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

Chapter 3-4 minor comments and suggestions
by unknown＠example.com 29 Apr '08

29 Apr '08

dear all, ome final, minor comments to chapters 3 and 4 jesper *

1 0

Chapter 5 comments/suggestions
by unknown＠example.com 29 Apr '08

29 Apr '08

hi all uggestions/changes for Chapter 5 jesper (and adam) *

1 0

Chapter 1 && 2 minor comments
by unknown＠example.com 29 Apr '08

29 Apr '08

Dear all, here are some minor typos and sugestiosn for chaps 1&2 Jesper --- General: "message-passing" or "message passing"? Chapter 1: p.4, line 3: electonically -> electronically p.4, line 38: "hybrid architectures," - add space p.5 (ff): Update "Miscellany chapter" to something with info object p.6, line 17: drop "more convenient methods of constructing intercommunicators" (not in that chapter) p.6, line 22: correct chapter name (add "Caching") Chapter 2: p.94).", line 14: MPI-2 -> MPI p.9, line 15-19: drop (or reformulate) p.9, line 40-45: reformulate: "In many cases MPI names for C functions are of the form Class_action_subset. This convention originated with MPI-1. Since MPI-2 an attempt has been made to standardize the names of MPI functions according to the following rules. The C++ bindings in particular follow these rules (see Sectoion 2.6)." p.10, line 15: MPI-1 functions -> some MPI functions (that were defined during the MPI-1 process) p.10, line 26-30: bullet unintelligible. Change: An argument whose value may be used by a call but may not be updated is marked as IN An argument that may be updated by a call but whose value will not be used is marked as OUT An argument whose value may be used and that may be updated is marked as INOUT p.10, line 47: too much blank LaTeX space p.11, line 38: add: ".it is freed." -> ".it is freed, and becomes inactive if it was persistent." p.16, line 11: since MPI-2. -> since better solutions were provided with MPI-2. p.16, line 14: replace with: "With MPI-2, these functions were given new names with new bindings for the address arguments." p.16, line 19: "error-prone, ." -> "error-prone. The MPI-2 ." p.16, line 28: Change to "Originally MPI-1.1 provided bindings for Fortran 77. These bindings are retained, but they are now." p.16, line 45: Give some reason or excuse: "Unfortunately but unavoidably, the MPI Fortran bindings." p.18, line 6: MPI-2 -> MPI p.18, line 10: can't -> cannot p.19, line 45: MPI-2 -> MPI p.19, line 46: "the new MPI-2." -> "an MPI function is related to a function defined already for MPI-1 p.19, line 48: MPI-2 -> MPI p.20, line 2: drop "for MPI-1" p.20, line 3: "MPI-1 name is" -> "MPI name may be" p.23, line 3: drop latter part of sentence, "are described in Chapter ". p.23, line 32: MPI-2 -> MPI

1 0

MPI-2.1: Review of Chapter 1 and buglets in Chapter 15
by unknown＠example.com 29 Apr '08

29 Apr '08

Dear Rolf, please find enclosed the changes to Chapter 1 in the file below. Also, please note, that in Chapter 15, starting with p441, l38 the word language is misspelled several times: p441, l38 -- laguage --> language. With best regards, Rainer -- ---------------------------------------------------------------- Dipl.-Inf. Rainer Keller http://www.hlrs.de/people/keller HLRS Tel: ++49 (0)711-685 6 5858 Nobelstrasse 19 Fax: ++49 (0)711-685 6 5832 70550 Stuttgart email: keller_at_[hidden] Germany AIM/Skype:rusraink *

