3 write "all processes" instead of "process i" (i is > never defined) > 5.t ERR p139.47 data is not necessarily placed in rank order. The > order is freely definably by the user in the displs[] > array. Or does the standard enforce the displs[] array > to preserve order? We might want to run this through > the forum (or just ignore it?)? > 5.u CLAR p136.42 add "The examples in this section are using > intracommunicators." as in p140.27 > 5.v EDIT p147.48 the last paragraph is redundant (has been stated > before for all collectives) > 5.w EDIT p157.35 why a separate subsection for Alltoallw but not for > Alltoallv? I'd remove this subsection > 5.x EDIT p157.36 the first two sentences should go to rationale (we > don't need to rationalize the operation in the > description) > 5.y EDIT p161.13 replace "Reduce" by "Reduction" or "Operations for > MPI_Reduce" > 5.z EDIT p162.13 state that the example is in Fortran (it's said for > all the C examples before) > 5.aa EDIT p163.10-12 I think we don't need to backref to MPI-1 here if we > consider MPI-2.1 a complete standard. > 5.ab CLAR p163.20 replace "bit" by "numeric" or "integer" > 5.ac EDIT p167.41 doubled sentence "The order of ..." > 5.ad CLAR p169.2 add: (without supporting the "in place" option) > 5.ae ERR p169.16-20 the example will deadlock if root == groupsize-1. Has > been there for a while :). Use Irecv/Wait instead. > 5.af EDIT p171.48 state that example is in Fortran > 5.ag ERR p174.12-20 this rationale should be erased completely because it > explains (for MPI-1), why only the inclusive is supported ;-) > 5.ah ERR p175.20 erase the last sentence in rationale, we don't need > this anymore (is MPI-1) > 5.ai EDIT p177.48 move orphaned Example header to next page (might > change after edits though -> check chapter for orphans) > 5.aj EDIT p179.24 the font suddenly changes? (sans-serif) > 5.ak CLAR p179.27 maybe we should add: "and to ensure appropriate > matching (for deterministic behavior) > > Best, > Torsten > > -- > bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- > Indiana University | http://www.indiana.edu > Open Systems Lab | http://osl.iu.edu/ > 150 S. Woodlawn Ave. | Bloomington, IN, 474045-7104 | USA > Lindley Hall Room 135 | +01 (812) 855-3608 > _______________________________________________ > mpi-21 mailing list > mpi-21_at_[hidden] > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-21 Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner_at_[hidden] 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-21] AUTHORS+REVIEWERS - Draft Apr. 12, 2008
by unknown＠example.com 13 Apr '08

13 Apr '08

Bronis, > Re: > > 5.p EDIT p136.37 "data is braodcasted" (not braodcast) > > This is not correct. The preferred spelling for the past tense > of "broadcast" is "broadcast" ("broadcasted" is an acceptable > alternative but I don't see why we would want to switch to it). ok, I agree - didn't know this! So this clearly is one of the items to discard (there might be more). Thanks, Torsten -- bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- Indiana University | http://www.indiana.edu Open Systems Lab | http://osl.iu.edu/ 150 S. Woodlawn Ave. | Bloomington, IN, 474045-7104 | USA Lindley Hall Room 135 | +01 (812) 855-3608

1 0

Re: [Mpi-21] AUTHORS+REVIEWERS - Draft Apr. 12, 2008
by unknown＠example.com 13 Apr '08

13 Apr '08

Torsten, Adam and Rolf: Re: > 5.p EDIT p136.37 "data is braodcasted" (not braodcast) This is not correct. The preferred spelling for the past tense of "broadcast" is "broadcast" ("broadcasted" is an acceptable alternative but I don't see why we would want to switch to it). Bronis

1 0

Re: [Mpi-21] REVIEWERS - Draft Apr. 12, 2008
by unknown＠example.com 13 Apr '08

13 Apr '08

Rolf, > if you find a bug, please write also the needed correction > in your review. done > Your reviews will be collected and have to go directly into the Forum > meeting and must be handled by the Forum. > Therefore, please use the format described below. oh, I did not know this before sending my review. I would not recommend to feed the results of my review directly to the forum because many things are small typos or minor editing comments that would be a waste of time for the whole forum. So we might want to filter my review or just apply the smaller changes in collaboration with the chapter author before presenting it to the forum. Thanks & Best, Torsten -- bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- Indiana University | http://www.indiana.edu Open Systems Lab | http://osl.iu.edu/ 150 S. Woodlawn Ave. | Bloomington, IN, 474045-7104 | USA Lindley Hall Room 135 | +01 (812) 855-3608

