1
1
Граф коммитов

4510 Коммитов

Автор SHA1 Сообщение Дата
George Bosilca
4925edb374 Do not believe the upper level about the total size to unpack. Check to insure we
respect the boundaries of the data as see by the user.

This commit was SVN r5853.
2005-05-24 19:54:46 +00:00
George Bosilca
15c3baaf3b Always set the bConverted to zero when we attach a new datatype.
This commit was SVN r5851.
2005-05-24 19:30:25 +00:00
George Bosilca
99640f35bc Composed predefined datatypes should be marked as DATA.
This commit was SVN r5849.
2005-05-24 17:50:08 +00:00
George Bosilca
29bd233d8d Correctly exit the loop. Imported from the beta branch.
This commit was SVN r5848.
2005-05-24 17:48:01 +00:00
Jeff Squyres
a96746f041 Remove the unignore file for now
This commit was SVN r5846.
2005-05-24 17:41:01 +00:00
Ralph Castain
a0e96e92ca Ensure that daemon debug flags get into the environment and propagated.
This commit was SVN r5844.
2005-05-24 15:02:50 +00:00
Ralph Castain
2451f3bdc9 Several fixes in this commit:
1. Fixed the GPR search engine so that keys AND worked, and so that multiple objects with the same key didn't mess up the search.

2. Added an orte_bitmap function based on the existing ompi_bitmap one, but minus the fortran "pollution"

3. Added a new name service function called create_my_name to remove the duplicate name creation that was happening with the RML. Basically, the RML has to assign a name when a process makes first contact if the process doesn't already have a name. For processes that get a name passed into them, this was okay - the name was already assigned. For other processes (e.g., singletons), this was not okay - the first message to the seed daemon was to create a name, which caused the RML to assign one, and then the name service to assign another.

4. Change orted so it gets its name the way everyone else does - during orte_init.

This commit was SVN r5842.
2005-05-24 13:39:15 +00:00
George Bosilca
7974c5b6b8 When the convertor is initialized a starting position set to -1 indicate
that the internal stack should not be constructed.

This commit was SVN r5840.
2005-05-23 23:13:04 +00:00
George Bosilca
487c4b9c6e As this convertor is never used to pack/unpack data we dont need to create the internal
stack. Using -1 as starting position specify that the internal stack should not be created.

This commit was SVN r5839.
2005-05-23 23:09:35 +00:00
George Bosilca
3278545a34 Remove useless static variable.
This commit was SVN r5838.
2005-05-23 23:08:28 +00:00
George Bosilca
4a7211468b Adapt to the last commits from Tim.
This commit was SVN r5837.
2005-05-23 23:05:15 +00:00
Ralph Castain
2739c0b0c5 Ensure that the name in the environment is cleared so that we pick up the name as components later.
This commit was SVN r5836.
2005-05-23 22:55:21 +00:00
Tim Woodall
649eeed4a7 rename
This commit was SVN r5835.
2005-05-23 22:22:20 +00:00
Tim Woodall
37b43eaccf small checkin :-)
- remove any ptl dependancy from mpi/pml interface
- beginings of a new pml
- new Bit Mover Interface (BMI) framework

This commit was SVN r5834.
2005-05-23 22:06:50 +00:00
Jeff Squyres
905bf85295 Be more careful with datatype names.
This commit was SVN r5833.
2005-05-23 21:45:20 +00:00
Jeff Squyres
913c0d2f87 Remove unused (and deprecated) 3rd argument to main
This commit was SVN r5830.
2005-05-23 21:02:39 +00:00
George Bosilca
8caf321317 All predefined types must have the DT_FLAG_BASIC set !
This commit was SVN r5827.
2005-05-23 19:43:44 +00:00
George Bosilca
878bf0f36d Move the AC_DEFUN outside the other one.
This commit was SVN r5822.
2005-05-23 16:43:37 +00:00
George Bosilca
fea0010aee For dist all AM_CONDITIONAL's have to be defined.
This commit was SVN r5821.
2005-05-23 16:26:55 +00:00
Tim Woodall
0138323c6e check for NULL process name and assign if allocated by server
This commit was SVN r5820.
2005-05-23 16:23:11 +00:00
George Bosilca
6740270ace This directory was supposed to be private ... it does not have to be published on the SVN.
This commit was SVN r5819.
2005-05-23 16:10:33 +00:00
Jeff Squyres
ca64115b14 Add logic to automatically determine if to use $ibdir/lib or
$ibdir/lib64.  May still need final testing / verification from Tim. 

This commit was SVN r5816.
2005-05-23 15:09:51 +00:00
Ralph Castain
4a83f4da02 Replace a missing init that was required.
Note that singleton operations are still down - we are investigating.

This commit was SVN r5815.
2005-05-23 14:50:52 +00:00
Ralph Castain
93eb0d4b40 Checkpoint
This commit was SVN r5814.
2005-05-23 14:22:35 +00:00
Jeff Squyres
131d49152f Need to prefix the svn version with "r" so that dist tarballs are made
correctly

This commit was SVN r5813.
2005-05-23 12:36:09 +00:00
Rainer Keller
0e211d5cfe Small buglet found by nightly built.
This commit was SVN r5812.
2005-05-23 11:25:24 +00:00
Ralph Castain
689a290711 Add one further degree of separation between opal and orte - allow separate init of the two systems. This allows the restart capability to avoid hitting opal utilities (e.g., mca_base_open, ompi_output_init) repeatedly.
Clean up the ignores as well.

