1
1

3527 Коммитов

Автор SHA1 Сообщение Дата
George Bosilca
5adaa88241 Cleanup the code and make it a little faster.
This commit was SVN r18038.
2008-03-31 17:12:03 +00:00
Matthias Jurenz
879fdc4feb merging VampirTrace-5.4.5 into the main branch
This commit was SVN r18030.
2008-03-31 12:48:35 +00:00
Matthias Jurenz
a33831c1f8 Pass OMPI's configure option '--[enable|disable]-binaries' to VT's configure
This commit was SVN r18029.
2008-03-31 12:46:27 +00:00
George Bosilca
60111ce66d Few less warnings.
This commit was SVN r18025.
2008-03-30 19:06:49 +00:00
George Bosilca
b4f828f389 We need a newline at the nd of the file, or some compiler bark.
This commit was SVN r18023.
2008-03-30 19:05:56 +00:00
Gleb Natapov
b42234461a Cleanup shared file creation on unix/linux.
This commit was SVN r18021.
2008-03-30 13:41:47 +00:00
Lenny Verkhovsky
7e45d7e134 Few updates due to RMAPS rank_file component changes
1. applied prefix rule to functions and variables of RMAPS rank_file component
2. cleaned ompi_mpi_init.c from paffinity code
3. paffinity code moved to new opal/mca/paffinity/base/paffinity_base_service.c file
4. added opal_paffinity_slot_list mca parameter

This commit was SVN r18019.
2008-03-30 11:52:11 +00:00
Jeff Squyres
d0f12f3df0 Make a better error message.
This commit was SVN r18014.
2008-03-29 12:54:24 +00:00
Rich Graham
3b42d2268d add functions to handle two different input buffers and a separate
output buffer.  User defined data types have not way to make use
of these.

This commit was SVN r18012.
2008-03-28 23:45:44 +00:00
Rich Graham
90e53ca9ee debug the pipeline algorithm.
This commit was SVN r18008.
2008-03-28 15:10:07 +00:00
Aurelien Bouteiller
77653ac787 Missing .h file in makefile breaked nightly tarball distcheck...
This commit was SVN r18006.
2008-03-28 14:36:56 +00:00
Aurelien Bouteiller
c16339944a Fix a coverity warning about using unsafe sprintf.
This commit was SVN r17999.
2008-03-27 21:24:27 +00:00
Aurelien Bouteiller
e11237aadb Introduction of the "progress" sender_based method to replace the slow isend-self method.
This commit was SVN r17998.
2008-03-27 21:19:45 +00:00
Aurelien Bouteiller
93db01871e This is part of the previous patch.
This commit was SVN r17997.
2008-03-27 21:06:14 +00:00
Aurelien Bouteiller
f8bf6f2c6a Code cleanup.
sender_based.h is now split in two files, to solve cyclic .h files inclusion. 
Most macros are now inline functions.
Variable names have been changed from places to places.
Various other small things... 

This commit was SVN r17996.
2008-03-27 21:05:44 +00:00
George Bosilca
691806680a I guess this wasn't really intended ...
This commit was SVN r17995.
2008-03-27 18:41:06 +00:00
George Bosilca
303941f642 Avoid a deadlock. The comment explain how this might happen.
This commit was SVN r17994.
2008-03-27 18:37:11 +00:00
George Bosilca
be4b153f0d Another patch for thread safety in the TCP BTL (thanks to Pierre).
This commit was SVN r17993.
2008-03-27 18:36:08 +00:00
Tim Prins
c5736e3f9a Remove old constants used with the registry.
This commit was SVN r17991.
2008-03-27 17:13:20 +00:00
Ralph Castain
6166278e18 Improve the scalability of the modex operation and fix a bug reported by Tim P
The bug was a race condition in the barrier operation that caused the barrier in MPI_Finalize to fail on very short programs.

Scalaiblity was improved by using the daemons to aggregate modex and barrier messages before sending them to the rank=0 proc. Improvement is proportional to ppn, of course, but there really wasn't a scaling problem at low ppn anyway. This modification also paves the way for better allgather operations since now all the data for each node is sitting at the daemon level, and the daemons are now aware that a collective operation on the OOB is underway (so they -can- participate in a collective of their own to support it).

Also added better diagnostics to map out the timing associated with MPI_Init - turned on by -mca orte_timing 1.