1 0

Re: [Mpi-21] AUTHORS+REVIEWERS - Draft Apr. 12, 2008
by unknown＠example.com 13 Apr '08

13 Apr '08

Hi Rolf, Adam, I understand from your mail Rolf, that Adam lost his write token and that only you can apply the changes/comments to the chapter (in my case the collectives chapter). Thus, I'll send my review results to the list instead the chapter author directly. My general comment: very well done Adam :). I have some smaller editing remarks and also found some small errors and clarification possibilities (see below). > ---------------------------------------------------------------------------------- > 1234567890123456789012345678901234567890123456789012345678901234567890123456789012 > All references are based on MPI-2.1 Draft Apr. 12, 2008 > ww.a __ pnnn.ll text (with nnn=page number and ll=line number > text continued > ... > ww.b __ pnnn.ll text ... and so on (total line length <= 82 characters.) > ---------------------------------------------------------------------------------- Ok, here my remarks, many of them (especially EDIT or CLAR) can probably be ignored, but I thought I'd bring it up: general: the line numbers are completely out of sync ... it's hard to classify a line that is between two line numbers. 5.a EDIT p129.19-41 "all group members" vs. "all members of a group" should use the same phrasing (parallel structure) 5.b NICE p129.41 Reference to 5.11 not 5.11.1 5.c EDIT p129.45 remove "," (comma) 5.d EDIT p129.43 replace "the same" with "a" 5.e EDIT p129.45 "the same group" in line 43 conflicts with "group or groups" 5.f CLAR p131.1 add "with exceptions stated in the following" to first sentence 5.g EDIT p131.4 no linebreak (Chapter~\ref{...}) 5.h ERR p131.31 replace "will" with "might" 5.i ERR p132. "special communicator must be created" this is wrong, later (in the bcast section) it is stated differently ... I would just remove this part of the advice or say "might be created" (see Bcast advice) 5.j CLAR p131.37 add "depending on the operation performed after last sentence (right now it sounds like ths user is free to choose which argument to replace) 5.k EDIT p133.32 "a" -> "an"? 5.l EDIT p136.22 start sentence with "If the communicator is an intracommunicator" as for all the other operations, remove "for intracommunicators" in line 23 5.m EDIT p136.24 replace "," by "and" 5.n EDIT p136.26-31 this paragraph is redundant (it's all already stated earlier, do we want this redundancy in the standard?) 5.o EDIT p136.32 move sentence to end of the previous paragraph (line 25) 5.p EDIT p136.37 "data is braodcasted" (not braodcast) 5.q CLAR p136.38-39 the text talks about send and receive buffer argument, but bcast only has a single buffer argument ... 5.r CLAR p136.38 "must be consistent with the ..." what does the word consistent mean in this context? This seems undefined. I think it means that the signature (size, count) is the same. We should say this explicitely. The same term "consistent" is used in all following operations (I'm not going to add it again) 5.s EDIT p138.3 write "all processes" instead of "process i" (i is never defined) 5.t ERR p139.47 data is not necessarily placed in rank order. The order is freely definably by the user in the displs[] array. Or does the standard enforce the displs[] array to preserve order? We might want to run this through the forum (or just ignore it?)? 5.u CLAR p136.42 add "The examples in this section are using intracommunicators." as in p140.27 5.v EDIT p147.48 the last paragraph is redundant (has been stated before for all collectives) 5.w EDIT p157.35 why a separate subsection for Alltoallw but not for Alltoallv? I'd remove this subsection 5.x EDIT p157.36 the first two sentences should go to rationale (we don't need to rationalize the operation in the description) 5.y EDIT p161.13 replace "Reduce" by "Reduction" or "Operations for MPI_Reduce" 5.z EDIT p162.13 state that the example is in Fortran (it's said for all the C examples before) 5.aa EDIT p163.10-12 I think we don't need to backref to MPI-1 here if we consider MPI-2.1 a complete standard. 5.ab CLAR p163.20 replace "bit" by "numeric" or "integer" 5.ac EDIT p167.41 doubled sentence "The order of ..." 5.ad CLAR p169.2 add: (without supporting the "in place" option) 5.ae ERR p169.16-20 the example will deadlock if root == groupsize-1. Has been there for a while :). Use Irecv/Wait instead. 5.af EDIT p171.48 state that example is in Fortran 5.ag ERR p174.12-20 this rationale should be erased completely because it explains (for MPI-1), why only the inclusive is supported ;-) 5.ah ERR p175.20 erase the last sentence in rationale, we don't need this anymore (is MPI-1) 5.ai EDIT p177.48 move orphaned Example header to next page (might change after edits though -> check chapter for orphans) 5.aj EDIT p179.24 the font suddenly changes? (sans-serif) 5.ak CLAR p179.27 maybe we should add: "and to ensure appropriate matching (for deterministic behavior) Best, Torsten -- bash$ :(){ :|:&};: --------------------- http://www.unixer.de/ ----- Indiana University | http://www.indiana.edu Open Systems Lab | http://osl.iu.edu/ 150 S. Woodlawn Ave. | Bloomington, IN, 474045-7104 | USA Lindley Hall Room 135 | +01 (812) 855-3608