1 0

Re: [Mpi-21] Formal MPI-2.1 Review
by unknown＠example.com 28 Apr '08

28 Apr '08

Hello Rolf, please find enclosed the changes to MPI-2.1 Chapter12 (compared to the version April 19 2008). Thanks, Rainer On Monday 21 April 2008 20:38, Rolf Rabenseifner wrote: > First, I want to thank all chapter authors for their fine work > cleaning up all the MPI-1/-2 problems and achieving a single monolithic > MPI-2.1 text. > > Dear reviewer, > > to handle about 20 chapters with in average 5 reviews per chapter, > I would like to formalize it: > > On next Monday (Apr. 28, 2008), I would like to ask you whether you can > make the following statement on the chapter(s) you voloteered for reviewing > (I'll ask independently each reviewer on each chapter): > > --------------------------------------------------------------------------- > Chap. ......... of MPI-2.1 Draft Apr. 19, 2008 represents fully and > exactly (as I can overlook) the text that was given by MPI-1.1 (June 1995), > MPI-2 (July 1997), MPI-1 errata (Oct. 12, 1998), Ballot 1+2 (May 15, 2002), > Ballot 3 (protocol from March 2008 meeting, 1st vote), > and Ballot 4 (protocol from March 2008 meeting, official reading) > if the following text-modifications (they are already implemented in the > current draft) are accepted by the MPI Forum: > - XXXXX > and if the following bugs will be corrected until first vote > in June 2008 meeting: > - YYYYY. > Further comments (not official part of this review): > - ZZZZZ > --------------------------------------------------------------------------- > > Please do not use ZZZZZ for MPI-2.2 items or for any further enhancement. > > If there are such XXXXX, YYYYY, and ZZZZZ items, then I can include them > into my slides, if you are sending it to me not later than > > Friday, April 25, 2008. > > All items that you detect later, you may bring to the MPI Forum > with your own slide, plugging in your own laptop at the meeting. > > And same procedure for MPI-1.3 with only MPI-1 input. > > I hope this review process is fine with you. > > Exception: Review of Chap. 3 and 4 will be based on the text > prepared by Rich Graham together with his reviewers this > week, i.e. final reviews at the meeting will be done > on MPI-2.1 Draft Apr. 26, 2008. > > Best regards from travel (currently Salishan conference, USA) > Rolf > > --------------------------------------------------------------------------- > Frontmatter: > - Bill Gropp, Rusty Lusk > Acknowledgements: > - Richard Graham, Bill Gropp, Rusty Lusk > Chap. 1: Introduction to MPI > - Bill Gropp, Rusty Lusk, Karl Feind, Adam Moody, Jesper Traeff > Chap. 2: MPI-2 Terms and Conventions > - Tony Skjellum, Bill Gropp, Richard Barrett, Jesper Traeff > Chap. 3: Point-to-Point Communication > - Rich Graham, Jespar Larsson Traeff, George Bosilca, > Steve Poole, Kannan Narasimhan, David Solt, B. Gropp, Matt Koop, > Adam Moody > Chap. 3: Datatypes > - Rich Graham, Jespar Larsson Traeff, George Bosilca, > Steve Poole, Kannan Narasimhan, David Solt, B. Gropp, Matt Koop, > Adam Moody > Chap. 5: Collective Communication > - Adam Moody, Steven Ericsson-Zenith, Edgar Gabriel, Rajeev > Thakur, Bill Gropp, Georg Bosilca, Torsten Hoefler, Jesper Traeff Chap. 6: > Groups, Context, and Communicators > - Richard Treumann, Steven Ericsson-Zenith, Edgar Gabriel, Tony > Skjellum, Bill Gropp, Georg Bosilca, Robert Blackmore > Chap. 7: Process Topologies > - Jesper Traeff, Rusty Lusk, Bill Gropp, Richard Barrett > Chap. 8: MPI Environmental Management (incl. sections from MPI-2 Misc.) > - George Bosilca, Rich Graham, Jespar Larsson Traeff, > Steve Poole, Kannan Narasimhan, David Solt, B. Gropp > Chap. 9: Miscellany > - Jespar Larsson Traeff, Rich Graham, George Bosilca, > Steve Poole, Kannan Narasimhan, B. Gropp > Chap.10: Process Creation and Management > - David Solt, Dries Kimpe, Rusty Lusk, Georg Bosilca, Bill Gropp, > Kalem Karian > Chap.11: One-Sided Communication > - Jespar Larsson Traeff, Ericsson-Zenith, Martin Schulz, > Bill Gropp, Darius Buntinas > Chap.12: External Interfaces > - Bronis de Supinski, Bill Gropp > Chap.13: I/O > - Rajeev Thakur, Joachim Worringen, Bill Gropp, Rainer Keller > Chap.14: Profiling Interface > - Bronis de Supinski, Bill Gropp, Jeff Brown > Chap.15: Deprecated Functions > - Rolf Rabenseifner, Bill Gropp > Chap.16: Language Bindings > - Jeff Squyres, Steve Poole, Purushotham Bangalore, > Bill Gropp, Erez Haba, Alexander Supalov > Bibliography: > - Rusty Lusk, Bill Gropp > Annex A: Language Bindings Summary > - Alexander Supalov, Jeff Squyres, Steve Poole, Purushotham > Bangalore, Bill Gropp, > Annex B: Change Log > - Rolf Rabenseifner, Bill Gropp > Index: > - Rusty Lusk, Bill Gropp > > MPI-1.3: Bill Gropp, Rolf Rabenseifner, Adam Moody, Puri Bangalore, > Terry Dontje > > > > > 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 -- ---------------------------------------------------------------- Dipl.-Inf. Rainer Keller http://www.hlrs.de/people/keller HLRS Tel: ++49 (0)711-685 6 5858 Nobelstrasse 19 Fax: ++49 (0)711-685 6 5832 70550 Stuttgart email: keller_at_[hidden] Germany AIM/Skype:rusraink *

