1
1
Граф коммитов

644 Коммитов

Автор SHA1 Сообщение Дата
George Bosilca
da4d993176 dd some Windows missing defines & macros.
This commit was SVN r14184.
2007-04-01 15:55:34 +00:00
Tim Prins
ce6b44d476 Fix a couple of minor memory leaks
This commit was SVN r14166.
2007-03-29 01:02:26 +00:00
Tim Prins
bc1b046e33 Change a bunch of 'OMPI' names to OPAL names, since this is opal.
This commit was SVN r14165.
2007-03-29 01:00:33 +00:00
George Bosilca
852d6b0b68 Including windows.h after opal_config.h was included break the precompiled
header feature of Visual C++.

This commit was SVN r14157.
2007-03-28 14:20:23 +00:00
Shiqing Fan
91cfb2f149 A few mismatched declearations are fixed, and several header files are added for Cygwin...
This commit was SVN r14151.
2007-03-27 14:17:25 +00:00
Sven Stork
548c511700 - export required symbol
This commit was SVN r14140.
2007-03-26 13:54:20 +00:00
Rainer Keller
1322f9f346 - Further attributes mainly for opal/* functions, marking
__opal_attribute_nonnull__, __opal_attribute_warn_unused_result__,
   __opal_attribute_malloc__, __opal_attribute_sentinel__ and
   __opal_attribute_format__

This commit was SVN r14078.
2007-03-20 13:01:32 +00:00
Sven Stork
d67565b042 - use include path relative to opal/include or this header file will not work when installed "--with-devel-headers"
This commit was SVN r14077.
2007-03-20 12:38:06 +00:00
Rainer Keller
249abd29c2 - Mark some deprecated functions (two still commented) and fix to
not use opal_cmd_line_make_opt anymore.

This commit was SVN r14072.
2007-03-20 10:08:58 +00:00
Josh Hursey
7ab741c1e2 - Add some debugging hooks for the CR runtime MCA params
- Add signal handler BLCR register (helps with debugging)
- ifdef out the cr_request_file section for checkpointing self.
  There is a bug with the 0.4.2 version of BLCR such that this
  does not handle moving checkpoint files around.
  I'm following up with the BLCR folks on this one (and checking
  the newest release).

This commit was SVN r14069.
2007-03-19 21:18:03 +00:00
Josh Hursey
a181c987cc Remove some old references to ft_enable parameter that no longer exists.
This was replaced by the "-am ft-enable-cr" AMCA parameter.

This commit was SVN r14055.
2007-03-17 20:02:42 +00:00
Josh Hursey
d03073e87d Make sure to protect the finalize call so tools like ompi_info
do not segv.

This commit was SVN r14054.
2007-03-17 19:47:54 +00:00
Josh Hursey
dadca7da88 Merging in the jjhursey-ft-cr-stable branch (r13912 : HEAD).
This merge adds Checkpoint/Restart support to Open MPI. The initial
frameworks and components support a LAM/MPI-like implementation.

This commit follows the risk assessment presented to the Open MPI core
development group on Feb. 22, 2007.

This commit closes trac:158

More details to follow.

This commit was SVN r14051.

The following SVN revisions from the original message are invalid or
inconsistent and therefore were not cross-referenced:
  r13912

The following Trac tickets were found above:
  Ticket 158 --> https://svn.open-mpi.org/trac/ompi/ticket/158
2007-03-16 23:11:45 +00:00
Brian Barrett
01d6121c7f * The MoreBacktrace code supplied by Apple doesn't work on 64 bit Intel
builds, so disable it there
  * On 10.4.8 (and possibly others), siginfo is NULL in the signal
    callback on 64 bit Intel builds, so account for that in the signal
    callback.

This commit was SVN r14045.
2007-03-16 21:27:19 +00:00
Brian Barrett
4e72a8b652 Need to use the right state variable when debugging 64 bit applicatiosn
This commit was SVN r14015.
2007-03-12 18:57:58 +00:00
Tim Prins
fe3ea0085f Fix minor memory leaks
This commit was SVN r13946.
2007-03-07 01:09:38 +00:00
George Bosilca
21f014f5f6 Apply a patch from Bert Wesarg.
This saves some memory for the constructors and destructors arrays of a
class by counting the constructors and destructors while we are counting
the cls_depth. And the reversion of the constructor array can now be done
without an extra loop.