This commit was SVN r17988.
2008-03-27 15:17:53 +00:00
Gleb Natapov
cf40674369 Decide if sends should be throttled at the receiver and pass this to the sender
in an ACK message. The decision can't be done reliably at the sender.

This commit was SVN r17987.
2008-03-27 08:56:43 +00:00
Rich Graham
e2ad9c4be2 adjust to change in orte_process_info.
This commit was SVN r17986.
2008-03-27 01:25:28 +00:00
Rich Graham
441fb9fb9e checkpoint.
This commit was SVN r17985.
2008-03-27 01:16:32 +00:00
Jeff Squyres
a2795fe43d Very minor modification against r17980: check the whole string against
"all", not just the first 3 chars (i.e., if someone sets the value
"allfoo", we should still error).

This commit was SVN r17981.

The following SVN revision numbers were found above:
  r17980 --> open-mpi/ompi@b3ef774d46
2008-03-26 19:10:02 +00:00
Josh Hursey
b3ef774d46 A fix for r17956.
r17956 broke the ability for the user to override the 'opal_event_include'
parameter. This commit checks to see if the user specified a value before
forcing the "all" value on the event engine.

This commit fixes Checkpoint/Restart support in the trunk which requires
this feature.

This commit was SVN r17980.

The following SVN revision numbers were found above:
  r17956 --> open-mpi/ompi@763218e754
2008-03-26 14:54:09 +00:00
Rainer Keller
b7efc2b18e - Coverity issues CID 42:
Event var_deref_model: Variable "array_of_integers" tracked as NULL was
   passed to a function that dereferences it. [model]
   The arrays passed down type_get_contents may be NULL, only iff max_* is 0...
   If the max_* parameter does not fit, an error is returned, anyhow.
   One could improve the checks of MPI_PARAM_CHECK, but to be on the
   safe side, fix in dt_args.c.

This commit was SVN r17974.
2008-03-26 09:07:06 +00:00
Rainer Keller
334b64e760 - Coverity issue CID 35:
Event var_deref_op: Variable "requests" tracked as NULL was
   dereferenced.
   Only check requests[i] for NULL, if requests is != NULL itself.

This commit was SVN r17973.
2008-03-26 08:19:55 +00:00
Rainer Keller
56f3d59f2a - Coverity issues 939, 940, 941:
Event uninit_use_in_call: Using uninitialized value "tag" in call to
   function "(ompi_dpm).connect_accept" and others
   The tag is set and used in get_rport only on root...

This commit was SVN r17972.
2008-03-26 08:09:11 +00:00
George Bosilca
a01f3f762c Check if extra is NULL or not ...
This commit was SVN r17967.
2008-03-25 22:43:46 +00:00
George Bosilca
bea5c0f734 Don't allocate anything if we don't really need it, and avoid leaking memory.
This commit was SVN r17966.
2008-03-25 22:43:11 +00:00
Jeff Squyres
763218e754 Fix #1253: default libevent to use select/poll and only use the other
mechanisms (such as epoll) if someone (ompi_mpi_init()) requests
otherwise.  See big comment in opal/event/event.c for a full
explanation.

This commit was SVN r17956.
2008-03-25 17:18:17 +00:00
Ralph Castain
90107f3c14 Fix an issue with comm_spawn over who sent/recv first in the modex. The modex assumes that the first name on the list is the "root" that will serve as the allgather collector/distributor. The dpm was putting that entity last, which forced us to pre-inform the parent procs of the child proc's contact info since the parent was trying to send to the child.
Clarify the setting of send_first in the mpi bindings (trivial, i know, but helpful)

Remove the extra xcast of child contact info to the parent job.

This commit was SVN r17952.
2008-03-25 14:57:34 +00:00
Ralph Castain
cca449e379 Move an OMPI RML tag to the OMPI layer
This commit was SVN r17950.
2008-03-25 13:30:48 +00:00
Jeff Squyres
5320c91ab3 Oops -- fix the constructor to also use opal_object_t instead of
opal_list_item_t.

This commit was SVN r17945.
2008-03-25 11:59:50 +00:00
Galen Shipman
0116041133 BTL shouldn't own the passive side's descriptor in the PML get protocol. The BTL
doesn't know when to free it on the passive side. 