1 0

problems with 19 Apr document
by unknown＠example.com 28 Apr '08

28 Apr '08

In reviewing the 19 April document for the Chicago meeting, I found a few issues: Chapter 15 (deprecated functions) - Seems to be missing some deprecated functions: MPI_DUP_FN, MPI_NULL_COPY_FN, MPI_NULL_DELETE_FN (there might be others?). Annex A - p498:41-42: The table header for the 2nd to last table on the page is a little weird: "C Constants specifying ignored input (no C++ or Fortran)", and the constants are MPI_F_STATUS[ES]_IGNORE. So the heading is technically correct, but a little confusing. - p556: this table represents (I think) all MPI-1 functions. Do we need to add in MPI-2 functions? Or do we still need this table anymore, since all the C++ functions are now integrated in the whole document? Examples Index - There are still a bunch of repeated names, some in all caps, some in mixed case, etc. - There are some examples listed just by MPI function name (e.g., MPI_SEND and MPI_Send) -- are we listing every MPI function in every MPI example? How was the selection for these example names given? Constant Index - The first entry is still "MPI::*_NULL" - Are we listing the C++ constants and Fortran constants here, too? Or just the C constants? Or just language neutral? (I only see C++ predefined datatypes listed -- should we list all or none of them?) Declarations Index - I don't see the MPI C++ subtypes (e.g., MPI::Intercomm). Callback function prototype index - C++ types are missing Function index - There are many entries that should not be in this index, because they are not functions. I suspect that someone simply used the wrong latex macro in the text. MPI_ MPI_*_FREE MPI_CHAR MPI_CHARACTER MPI_COMM_NULL MPI_COMM_SELF MPI_COMM_WORLD MPI_DUP_FN MPI_ERROR MPI_ERRORS_RETURN MPI_FILE MPI_FILE_NULL MPI_MAX MPI_STATUS MPI_SUCCESS MPI_THREAD_INIT (!! looks like a typo page p479) MPI_TYPE_UB, MPI_TYPE_LB (these two functions are on one line) MPI_UNSIGNED_CHAR MPI_WCHAR MPI_WIN_BASE MPI_XXX_GET_ERRHANDLER MPI_YYY_DELETE_ATTR MPI_YYY_FREE_KEYVAL MPI_YYY_GET_ATTR MPI_YYY_SET_ATTR -- Jeff Squyres Cisco Systems

