ports/openmpi
ports/openmpi
1
1
Форкнуть 0
Код Релизы Активность
13 938 коммитов 3 Ветки 0 Теги 130 MiB
3a2f071018
Граф коммитов

13938 Коммитов

This Branch
This Branch
Все ветки
Автор SHA1 Сообщение Дата
George Bosilca
3a2f071018 If the user asked for dynamic rules but "forget" to provide them, nicely 
complain and switch back to the default behavior (fixed rules).

This commit was SVN r22109.
 2009-10-19 17:58:47 +00:00
Ralph Castain
941b8722b4 Remove an errant entry in the platform files 
This commit was SVN r22108.
 2009-10-16 16:12:37 +00:00
Ralph Castain
ff9d72b3ab Add a new multicast tag for collecting ps data 
This commit was SVN r22107.
 2009-10-16 04:21:22 +00:00
Terry Dontje
13907781b2 missed adding report-uri option 
This commit was SVN r22106.
 2009-10-15 18:05:24 +00:00
Ralph Castain
49ce2b4342 Add a new interface to the rmcast framework to query the output channel for the proc 
This commit was SVN r22105.
 2009-10-15 17:47:42 +00:00
Terry Dontje
c96af5654c correct options and wording that were dropped in the last change due to committing v1.3 manpage to the trunk 
This commit was SVN r22104.
 2009-10-15 15:03:21 +00:00
Ralph Castain
99c67183d2 Minor cleanups, mainly to ensure we correctly block on blocking sends 
This commit was SVN r22102.
 2009-10-15 02:39:15 +00:00
Ralph Castain
2f91a4833b Have the trigger event return the event itself in the callback function so it can be reset, if desired 
This commit was SVN r22101.
 2009-10-15 02:35:53 +00:00
Ralph Castain
2665825693 Correct an error that causes the system to "bounce" when we order a job killed. We didn't used to discriminate between a process being ordered to die, and a process that was aborted by an external signal. Unfortunately, that means the error mgr gets called and told a process abnormally aborted when we order termination, thus causing the errmgr to send out a "kill procs" command again. 
Wouldn't be so bad, except...the errmgr orders the termination of ALL procs, which kills any other job that should have been left alone.

Add a new proc and job state indicating "killed_by_cmd" so we can tell the difference between a proc/job that was deliberately terminated by us vs one that is killed by external signal.

This change was tested to ensure it didn't interfere with ctrl-c operation (it doesn't - we order termination of all jobs when we get a ctrl-c).

This commit was SVN r22100.
 2009-10-14 22:49:56 +00:00
Ralph Castain
18960a9c5a Refactor the multicast support so the data type objects can be accessed beyond just the one component 
Ensure that the local node is included in the allocation prior to bootstrap discovery

This commit was SVN r22099.
 2009-10-14 17:43:40 +00:00
Terry Dontje
0a8645a411 This commit fixes trac:2017 
This commit was SVN r22098.

The following Trac tickets were found above:
  Ticket 2017 --> https://svn.open-mpi.org/trac/ompi/ticket/2017
 2009-10-14 11:40:47 +00:00
Ralph Castain
60c4ebab45 Update makefile so it finds new platform file directory 
This commit was SVN r22097.
 2009-10-14 01:48:30 +00:00
Ralph Castain
3a046b7262 Update platform files for cisco clusters 
This commit was SVN r22096.
 2009-10-14 01:47:32 +00:00
Ralph Castain
bc869636be Reset the verbosity levels to suppress debug output 
This commit was SVN r22095.
 2009-10-13 15:29:38 +00:00
Ralph Castain
e501589b3b Cleanup the bootstrap procedure for multiple daemons starting up 
This commit was SVN r22094.
 2009-10-13 15:14:54 +00:00
Ralph Castain
c25dd14440 Correctly set the multicast interface, cleanup a comment 
This commit was SVN r22093.
 2009-10-13 15:14:28 +00:00
Ralph Castain
d8d80d6f1a Closes trac:2054. Check if a user specifies more cpus-per-rank than there are cpus in a socket - if so, politely tell them "you are stupid" and abort. 
This commit was SVN r22091.

