Re: [mpi-21] Provide one MPI-1.2 / MPI-1.3 standard document

["Rolf Rabenseifner" <rabenseifner@xxxxxxx>]

MPI 1.3 will be handled by Rainer.
Rainer will give the information where to find the dcuments
forreview when they are prepared for review.

MPI 2.1 is not yet ready for review.

Best regard
Rolf


On Fri, 8 Feb 2008 14:27:02 -0800 (PST)
 "Bronis R. de Supinski" <bronis@xxxxxxxx> wrote:
>
>Rolf:
>
>This reminds me that I signed up to review chapters from both MPI 1.2
>(I guess 1.3 now) and from MPI 2.0. I sort of remember which but not
>precisely. Are those assignments posted somewhere?
>
>Also, which version(s) are we supposed to review exactly? Are they
>available yet? I know you wanted us to wait until at least 1.3 was
>available but I seem to recall that we are waiting for a first draft
>of MPI 2.1. Is that correct?
>
>Thanks,
>
>Bronis
>
>
>
>On Fri, 8 Feb 2008, Rainer Keller wrote:
>
>> Hello dear all,
>> the MPI-1 and MPI-2 sources have been checked in to the SVN-Repository.
>>
>> Into the MPI-1 document on the SVN, I have integrated the changes of chapter 3
>> of "MPI-2: Extensions to the Message-Passing Interface", from July 18, 1997
>> and checked the changes of the "Errata for MPI-1.1", from October 12, 1998.
>>
>> Indeed of the 55 listed changes to MPI-1.1 of the latter document, *eight*
>> were not yet in chapter 3 of MPI-2.0:
>>    1. pI, l7: Change date of Version 1.0 -- from p1, l21
>>    2. pI, l16: Change date as well of when draft was processed -- from p1, l27
>>    16. p77, l41-42: Get rid of superfluous argument (1) to MPI_TYPE_HVECTOR --
>> from p4, l8-19
>>    28. p153, l39: Replace comm_a by comm_b in check -- from p6, l30
>>    29. p172, l26: Fix name when calling MPI_Keyval_create -- from p6, l40
>>    31. p182, l18: Change output description of status in MPI_TOPO_TEST from
>> (choice) to (state) -- from p6, l6-19
>>    32. p184, l31-32: Change output description of coords in MPI_CART_COORDS
>> from (integer) to (array of integers) -- from p6, l20-36
>>    37. p200, l27-29: Makes an attempt to better the explanation for MPI_Abort
>> -- from p8, l26-35
>>
>> Additionally, I have changed the build process so that one may choose between
>> latex/pdflatex for Postscript / PDF.
>> The PDF now features hyperlinks to the actual functions on the *correct* page,
>> as well as the function index, references and PDF-Metainformation. The
>> results may be seen below.
>>
>> It does contain the ballot 3 items (in the latex marked with
>> %MPI-1.2-Ballot3).
>> May I ask the MPI-2.1 reviewers to please send comments / suggestions /
>> further clarifications?
>>
>> Until February 15th, all the errata regarding the MPI-1.1/MPI-1.2 standard
>> should be send and provisionally included to be read & voted upon, until
>> March 15th, all of the MPI-1.1 errata parts should be integrated.
>> I then will send the final MPI-1.3 draft document with and without \CHANGE,
>> \INTO/,\ADD,\DELETE macros.
>>
>> With best regards,
>> Rainer
>>
>>
>> PS: The current changes-file is attached below as well.
>> --
>> ----------------------------------------------------------------
>> 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@xxxxxxx
>>  Germany                             AIM/Skype:rusraink
>>



Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner@xxxxxxx
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)