1 0

Meeting Apr. 28-30 - MPI-1.3 and MPI-2.1
by unknown＠example.com 27 Apr '08

27 Apr '08

Hi all, latest MPI-1.3 and MPI-2.1 preparation of the meeting is done. All relelant files can be found in the first section of http://www.hlrs.de/mpi/mpi21/doc/ Username and pw as usual: mpi21 I'm leaving now my office and the internet. All further reviews: please have them available on your laptop and (if possible/necessary) a printed version for me. See you on Monday - 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)

1 0

Re: [Mpi-21] Formal MPI-2.1 Review
by unknown＠example.com 26 Apr '08

26 Apr '08

Rolf, As you have the write token I didn't want to commit the changes suggested by Kannan. Attached you will find the diff regarding the chapter 8 (inquiry.tex). george. On Apr 26, 2008, at 1:13 PM, Narasimhan, Kannan wrote: > Rolf, > > Apologies for missing the April 25 deadline. I've reviewed Chapters > 8 and 9 in the MPI 2.1 draft, and it looks good -- the authors have > done a wonderful job with the merge. > > Here is the "informal" review statement (since I wasn't sure if my > minor edits fall under XXXX or YYYY category) of the formal > statement :). > > The following is the list of edits required in each of these chapters: > > Chapter 8 ( MPI Environment Management ) > > 1. Pg 265, 8.3 Error Handling (Line 12): Remove the phrase "On the > other hand" and reword the sentence as follows: C and C++ have > distinct typedefs for user defined error handling callback functions > that accept communicator, file, and window arguments. > > 2. Pg 266, 8.3.1 Error Handlers for Communicators (Lines 22-26): > Change all references to HANDLER_FUNCTION to > {COMM,WIN,FILE}_ERRHANDLER_FN in the "Advice to users" para. > > 3. Pg 268, 8.3.2 Error Handlers for Windows (Line 13): ADD: Creates > an error handler that can be attached to a window object. > > 4. Pg 269, 8.3.3 Error Handlers for Files (Line 24): ADD: Creates an > error handler that can be attached to a file object. > > 5. Pg 270, 8.3.4 Freeing Error handlers and Retrieving Error > Strings(Line 27-28): Change the sentence to: The error handler will > be deallocated after all of the objects associated with it > (communicator, window, or file) are deallocated. > > Chapter 9 (Miscellany) > > 1. Pg 283, 9.1 The Info Object (Line 26): Change the first sentence > to: Many of the routines in MPI take an info argument. > > -Kannan- > > > -----Original Message----- > From: mpi-21-bounces_at_[hidden] [mailto:mpi-21-bounces_at_[hidden] > ] On Behalf Of Rolf Rabenseifner > Sent: Monday, April 21, 2008 1:39 PM > To: MPI 2.1 Mailing List > Subject: [Mpi-21] Formal MPI-2.1 Review > > First, I want to thank all chapter authors for their fine work > cleaning up all the MPI-1/-2 problems and achieving a single > monolithic > MPI-2.1 text. > > Dear reviewer, > > to handle about 20 chapters with in average 5 reviews per chapter, I > would like to formalize it: > > On next Monday (Apr. 28, 2008), I would like to ask you whether you > can make the following statement on the chapter(s) you voloteered > for reviewing (I'll ask independently each reviewer on each chapter): > > --------------------------------------------------------------------------- > Chap. ......... of MPI-2.1 Draft Apr. 19, 2008 represents fully and > exactly (as I can overlook) the text that was given by MPI-1.1 (June > 1995), > MPI-2 (July 1997), MPI-1 errata (Oct. 12, 1998), Ballot 1+2 (May 15, > 2002), Ballot 3 (protocol from March 2008 meeting, 1st vote), and > Ballot 4 (protocol from March 2008 meeting, official reading) if the > following text-modifications (they are already implemented in the > current draft) are accepted by the MPI Forum: > - XXXXX > and if the following bugs will be corrected until first vote in June > 2008 meeting: > - YYYYY. > Further comments (not official part of this review): > - ZZZZZ > --------------------------------------------------------------------------- > > Please do not use ZZZZZ for MPI-2.2 items or for any further > enhancement. > > If there are such XXXXX, YYYYY, and ZZZZZ items, then I can include > them into my slides, if you are sending it to me not later than > > Friday, April 25, 2008. > > All items that you detect later, you may bring to the MPI Forum with > your own slide, plugging in your own laptop at the meeting. > > And same procedure for MPI-1.3 with only MPI-1 input. > > I hope this review process is fine with you. > > Exception: Review of Chap. 3 and 4 will be based on the text > prepared by Rich Graham together with his reviewers this > week, i.e. final reviews at the meeting will be done > on MPI-2.1 Draft Apr. 26, 2008. > > Best regards from travel (currently Salishan conference, USA) Rolf > > --------------------------------------------------------------------------- > Frontmatter: > - Bill Gropp, Rusty Lusk > Acknowledgements: > - Richard Graham, Bill Gropp, Rusty Lusk Chap. 1: > Introduction to MPI > - Bill Gropp, Rusty Lusk, Karl Feind, Adam Moody, Jesper > Traeff Chap. 2: MPI-2 Terms and Conventions > - Tony Skjellum, Bill Gropp, Richard Barrett, Jesper Traeff > Chap. 3: Point-to-Point Communication > - Rich Graham, Jespar Larsson Traeff, George Bosilca, > Steve Poole, Kannan Narasimhan, David Solt, B. Gropp, Matt > Koop, > Adam Moody > Chap. 3: Datatypes > - Rich Graham, Jespar Larsson Traeff, George Bosilca, > Steve Poole, Kannan Narasimhan, David Solt, B. Gropp, Matt > Koop, > Adam Moody > Chap. 5: Collective Communication > - Adam Moody, Steven Ericsson-Zenith, Edgar Gabriel, Rajeev > Thakur, > Bill Gropp, Georg Bosilca, Torsten Hoefler, Jesper Traeff > Chap. 6: Groups, Context, and Communicators > - Richard Treumann, Steven Ericsson-Zenith, Edgar Gabriel, > Tony > Skjellum, Bill Gropp, Georg Bosilca, Robert Blackmore > Chap. 7: Process Topologies > - Jesper Traeff, Rusty Lusk, Bill Gropp, Richard Barrett > Chap. 8: MPI Environmental Management (incl. sections from MPI-2 > Misc.) > - George Bosilca, Rich Graham, Jespar Larsson Traeff, > Steve Poole, Kannan Narasimhan, David Solt, B. Gropp Chap. > 9: Miscellany > - Jespar Larsson Traeff, Rich Graham, George Bosilca, > Steve Poole, Kannan Narasimhan, B. Gropp > Chap.10: Process Creation and Management > - David Solt, Dries Kimpe, Rusty Lusk, Georg Bosilca, Bill > Gropp, > Kalem Karian > Chap.11: One-Sided Communication > - Jespar Larsson Traeff, Ericsson-Zenith, Martin Schulz, > Bill Gropp, Darius Buntinas > Chap.12: External Interfaces > - Bronis de Supinski, Bill Gropp > Chap.13: I/O > - Rajeev Thakur, Joachim Worringen, Bill Gropp, Rainer Keller > Chap.14: Profiling Interface > - Bronis de Supinski, Bill Gropp, Jeff Brown > Chap.15: Deprecated Functions > - Rolf Rabenseifner, Bill Gropp > Chap.16: Language Bindings > - Jeff Squyres, Steve Poole, Purushotham Bangalore, > Bill Gropp, Erez Haba, Alexander Supalov > Bibliography: > - Rusty Lusk, Bill Gropp > Annex A: Language Bindings Summary > - Alexander Supalov, Jeff Squyres, Steve Poole, Purushotham > Bangalore, > Bill Gropp, > Annex B: Change Log > - Rolf Rabenseifner, Bill Gropp > Index: > - Rusty Lusk, Bill Gropp > > MPI-1.3: Bill Gropp, Rolf Rabenseifner, Adam Moody, Puri Bangalore, > Terry Dontje > > > > > 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] Formal MPI-2.1 Review
by unknown＠example.com 26 Apr '08

