1
1
Граф коммитов

24 Коммитов

Автор SHA1 Сообщение Дата
Rainer Keller
64aa384745 - Fix comment to reflect current code.
This commit was SVN r20479.
2009-02-09 00:52:57 +00:00
George Bosilca
633ee3c2ce Small optimizations in order to force the compiler to inline some critical functions.
This commit was SVN r14317.
2007-04-12 04:29:43 +00:00
Brian Barrett
75bfb1f0be Make BEGIN_C_DECLS and END_C_DECLS publically available so that they can see
wider use

This commit was SVN r13869.
2007-03-01 17:25:21 +00:00
Brian Barrett
c90ad01220 set object information in the internal macro, as that's what is used by
the free list grow functions

This commit was SVN r13402.
2007-01-31 21:51:06 +00:00
George Bosilca
9dead6382b What we need is to have unsigned long long constants.
This commit was SVN r13378.
2007-01-31 03:00:33 +00:00
George Bosilca
38fd8ffb54 The intended patch.
This commit was SVN r13342.
2007-01-28 04:55:05 +00:00
Jeff Squyres
c522f6f997 Tim pointed out that the original value was probably not what was
intended.  In truth, any unique value will do, but he's right that
having a regular, recognizeable, repeatable value is probably much
better for debugging purposes.

This commit was SVN r13341.
2007-01-28 00:46:25 +00:00
Jeff Squyres
3c5c8c3c4c Refinement of Rainer's r13227 and r13228 (worked with Rainer, Ralph,
and George on these refinements):

 * Rename the static OBJ initializer macro to be
   OPAL_OBJ_STATIC_INIT(class)
 * Ensure that all static OBJ initializations get a refcount of 1
   (doesn't ''really'' matter, since they're static, it should never
   get to the point where the OBJ is DESTRUCTed, but more correct
   nonetheless)
 * Add a "magic number" to the OBJ when compiling with debug support.
   The magic number does some rudimentary support to ensure that
   you're operating on a valid OBJ (and fails an assertion if you're
   not).  Check to ensure that the memory contains the magic number
   when performing actions of OBJ's.  Also remove the magic number
   when DESTRUCTing OBJs, so that if, for example, an OBJ is
   DESTRUCTed more than once, we'll fail the magic number assert.

This commit was SVN r13338.

The following SVN revision numbers were found above:
  r13227 --> open-mpi/ompi@96030de97b
  r13228 --> open-mpi/ompi@c2e9075d29
2007-01-27 13:44:03 +00:00
George Bosilca
c0b22678a7 OMPI_ENABLE_DEBUG is always defined. Remove all warnings when compiled in
non debug mode.

This commit was SVN r13230.
2007-01-21 18:14:06 +00:00
Rainer Keller
c2e9075d29 - Define a OPAL_CLASS_EMPTY to be used for initialization.
Similar within the dt_module for the predefined datatypes.

This commit was SVN r13228.
2007-01-21 15:52:06 +00:00
George Bosilca
3ecdb1678e Rollback the opal_object commit. Except on my initial test environment it
do not produce anything usefull.

This commit was SVN r12487.
2006-11-08 07:10:16 +00:00
George Bosilca
ad5d5b95c8 Help the compiler to understand what we're doing here. It generate less
memory accesses at least with gcc.

This commit was SVN r12483.
2006-11-08 05:08:11 +00:00
Rainer Keller
528329c30f - Have two separate versions of OBJ_RETAIN and OBJ_RELEASE to
get readable preprocessed output.

This commit was SVN r12162.
2006-10-18 08:24:38 +00:00
George Bosilca
5e280cda19 Latest and greatest. Now OPAL is ready for the Windows prime-time.
The same treatement will happens on all sub-projects. The .h files
have to be C++ compatibles and all symbols with an external visibility
have to get the {PROJECT}_DECLSPEC in front of the prototype.

This commit was SVN r11340.
2006-08-23 00:29:35 +00:00
George Bosilca
6afa4c6c64 Windows friendly version. We have to split the OMPI_DECLSPEC in at least 3
different macros, one for each project. Therefore, now we have OPAL_DECLSPEC,
ORTE_DECLSPEC and OMPI_DECLSPEC. Please use them based on the sub-project.

This commit was SVN r11270.
2006-08-20 15:54:04 +00:00
Galen Shipman
5271948ec0 --- opal object changes
add object size to opal class
no longer need the size when allocating a new object as this is stored in
the class structure

--- dr changes 
Previous rev. maintained state on the communicator used for acking duplicate
fragments, but the communicator may be destroyed prior to successfull
delivery of an ack to the peer. We must therefore maintain this state
globally on a per peer, not a per peer, per communicator basis. 
This requires that we use a global rank on the wire and translate this as
appropriate to a local rank within the communicator. 

This commit was SVN r9454.
2006-03-29 16:19:17 +00:00
George Bosilca
c8842181d1 Correct the obj_update for the threaded mode.
This commit was SVN r8988.
2006-02-12 05:36:00 +00:00
Brian Barrett
566a050c23 Next step in the project split, mainly source code re-arranging
- move files out of toplevel include/ and etc/, moving it into the
    sub-projects
  - rather than including config headers with <project>/include, 
    have them as <project>
  - require all headers to be included with a project prefix, with
    the exception of the config headers ({opal,orte,ompi}_config.h
    mpi.h, and mpif.h)

This commit was SVN r8985.
2006-02-12 01:33:29 +00:00
George Bosilca
d56656580a Profiling shows that we spend way to much time (it's relative of course but it's still more
than I want to) on the constructor and destructor loops. Compress the constructors and
destructors array to remove all the useless tests against NULL. Now we use the cache a
little bit more efficient.

Correct a threading problem in the obj_update function. In the threaded case we have to
return the value returned by the ompi_atomic_add function otherwise there are ways
leading to the release of an object twice.

This commit was SVN r8966.
2006-02-10 21:04:26 +00:00
George Bosilca
bd0ee62e62 Protect headers and use __WINDOWS__ for Windows code.
This commit was SVN r8468.
2005-12-12 22:01:51 +00:00
Jeff Squyres
42ec26e640 Update the copyright notices for IU and UTK.
This commit was SVN r7999.
2005-11-05 19:57:48 +00:00
George Bosilca
3f466e2915 Small change with large benefits. Until now, in DEBUG mode we keep trace of the file and line number where
an object has been created. I slightly change this behaviour. Now, during the life time of an object (when
the reference count is > 0) we keep the place where the object was created and just before freeing the object
we set the location where the object get destroyed. So when we try to release an already destroyed object
we have the location where it happens (where the object get destroyed for the first time).

This commit was SVN r6428.
2005-07-12 05:26:37 +00:00
Brian Barrett
9da0b4fe1d * rename all the atomic functions from ompi to opal
This commit was SVN r6325.
2005-07-03 21:38:51 +00:00
Brian Barrett
499e4de1e7 * rename ompi_object and ompi_class to opal_object and opal_class
This commit was SVN r6321.
2005-07-03 16:06:07 +00:00