This commit was SVN r13939.
2007-03-06 16:38:06 +00:00
Gleb Natapov
90fb58de4f When frags are allocated from mpool by free_list the frag structure is also
allocated from mpool memory (which is registered memory for RDMA transports)
This is not a problem for a small jobs, but for a big number of ranks an
amount of waisted memory is big.

This commit was SVN r13921.
2007-03-05 14:17:50 +00:00
George Bosilca
36e0e6f074 va_args don't like char. Anyway, all chars get promoted to int
when they go through the ...

This commit was SVN r13878.
2007-03-01 23:26:20 +00:00
George Bosilca
04d84aefb4 Force the Windows registry to be set with a higher
priority than the default value.

This commit was SVN r13877.
2007-03-01 23:25:14 +00:00
George Bosilca
caa1522a22 Don't duplicate the component name for the MCA parameters on Windows when
we read the keys from the registry.

This commit was SVN r13875.
2007-03-01 21:50:27 +00:00
George Bosilca
be24d5d683 Allow guess_strlen to understand %c, as it is now required.
This commit was SVN r13874.
2007-03-01 21:48:31 +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
Josh Hursey
0404444dbe * Added 2 new MCA parameters
- mca_base_param_file_prefix
     (Default: NULL)
     This is the fullname of the "-am" mpirun option. Used to specify a ':'
     separated list of AMCA parameter set files.
  - mca_base_param_file_path
     (Default: $SYSCONFDIR/amca-param-sets/:$CWD)
     The path to search for AMCA files with relative paths. A warning will be
     printed if the AMCA file cannot be found.

* Added a new function "mca_base_param_recache_files" the re-reads the file
configurations. This is used internally to help bootstrap the MCA system.

* Added a new orterun/mpirun command line option '-am' that aliases for the
mca_base_param_file_prefix MCA parameter

* Exposed the opal_path_access function as it is generally useful in other
places in the code.

* New function "opal_cmd_line_make_opt_mca" which will allow you to append a
new command line option with MCA parameter identifiers to set at the same
time. Previously this could only be done at command line declaration time.

* Added a new directory under the $pkgdatadir named "amca-param-sets" where all
the 'shipped with' Open MPI AMCA parameter sets are placed. This is the first
place to search for AMCA sets with relative paths.

* An example.conf AMCA parameter set file is located in
contrib/amca-param-sets/.

* Jeff Squyres contributed an OpenIB AMCA set for benchmarking.

Note: You will need to autogen with this commit as it adds a configure param.
  Sorry :(

This commit was SVN r13867.
2007-03-01 13:39:20 +00:00
Sven Stork
ddff458f3f - this symbol is required by components
This commit was SVN r13850.
2007-02-28 16:38:59 +00:00
Rainer Keller
0889ebd59f - Eliminate warnings, that PGI-6.2.5 issues with -Minform=inform
This commit was SVN r13840.
2007-02-28 08:36:34 +00:00
Sven Stork
d8a369936e - Fix more symbols that should be exported.
This commit was SVN r13824.
2007-02-27 15:17:17 +00:00
Jeff Squyres
f02f8297a7 Minor fix for verbose message when a component is rejected and closed.
This commit was SVN r13807.
2007-02-27 02:07:33 +00:00
Ralph Castain
edcf25e4bb Use OPAL version numbers
This commit was SVN r13728.
2007-02-21 14:32:41 +00:00
Ralph Castain
8022081a85 Cleanup a compiler warning
This commit was SVN r13726.
2007-02-21 14:26:30 +00:00
George Bosilca
386fe713ef Define the windows registry static function only on Windows.
This commit was SVN r13691.
2007-02-16 21:48:02 +00:00
George Bosilca
924224e66d On Windows allow the MCA params to come from the Windows Registry. The keys
are readed recursively, and their name built-up based on the definition
tree.

This commit was SVN r13688.
2007-02-16 18:39:40 +00:00
George Bosilca
77a8a40d6a This function is supposed to be static.
This commit was SVN r13685.
2007-02-16 18:17:50 +00:00
Brian Barrett
4890636581 Work around what appears to be an odd error in the pathscale compiler where
the dispatch function opal_atomic_cmpset_acq_xx was used to call
opal_atomic_cmpset_acq_32  would cause the compiler / linker to fix up a
jump address to be itself, leading to an infinite loop.