26 Apr '08

Rolf, Apologies for missing the April 25 deadline. I've reviewed Chapters 8 and 9 in the MPI 2.1 draft, and it looks good -- the authors have done a wonderful job with the merge. Here is the "informal" review statement (since I wasn't sure if my minor edits fall under XXXX or YYYY category) of the formal statement :). The following is the list of edits required in each of these chapters: Chapter 8 ( MPI Environment Management ) 1. Pg 265, 8.3 Error Handling (Line 12): Remove the phrase "On the other hand" and reword the sentence as follows: C and C++ have distinct typedefs for user defined error handling callback functions that accept communicator, file, and window arguments. 2. Pg 266, 8.3.1 Error Handlers for Communicators (Lines 22-26): Change all references to HANDLER_FUNCTION to {COMM,WIN,FILE}_ERRHANDLER_FN in the "Advice to users" para. 3. Pg 268, 8.3.2 Error Handlers for Windows (Line 13): ADD: Creates an error handler that can be attached to a window object. 4. Pg 269, 8.3.3 Error Handlers for Files (Line 24): ADD: Creates an error handler that can be attached to a file object. 5. Pg 270, 8.3.4 Freeing Error handlers and Retrieving Error Strings(Line 27-28): Change the sentence to: The error handler will be deallocated after all of the objects associated with it (communicator, window, or file) are deallocated. Chapter 9 (Miscellany) 1. Pg 283, 9.1 The Info Object (Line 26): Change the first sentence to: Many of the routines in MPI take an info argument. -Kannan- -----Original Message----- From: mpi-21-bounces_at_[hidden] [mailto:mpi-21-bounces_at_[hidden]] On Behalf Of Rolf Rabenseifner Sent: Monday, April 21, 2008 1:39 PM To: MPI 2.1 Mailing List Subject: [Mpi-21] Formal MPI-2.1 Review First, I want to thank all chapter authors for their fine work cleaning up all the MPI-1/-2 problems and achieving a single monolithic MPI-2.1 text. Dear reviewer, to handle about 20 chapters with in average 5 reviews per chapter, I would like to formalize it: On next Monday (Apr. 28, 2008), I would like to ask you whether you can make the following statement on the chapter(s) you voloteered for reviewing (I'll ask independently each reviewer on each chapter): --------------------------------------------------------------------------- Chap. ......... of MPI-2.1 Draft Apr. 19, 2008 represents fully and exactly (as I can overlook) the text that was given by MPI-1.1 (June 1995), MPI-2 (July 1997), MPI-1 errata (Oct. 12, 1998), Ballot 1+2 (May 15, 2002), Ballot 3 (protocol from March 2008 meeting, 1st vote), and Ballot 4 (protocol from March 2008 meeting, official reading) if the following text-modifications (they are already implemented in the current draft) are accepted by the MPI Forum: - XXXXX and if the following bugs will be corrected until first vote in June 2008 meeting: - YYYYY. Further comments (not official part of this review): - ZZZZZ --------------------------------------------------------------------------- Please do not use ZZZZZ for MPI-2.2 items or for any further enhancement. If