1 0

REVIEWERS - Draft Apr. 12, 2008
by unknown＠example.com 13 Apr '08

13 Apr '08

Dear reviewer, if you find a bug, please write also the needed correction in your review. (If the correction is not obvious, it may be a MPI-2.2 problem). Your reviews will be collected and have to go directly into the Forum meeting and must be handled by the Forum. Therefore, please use the format described below. Best regards Rolf On Sun, 13 Apr 2008 00:05:37 +0200 "Rolf Rabenseifner" <rabenseifner_at_[hidden]> wrote: > Dear chapter author, dear reviewer, > > This e-mail is the first preparation of the official reading. > Please read carefully. > > Again,we have a stable and consistent draft. > It is SVN revision 108 and pdf and ps are on > http://www.hlrs.de/mpi/mpi21/doc/ > (username and password are mpi21) > > Nearly all chapters have received an update by the chapters' authors. > Not finished is - as far as I can see > - Point-to-point (may-be without MPI-1/2 problems] (Rich Graham) > - Datatypes [it was split from Point-to-point] (Rich Graham) > and full MPI-2 based chapters (maybe without need for updates) > - Terms (Tony Skjellum) > - External Interfaces (Bronis de Supinski) > - Profiling (Bronis de Supinski) > Therefore, review can start in all chapters, except the datatype chapter. > > IMPORTANT for authors: > If you make further modifications, please be sure that you use > the macros and that your modification (or at least an delete-anchor) > is printed **red**. (Details on using the macros, see below.) > Please check red color in the resulting mpi-report.pdf before you > svn commit your changes. > As far as I can see, all changes up to now may have no problem > for acceptance by the Forum. > At this late stage, I would recommend to do only changes that do > not need intensive debates at the Forum meeting. Such things > would be MPI-2.2 modifications. > > IMPORTANT for reviewers: > - Please check very carefully all red or magenta changes. > Modifications are done by the chapter's author (red) or > by me (magenta). > You may be the first reviewer who read and check this text. > - Please check carefully the cyan text. > Those modifications are proposed by a reviewer and have been > implemented by my > You may be the second reviewer for those modifications. > - All green stuff should be fine, because it is based on > Forum's decisions (Ballot 1-4). > - Original text is black (MPI-1.3) or blue (MPI-2.0). > - Please make all comments in the format > > ---------------------------------------------------------------------------------- > 1234567890123456789012345678901234567890123456789012345678901234567890123456789012 > All references are based on MPI-2.1 Draft Apr. 12, 2008 > ww.a __ pnnn.ll text (with nnn=page number and ll=line number > text continued > ... > ww.b __ pnnn.ll text ... and so on (total line length <= 82 characters.) > ---------------------------------------------------------------------------------- > > - Please send only review items that are in the category > -- ERROR, must be fixed. > - If you want to send an item in the categories > -- Nice to have, > -- Editorial proposal (like move xxx to yyy), > -- Clarification which needs Forum decisions, > -- ... > then please substitute the __ by 22, because then it should be > handled in MPI-2.2. > > This format is used in > http://www.hlrs.de/mpi/mpi21/doc/MPI-2.1draft-2008-02-23-review.txt > where all reviews are filed in. > (I'll substitute your xx by the next number. If you have more than > 26 items, then please continue with xx.a -- xx.z, yy.a - ...) > > Please use this format. I'll have only a few minutes between two flights > to put all reviews together. > > WORKPLAN: > - The authors get back the write-token until Thursday, Apr. 17. > (If you need also Friday Apr. 18, then please inform me!!) > - All stuff that is not done until that point, will be defered to MPI-2.2! > - I'm on travel and may not be able to respond on e-mails this week. > - On Friday, Apr. 18 (or Saturday, Apr. 19), I will make the > final version for all final reviews and for the official