This commit was SVN r5811.
2005-05-22 18:40:03 +00:00
Josh Hursey
283bbb12f0 checkpoint for Windows src/tools build
This commit was SVN r5810.
2005-05-22 18:03:27 +00:00
Josh Hursey
976006805f forgot one. need to make sure this is OMPI_DECLSPECed for Windows
This commit was SVN r5809.
2005-05-22 17:01:26 +00:00
Josh Hursey
4f215294d4 A bit of cleanup to build under windows.
Change a couple function definitions in datatype.h from extern
to OMPI_DECLSPEC since they are never defined anywhere else, and 
need to be for the build.

Added the datatype.h file to ptl_self so it can see the inline
convertor_pack routines in that file when not in Debug mode.

This commit was SVN r5808.
2005-05-22 17:00:25 +00:00
Josh Hursey
b1f1786d51 fixed a define around a unistd include
This commit was SVN r5807.
2005-05-22 15:46:40 +00:00
Ralph Castain
7b6db8a18f Can now start/finalize/restart the run-time without crashing.
Add a unit test for that functionality - will test more fully next week.

This commit was SVN r5806.
2005-05-22 03:11:33 +00:00
Jeff Squyres
a419d9be72 Use an explicit TMPDIR so that we don't get stale universe directories
This commit was SVN r5805.
2005-05-21 11:44:16 +00:00
Thara Angskun
c19562ec8d Clean up code
This commit was SVN r5804.
2005-05-21 06:35:45 +00:00
Jeff Squyres
a6dbf68357 Ensure to pass NULL as the 4th parameter so that we don't end up with
3 IO parameters with blank names.

This commit was SVN r5801.
2005-05-20 20:56:54 +00:00
Jeff Squyres
1bdacc0c54 Add in display for all the INTEGER*n, REAL*n, and COMPLEX*n sizes and
alignments.

This commit was SVN r5800.
2005-05-20 20:56:15 +00:00
Tim Woodall
031f687e5a - global debug parameter was renamed to orte_debug
- pass an environment through to app/ssh session

This commit was SVN r5799.
2005-05-20 19:33:22 +00:00
George Bosilca
56d4c15047 If somehow we fail to open a connection print the IP address and the port used.
Submitted: George Bosilca
Reviewed: Tim Woodall

This commit was SVN r5798.
2005-05-20 17:47:29 +00:00
Ralph Castain
cabcd783df Put the memset in the correct place....
This commit was SVN r5797.
2005-05-20 16:26:42 +00:00
Ralph Castain
7a28a1a9a8 Remove a totally useless field that somehow got into the universe info structure.
Re-sync the orte and ompi constants (latter was missing a few new ones).

This commit was SVN r5796.
2005-05-20 15:50:27 +00:00
Jeff Squyres
d7e9fd635c This license has tenatively been agreed to by all parties. The
agreements are not yet final -- I'll probably have a no-change commit
with a log message when the agreements are final -- but I'm tired of
having to go lookup the license text in my mail archives.  :-)

Note that this license is vanilla BSD except for the 2nd-to-last
paragraph.

This commit was SVN r5794.
2005-05-20 10:42:59 +00:00
Thara Angskun
42d6434337 - some cleanup
This commit was SVN r5793.
2005-05-20 09:18:28 +00:00
Jeff Squyres
bb3009e6f3 Re-add "datatype/datatype.h" because it disappeared from some
upper-level .h files.

This commit was SVN r5792.
2005-05-20 00:40:05 +00:00
Jeff Squyres
eebda3b036 More ompi_datatype_t -> struct ompi_datatype_t changes
This commit was SVN r5791.
2005-05-20 00:29:45 +00:00
Jeff Squyres
0dd11c8c32 Add <unistd.h> for getcwd() on Linux.
This commit was SVN r5790.
2005-05-20 00:16:48 +00:00
George Bosilca
2f85bc409b Hmmm there is no way to don't have to compile half of the files when I touch to the datatype.h file ...
This commit was SVN r5789.
2005-05-19 23:57:56 +00:00
Jeff Squyres
f28095e632 Bunches of fixes for Fortran support:
- Fully support REAL*N, INTEGER*N, and COMPLEX*N in the MPI_Op
  reduction operations.
- Update ddt to fully support these types as well, to include using
  the results of sizes and alignments determined by configure
- Discover the goodness of m4 and consolidate a LOT of configure code
  (i.e., remove a lot of essentially duplicated code and
  m4-subroutine-ize it).  The big kicker was figuring out how to
  parameterize AC_DEFINE_UNQUOTED, which you can do if you use m4
  properly.
- If we don't support a given INTEGER*N, REAL*N, or COMPLEX*N, don't
  error.  Just set the right flags so that we don't support them in
  the MPI layer.

This commit was SVN r5788.
2005-05-19 23:56:02 +00:00
Jeff Squyres
08082be721 Silly mistake (how on earth did it live this long?): adding -DNDEBUG
to CXXFLAGS should use $CXXFLAGS, not $CFLAGS.

This commit was SVN r5787.
2005-05-19 23:52:13 +00:00
George Bosilca
84daefbfc0 We need a prototype for one of the functions in datatype.h
This commit was SVN r5785.
2005-05-19 23:28:07 +00:00
George Bosilca
a47d0c106e Still tracking useless dependencies ... (here comm does not need to know what's in the ompi_op_t structure).
This commit was SVN r5784.
2005-05-19 23:16:06 +00:00