there are such XXXXX, YYYYY, and ZZZZZ items, then I can include them into my slides, if you are sending it to me not later than Friday, April 25, 2008. All items that you detect later, you may bring to the MPI Forum with your own slide, plugging in your own laptop at the meeting. And same procedure for MPI-1.3 with only MPI-1 input. I hope this review process is fine with you. Exception: Review of Chap. 3 and 4 will be based on the text prepared by Rich Graham together with his reviewers this week, i.e. final reviews at the meeting will be done on MPI-2.1 Draft Apr. 26, 2008. Best regards from travel (currently Salishan conference, USA) Rolf --------------------------------------------------------------------------- Frontmatter: - Bill Gropp, Rusty Lusk Acknowledgements: - Richard Graham, Bill Gropp, Rusty Lusk Chap. 1: Introduction to MPI - Bill Gropp, Rusty Lusk, Karl Feind, Adam Moody, Jesper Traeff Chap. 2: MPI-2 Terms and Conventions - Tony Skjellum, Bill Gropp, Richard Barrett, Jesper Traeff Chap. 3: Point-to-Point Communication - Rich Graham, Jespar Larsson Traeff, George Bosilca, Steve Poole, Kannan Narasimhan, David Solt, B. Gropp, Matt Koop, Adam Moody Chap. 3: Datatypes - Rich Graham, Jespar Larsson Traeff, George Bosilca, Steve Poole, Kannan Narasimhan, David Solt, B. Gropp, Matt Koop, Adam Moody Chap. 5: Collective Communication - Adam Moody, Steven Ericsson-Zenith, Edgar Gabriel, Rajeev Thakur, Bill Gropp, Georg Bosilca, Torsten Hoefler, Jesper Traeff Chap. 6: Groups, Context, and Communicators - Richard Treumann, Steven Ericsson-Zenith, Edgar Gabriel, Tony Skjellum, Bill Gropp, Georg Bosilca, Robert Blackmore Chap. 7: Process Topologies - Jesper Traeff, Rusty Lusk, Bill Gropp, Richard Barrett Chap. 8: MPI Environmental Management (incl. sections from MPI-2 Misc.) - George Bosilca, Rich Graham, Jespar Larsson Traeff, Steve Poole, Kannan Narasimhan, David Solt, B. Gropp Chap. 9: Miscellany - Jespar Larsson Traeff, Rich Graham, George Bosilca, Steve Poole, Kannan Narasimhan, B. Gropp Chap.10: Process Creation and Management - David Solt, Dries Kimpe, Rusty Lusk, Georg Bosilca, Bill Gropp, Kalem Karian Chap.11: One-Sided Communication - Jespar Larsson Traeff, Ericsson-Zenith, Martin Schulz, Bill Gropp, Darius Buntinas Chap.12: External Interfaces - Bronis de Supinski, Bill Gropp Chap.13: I/O - Rajeev Thakur, Joachim Worringen, Bill Gropp, Rainer Keller Chap.14: Profiling Interface - Bronis de Supinski, Bill Gropp, Jeff Brown Chap.15: Deprecated Functions - Rolf Rabenseifner, Bill Gropp Chap.16: Language Bindings - Jeff Squyres, Steve Poole, Purushotham Bangalore, Bill Gropp, Erez Haba, Alexander Supalov Bibliography: - Rusty Lusk, Bill Gropp Annex A: Language Bindings Summary - Alexander Supalov, Jeff Squyres, Steve Poole, Purushotham Bangalore, Bill Gropp, Annex B: Change Log - Rolf Rabenseifner, Bill Gropp Index: - Rusty Lusk, Bill Gropp MPI-1.3: Bill Gropp, Rolf Rabenseifner, Adam Moody, Puri Bangalore, Terry Dontje 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