reading. > - Version Apr. 17 together with all your review mails will be the > basis for the official reading. > - Reviews until Friday, Apr. 25. according the format above. > - Saturday, Apr. 26: > I will put all reviews together to have unique numbers. > > Best regards > Rolf > > > Annex: The rules for authors: > > The goals of your work as chapter author: > ----------------------------------------- > > - Remove references to MPI-1 and MPI-2. > - Substitute by MPI. > - Necessary modifications to achieve this goal. > > With this, we should have a single MPI-2.1 standard that does > not "know" the MPI-1 or MPI-2 history of individual functions. > > ** This should be mainly a task in the area of chapter-introductions. > ** Please, never change the wording of function definitions. > > Exceptions: > - There are routines that are deprecated and that are > already referenced in a consistent way. > My recommendation: > Current wording: > There are *new* function, and existing are now deprecated. > Proposal: > There are functions, and there exist also deprecated functions > with (nearly) same functionality but deprecated due to some lack > in the bindings (or functionality) > - There is history information in the frontmatter. > - there is a change-log annex with limited memory (only previous version) > > Technical editing rule: > ----------------------- > > You must identify all of your changes: > > a) new/modified wording - you are highlighting your new/modified wording with: > > \mpiiidotiMergeNEWforSINGLEbegin% MPI-2.1 round-two - begin of modification > .... your modified / new wording .... > \mpiiidotiMergeNEWforSINGLEendI% MPI-2.1 round-two - end of modification > > b) Moved paragraphs, sentences, ... - you are highlighting the first word > of the moved text with: > > \mpiiidotiMergeNEWforSINGLEbegin% MPI-2.1 round-two - begin of text-move > First-word > \mpiiidotiMergeNEWforSINGLEendI% MPI-2.1 round-two - ongoing text-move > ... rest of the moved text.... > % MPI-2.1 round-two - end of text-move > > c) If you want to delete text (i.e. not only a move to another location), > the please us > > \mpiiidotiMergeNEWforSINGLEbegin% MPI-2.1 round-two - deleted-text-anchor > Last word of the previous text > \mpiiidotiMergeNEWforSINGLEendI% MPI-2.1 round-two - end of anchor > > % MPI-2.1 round-two - begin of deteted text - Draft Apr.1, 2008 page/line-page/line > % the text you delete must be preceded with "% " > % deprecated text > % ... > % deprecated text > % MPI-2.1 round-two - end of deteted text > > Additionally, you must add this outcommented text to the end of > removed_source.tex > > Caution: These macros work only in black parts of the text. > > In blue parts you must substitute > \mpiiidotiMergeNEWforSINGLEendI% > by > \mpiiidotiMergeNEWforSINGLEendII% > > > The highlighting is done with red color. > (There are only a few other sentences from me magenta. > Therefore red should be the best to highlight the round-two modifications.) > > ----------------------------------------------------- > Using svn: > Before you start to make changes: > > cd ....your_directory..../trunk/MPI-2.1/ > svn update > > to get the newest sources. > > After you made changes, please call make and check mpi-report.pdf > to guarantee that all your changes are really highlighted with *red* color. > > When you make the first svn commit after you received a new username with > write-permission, you must use the aditional option > > --username your_new_user_name > > Please commit with > > svn commit chap-xxxx/xxxx.tex -m 'MPI-2.1 - your name - what you have done' > > Please commit as soon as you finished the changes and after you have > checked the red color. > > -------------------------------------------------- > > Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner_at_[hidden] > 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-21 mailing list > mpi-21_at_[hidden] > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-21 Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner_at_[hidden] 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