This commit was SVN r17943.
2008-03-25 01:43:41 +00:00
Ralph Castain
ebea4d04e4 Remove defunct error constant - we no longer have a GPR that can hold corrupt data!
This commit was SVN r17942.
2008-03-24 21:05:14 +00:00
Jeff Squyres
ebfdd133f5 AFACT, we never put endpoints on a list.
This commit was SVN r17940.
2008-03-24 18:32:55 +00:00
Jeff Squyres
004c3a5b09 Ensure to cover all cases when either ORTE or OMPI is not yet
initialized.  For example, there is a period of time during
ompi_mpi_init when orte_initialized==true, but
ompi_mpi_initialized==false (and therefore communicators are not setup
yet, etc.).

This commit was SVN r17937.
2008-03-24 16:25:14 +00:00
Ralph Castain
dc7f45dafd Remove the obsolete and largely unused orte_system_info structure. The only fields that were used in that struct were nodeid and nodename - these have been transferred to the orte_process_info structure.
Only one place used the user name field - session_dir, when formulating the name of the top-level directory. Accordingly, the code for getting the user's id has been moved to the session_dir code.

This commit was SVN r17926.
2008-03-23 23:10:15 +00:00
Jeff Squyres
dee561d29e Per recent off-list discussions about the build system, I have done
some cleanups and standardizations in the various */tools/*/ 
Makefile.am files.  This commit:

 * Somewhat simplify the tool Makefile.am's 
 * Makes the tool Makefile.am's consistent with each other (do similar
   actions in similar ways)
 * Update the tool Makefile.am's to remove old kruft that was required
   by older versions of AM (trunk requires AM >=1.10)

This commit was SVN r17921.
2008-03-22 02:04:05 +00:00
Brian Barrett
f176c67cd2 Set the nodeid to something somewhat sane if we're not using modex, and
don't set the LOCAL flag just because both procs have an invalid nodeid.

This commit was SVN r17917.
2008-03-21 20:20:00 +00:00
Brian Barrett
5a7ebf5f25 Do not try to update the local process with modex information (from the local
process) as it stomps on information if the modex doesn't exist for the
current platform

This commit was SVN r17916.
2008-03-21 19:20:47 +00:00
Jeff Squyres
4fbcb75ce8 With 5 commits over a 16 hour period and 3 broken tarball builds and a
still-broken trunk build on common platforms (e.g., 64 bit Linux
RHEL4U4), I think it's clear that this code is not ready for
prime-time.

I'm backing out all the commits in the trunk/ompi/op tree from r17901
onwards.  This code can be re-committed when compiles and runs on
common platforms.

cd ompi/op
svn merge -r 17907:17900 https://svn.open-mpi.org/svn/ompi/trunk/ompi/op .

This commit was SVN r17908.

The following SVN revision numbers were found above:
  r17901 --> open-mpi/ompi@b9520e61dc
2008-03-21 14:47:01 +00:00
Jeff Squyres
8284f64af1 With r17906, this commit should make the trunk compile again.
This commit was SVN r17907.

The following SVN revision numbers were found above:
  r17906 --> open-mpi/ompi@df4a6c3fc5
2008-03-21 13:49:23 +00:00
Rich Graham
df4a6c3fc5 fix function prototypes for new 3 buffer routines.
This commit was SVN r17906.
2008-03-21 13:44:15 +00:00
Ralph Castain
b2655ab585 Per Brian's suggestion, remove unnecessary library dependency - libtool automagically picks up the other libraries when we include libmpi
This commit was SVN r17905.
2008-03-21 12:47:04 +00:00
Rich Graham
0974160e29 correct several of the new macros.
This commit was SVN r17904.
2008-03-21 03:45:43 +00:00
Rich Graham
a7c836a2b0 fix location of the restrict key word.
Make the tag in the fan-in/fan-out algorithm be fragment based.

This commit was SVN r17903.
2008-03-21 01:40:36 +00:00
Rich Graham
2c66d396b7 take care of some bit-rot with the fanin-fanout method.
This commit was SVN r17902.
2008-03-21 01:08:49 +00:00
Rich Graham
b9520e61dc get the sm optimized allreduce working for all but user defined
operations.  Added to the reduction operations a set of reduction
functions that take 2 input buffers and one output buffer to avoid
some extra memory copies.  These can't be used with user defined
operations.  The intel c collective suite passes both original, and
new (new, not the user defined operations).

This commit was SVN r17901.
2008-03-20 23:51:16 +00:00