We're still looking into exactly what caused this, but during the
investigation into the hang, we determined that the compiler (both pathcc
and gcc) weren't always inlining both the call to opal_atomic_cmpset_acq_xx
and opal_atomic_cmpset_acq_32, meaning there was a function call in
opal_atomic_lock.  The atomic lock will always be 32 bit, so there's no
need for the dispatch function, so might as well remove the dispatch
function that may or may not be inlined.

A bug fix that leads to potentially better performance.  Gotta love the
few times that happens...

This commit was SVN r13651.
2007-02-14 16:57:39 +00:00
Brian Barrett
2a16c094f7 A slightly more correct fix for issues with reordering unlock calls. We want
to flush all writes pending (ie, the data being protected) out of the memory
manager before we write the spinlock unlock. Only need a wmb instead of
full mb, which is at least slightly less intrusive.  Also, after much
thought, no need for a memory barrier in init.

This commit was SVN r13649.
2007-02-14 16:37:31 +00:00
Galen Shipman
7365aeaae2 use memory barriers for lock init and unlock
This commit was SVN r13644.
2007-02-13 21:20:59 +00:00
Brian Barrett
5a3293c0ad The extra includes should prefix the includedir, not the compile-time INCLUDEDIR when creating the -I arguments.
This commit was SVN r13623.
2007-02-12 22:43:50 +00:00
Brian Barrett
68cc044162 Clean up some support issues for platforms that don't have hton / ntoh
This commit was SVN r13615.
2007-02-12 16:29:59 +00:00
George Bosilca
de1cbab1fd Don't complain on 64 bits architectures.
This commit was SVN r13574.
2007-02-09 03:17:02 +00:00
Rainer Keller
38696e8d32 - Check for attributes, after C and C++ have been set up.
- Set ompi-specific autoconf cache-variables
 - Implement one function to check for availability of an
   attribute with the possibility for a cross-check.
 - Do cross-checks for
     __attribute__(format)
     __attribute__(nonnull)
     __attribute__(sentinel)
     __attribute__(warn_unused_result)
 - Grep the compilers warnings for keywords regarding ignored
   attributes.
 - Include also the no_instrument_function

This commit was SVN r13556.
2007-02-08 13:34:44 +00:00
George Bosilca
4439e44179 Be even more Windows friendly.
This commit was SVN r13549.
2007-02-07 22:39:03 +00:00
George Bosilca
2a8f348537 No more warnings on Windows.
This commit was SVN r13548.
2007-02-07 22:38:09 +00:00
George Bosilca
6724833eef Protect unistd.h usage.
This commit was SVN r13546.
2007-02-07 22:36:50 +00:00
Ralph Castain
c7be9a7121 Complete backout of prior sched_yield and paffinity changes
This commit was SVN r13530.
2007-02-07 14:22:37 +00:00
Pak Lui
b99b200aeb * fix a typo
This commit was SVN r13525.
2007-02-06 21:59:55 +00:00
George Bosilca
ebbc37584f Add Windows support for finding the number of available processors, and
return the same value if there is no registered paffinity module.

This commit was SVN r13509.
2007-02-06 16:53:13 +00:00
Jeff Squyres
0732c555de Refs trac:853
Add new function opal_get_num_processors() that will return the number
of processors on the local host.  Does the Right thing in POSIX
environments (to include a special case for OS X), and will shortly do
the Right Thing for Windows (this commit includes a change to
configure, so I wanted to get that in before the US workday -- the
Windows code can some shortly because it won't involve configury
changes).

This commit was SVN r13506.

The following Trac tickets were found above:
  Ticket 853 --> https://svn.open-mpi.org/trac/ompi/ticket/853
2007-02-06 12:03:56 +00:00
Jeff Squyres
c91fcd7fbd Fix a bunch of minor typos submitted by Bernhard Fischer.
This commit was SVN r13505.
2007-02-06 12:00:30 +00:00
George Bosilca
d5982eb5cb The whole affinity game for the Microsoft Windows platform.
This commit was SVN r13503.
2007-02-06 00:23:51 +00:00
Brian Barrett
cf8bc2ad0b print debugging information about switing the event flag and print the
initialization information

This commit was SVN r13497.
2007-02-05 19:38:54 +00:00