Dear Bill, George, Rolf, here my review about all ticket 0 up to svn r1291. All pages:lines relative to
Bill (Frontmatter) - ii:45 2010 --> 2012 - iii:3-9 you should list the real new things, as in iii:29-32 - xxx:8-26: Please add Reinhold Bader, Nick M. Maclaren, John Reid and Bill Long to the list. They activly supported the Fortran discusion. Bill (Introduction) - p4:37-39 Details missing, Should be similar to p3:13-14 - p7:32 If we proceed with removing something, then item abbout new chapter "Removed Functions" must be added. - p7:43 If we change the way of our indexes, then this text must be changed. - p8: We should mention that the JOD was part of MPI-2.0 because several chapters in the JOD may be now obsolety, e.g., about Fortran p8:22-23 Torsten (chap-coll) + Rolf (chap-changes) - p198:5 The C-Binding of MPI_OP_FREE was changed from incorrect MPI_op_free to correct MPI_Op_free. This may be a very old MPI-1.1 errata, but may be at least mentioned as a ticket-0 change. - p861:22 may be mentioned as a late MPI-1.1 errata Bill (Groups chpa-context) - p262:40, sorry my indent was wrong. George (Environm chap-iquiry) - p351:43 please make also the correction 240-L as on p351:33, i.e., "::" - p352:17-18 "before MPI_INIT and after MPI_FINALIZE." should be also part of the removal. I personally would prefer to have this sentences at each routine, instead of the summarizing one sentence on p353:2.3 about two routines. - p433:11 (MPI_Win_attach) reads TYPE(*), ASYNCHRONOUS :: base but should read TYPE(*), DIMENSION(..), ASYNCHRONOUS :: base - p434:16 (MPI_Win_detach) reads TYPE(*), ASYNCHRONOUS :: base but should read TYPE(*), DIMENSION(..), ASYNCHRONOUS :: base Rolf (ADD-F2008.sed and all accompanying files) - MPI_Win_attach with "DIMENSION(..)" for "base": TYPE(*), DIMENSION(..), ASYNCHRONOUS :: base - MPI_Win_detach with "DIMENSION(..)" for "base": TYPE(*), DIMENSION(..), ASYNCHRONOUS :: base - MPI_Get-address with "ASYNCHRONOUS" for "location": TYPE(*), DIMENSION(..), ASYNCHRONOUS :: location Rolf/Jeff (Binding) - p635:45 is correct, but the \const should not contain the correction macro: therefore \constskip{.....with correction...}\cdeclindex{MPI::LONG\_LONG} (done in r1304) Best regards Rolf ----- Original Message -----
From: "Rolf Rabenseifner" <[email protected]> To: "Main MPI Forum mailing list" <[email protected]> Sent: Monday, July 16, 2012 5:00:10 PM Subject: [Mpi-forum] New pdf version from Torsten Torsten has done a new pdf:
http://www.unixer.de/sec/mpi-report-r1300.pdf
Thanks to Torsten
Best regards Rolf
-- Dr. Rolf Rabenseifner . . . . . . . . . .. email [email protected] 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-forum mailing list [email protected] http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum
-- Dr. Rolf Rabenseifner . . . . . . . . . .. email [email protected] 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)