Formal MPI-2.1 Review
by unknown＠example.com 21 Apr '08

21 Apr '08

First, I want to thank all chapter authors for their fine work cleaning up all the MPI-1/-2 problems and achieving a single monolithic MPI-2.1 text. Dear reviewer, to handle about 20 chapters with in average 5 reviews per chapter, I would like to formalize it: On next Monday (Apr. 28, 2008), I would like to ask you whether you can make the following statement on the chapter(s) you voloteered for reviewing (I'll ask independently each reviewer on each chapter): --------------------------------------------------------------------------- Chap. ......... of MPI-2.1 Draft Apr. 19, 2008 represents fully and exactly (as I can overlook) the text that was given by MPI-1.1 (June 1995), MPI-2 (July 1997), MPI-1 errata (Oct. 12, 1998), Ballot 1+2 (May 15, 2002), Ballot 3 (protocol from March 2008 meeting, 1st vote), and Ballot 4 (protocol from March 2008 meeting, official reading) if the following text-modifications (they are already implemented in the current draft) are accepted by the MPI Forum: - XXXXX and if the following bugs will be corrected until first vote in June 2008 meeting: - YYYYY. Further comments (not official part of this review): - ZZZZZ --------------------------------------------------------------------------- Please do not use ZZZZZ for MPI-2.2 items or for any further enhancement. If there are such XXXXX, YYYYY, and ZZZZZ items, then I can include them into my slides, if you are sending it to me not later than Friday, April 25, 2008. All items that you detect later, you may bring to the MPI Forum with your own slide, plugging in your own laptop at the meeting. And same procedure for MPI-1.3 with only MPI-1 input. I hope this review process is fine with you. Exception: Review of Chap. 3 and 4 will be based on the text prepared by Rich Graham together with his reviewers this week, i.e. final reviews at the meeting will be done on MPI-2.1 Draft Apr. 26, 2008. Best regards from travel (currently Salishan conference, USA) Rolf --------------------------------------------------------------------------- Frontmatter: - Bill Gropp, Rusty Lusk Acknowledgements: - Richard Graham, Bill Gropp, Rusty Lusk Chap. 1: Introduction to MPI - Bill Gropp, Rusty Lusk, Karl Feind, Adam Moody, Jesper Traeff Chap. 2: MPI-2 Terms and Conventions - Tony Skjellum, Bill Gropp, Richard Barrett, Jesper Traeff Chap. 3: Point-to-Point Communication - Rich Graham, Jespar Larsson Traeff, George Bosilca, Steve Poole, Kannan Narasimhan, David Solt, B. Gropp, Matt Koop, Adam Moody Chap. 3: Datatypes - Rich Graham, Jespar Larsson Traeff, George Bosilca, Steve Poole, Kannan Narasimhan, David Solt, B. Gropp, Matt Koop, Adam Moody Chap. 5: Collective Communication - Adam Moody, Steven Ericsson-Zenith, Edgar Gabriel, Rajeev Thakur, Bill Gropp, Georg Bosilca, Torsten Hoefler, Jesper Traeff Chap. 6: Groups, Context, and Communicators - Richard Treumann, Steven Ericsson-Zenith, Edgar Gabriel, Tony Skjellum, Bill Gropp, Georg Bosilca, Robert Blackmore Chap. 7: Process Topologies - Jesper Traeff, Rusty Lusk, Bill Gropp, Richard Barrett Chap. 8: MPI Environmental Management (incl. sections from MPI-2 Misc.) - George Bosilca, Rich Graham, Jespar Larsson Traeff, Steve Poole, Kannan Narasimhan, David Solt, B. Gropp Chap. 9: Miscellany - Jespar Larsson Traeff, Rich Graham, George Bosilca, Steve Poole, Kannan Narasimhan, B. Gropp Chap.10: Process Creation and Management - David Solt, Dries Kimpe, Rusty Lusk, Georg Bosilca, Bill Gropp, Kalem Karian Chap.11: One-Sided Communication - Jespar Larsson Traeff, Ericsson-Zenith, Martin Schulz, Bill Gropp, Darius Buntinas Chap.12: External Interfaces - Bronis de Supinski, Bill Gropp Chap.13: I/O - Rajeev Thakur, Joachim Worringen, Bill Gropp, Rainer Keller Chap.14: Profiling Interface - Bronis de Supinski, Bill Gropp, Jeff Brown Chap.15: Deprecated Functions - Rolf Rabenseifner, Bill Gropp Chap.16: Language Bindings - Jeff Squyres, Steve Poole, Purushotham Bangalore, Bill Gropp, Erez Haba, Alexander Supalov Bibliography: - Rusty Lusk, Bill Gropp Annex A: Language Bindings Summary - Alexander Supalov, Jeff Squyres, Steve Poole, Purushotham Bangalore, Bill Gropp, Annex B: Change Log - Rolf Rabenseifner, Bill Gropp Index: - Rusty Lusk, Bill Gropp MPI-1.3: Bill Gropp, Rolf Rabenseifner, Adam Moody, Puri Bangalore, Terry Dontje 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