From jsquyres at [hidden] Thu May 1 14:15:50 2008 From: jsquyres at [hidden] (Jeff Squyres) Date: Thu, 1 May 2008 15:15:50 -0400 Subject: [Mpi-21] MPI 2.1 doc changes Message-ID: <874CD436-AA53-4C77-9A59-7D51BA94B772@cisco.com> Rolf - Per the April meeting, I have made the following changes in mpi-forum- docs/MPI-2.1: 1. You likely have these noted somewhere already -- this is just to confirm to you that I performed these changes in the latex already (sections / line numbers based on the April 19 doc): - 5.9.5: p168.21: add missing C++ binding for User_function - 16.1.1: p449.25-26: delete sentence "See Annex ... for a function cross reference" because we deleted that table from the annex. --> I did not remove this table from Annex A; I think you were going to do it 2. Made the changes with regards to the beginning of Annex A: - Added the missing MPI-2 types - Added missing C++ types (all) - Updated to use the MPI-2 names and typedefs for attribute and errhandler callbacks - Added the missing C++ typedefs for attribute and errhandler callbacks - Split list of handle/opaque types and attribute/errhandler callback typedefs into two subsections - Created new subsection for the deprecated MPI-1 attribute and errhandler typedef names 3. Added many examples to the Examples Index. I'm sure we can always add more here. 4. Added all C++ types to the Declaration Index. 5. Added missing types from the Callback Function Prototype Index --> Do we want the deprecated typedefs listed here? Currently, they *are*. 6. Committed a change from Rainer to the MAKE-APPLANG script that enables building the doc properly on both Linux and OS X. 7. Committed new text introductions to many of the new indexes. 8. Fixed most (all?) of the non-functions listed in the MPI function index. There's at least one left that I'm aware of -- MPI_THREAD_INIT -- but I know that that item is already on your list to fix, so I left it for you. 9. I checked: all the pre-defined attribute callback functions (both old and new) show up properly in the main function index. I then stole a LaTeX trick from the bindings chapter to have functions show up in the C/Fortran indexes, but not have them show up in the chapter itself. That means that I didn't have to modify the index-generating scripts at all to make the 3 functions show up in the C/Fortran indexes properly (MPI_NULL_COPY_FN, MPI_DUP_FN, MPI_NULL_DELETE_FN). Woo hoo! -------------------------------------------------------- You can see a summary of all my changes this way: svn log -v -r119:132 To see the details of exactly what I changed in any given SVN commit, use this: svn diff -r N-1:N where N is an individual r number. For example, to see exactly what I did in r124: svn diff -r 123:124 With this mail, I am giving the SVN write token back to you. Please let me know if you have any questions, comments, etc. If I did something that you think is "bad", please let me know and I'll back it out. -- Jeff Squyres Cisco Systems From rabenseifner at [hidden] Sat May 3 11:08:25 2008 From: rabenseifner at [hidden] (Rolf Rabenseifner) Date: Sat, 03 May 2008 18:08:25 +0200 Subject: [Mpi-21] Official reading for MPI-1.3 and MPI-2.1 is finished Message-ID: At the meeting April 28-30, 2008, the MPI Forum has finished the review process for MPI-1.3 and MPI-2.1. The official reading is hereby finished. The final bug-fixes (discussed and voted at the meeting) will be done in the next weeks. All protocol files can be found at http://www.hlrs.de/mpi/mpi21/doc/ (username and pw is mpi21) and soon at the official protocol page at http://meetings.mpi-forum.org/secretary/2008/04/ All further enhancements will be part of MPI-2.2 (except obvious typos detected before final document assembly) I want to thank all who contributed to the ballots and the combined documents. 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)