The following Trac tickets were found above:
  Ticket 2054 --> https://svn.open-mpi.org/trac/ompi/ticket/2054
 2009-10-13 04:19:07 +00:00
Ralph Castain
1475d34c13 Ensure we default to byslot mapping 
This commit was SVN r22090.
 2009-10-11 23:50:42 +00:00
Ralph Castain
67625d5df6 Esure that mca params have a chance to be used in the standard hierarchy - in this case, allowing a choice of mapping policy to be made at the rmaps mca level. 
This commit was SVN r22089.
 2009-10-11 03:44:39 +00:00
Jeff Squyres
ad62148bc2 Trunk will now eventually become v1.7 
This commit was SVN r22088.
 2009-10-10 19:26:55 +00:00
Ralph Castain
84cc847be8 Next phase of auto-wireup using multicast. Enable use of multicast groups to separate comm from different application groups. Have the orted bootstrap message go to a different rml tag so the node can be added to the pool. 
This commit was SVN r22083.
 2009-10-10 01:19:56 +00:00
Ralph Castain
40e2299fa7 Test to ensure that num_procs was provided for the resilient mapper - it cannot be used with options like npernode. 
Cleanup the show_help text file

This commit was SVN r22082.
 2009-10-09 15:26:23 +00:00
Ralph Castain
b7a0125bb7 Add a test for the new opal if.c functions. Modify the multicast test 
This commit was SVN r22081.
 2009-10-09 15:25:18 +00:00
Ralph Castain
c58a30ea10 Add two new functions: 
1. check for loopback interface

2. convert tuple addresses to ip addrs + mask

This commit was SVN r22080.
 2009-10-09 15:24:41 +00:00
Jeff Squyres
c4f2db926f Add missing semicolons. Wow. 
This commit was SVN r22079.
 2009-10-08 19:50:19 +00:00
Terry Dontje
0828945eea Fix an issue with #2048 fix that did not goto the error case. 
This commit was SVN r22076.
 2009-10-08 13:27:32 +00:00
Terry Dontje
58c864699c This commit fixes trac:2048 
This commit was SVN r22075.

The following Trac tickets were found above:
  Ticket 2048 --> https://svn.open-mpi.org/trac/ompi/ticket/2048
 2009-10-08 12:54:53 +00:00
Jeff Squyres
3dc84e9d0b Change the default value of shell_scripts_basename to not include the 
version because they're installed in bindir by default, where you can
only have one Open MPI installation at a time.  Plus, without the
version numbers is what mpi-selector expects.

Thanks to Bill Johnstone for pointing out the problem.

This commit was SVN r22074.
 2009-10-08 11:47:53 +00:00
Jeff Squyres
9afe50d886 Update Cisco copyrights for consistency 
This commit was SVN r22072.
 2009-10-07 22:02:32 +00:00
Jeff Squyres
0d1e177453 Remove 2 extraneous ORTE_ERROR_LOGs and 1 extraneous opal_output. 
This commit was SVN r22071.
 2009-10-07 20:12:37 +00:00
Jeff Squyres
d317ce0367 Fix CID 1381: don't bother checking for (NULL == p); it's overkill. 
posix_memalign() will either return 0 or not, indicating success.  And
if posix_memalign() fails, it's not always going to be due to
out-of-memory -- just return ERR_IN_ERRNO.

This commit was SVN r22070.
 2009-10-07 20:01:50 +00:00
Jeff Squyres
7900451e4e Fix CID 1326: for the (unlikely) case where 
opal_paffinity_base_get_processor_info() returns failure.

This commit was SVN r22069.
 2009-10-07 19:52:08 +00:00
Jeff Squyres
5c1af9c2ba Fix CID 1355: ensure that mca_base_param_reg_int() actually 
succeeded.

This commit was SVN r22068.
 2009-10-07 19:43:35 +00:00
Jeff Squyres
d56b8d9183 Fix CID 1369: minor memory leak. 
This commit was SVN r22067.
 2009-10-07 19:40:00 +00:00
Jeff Squyres
de59a24593 Fix CID 1384. Also remove some opal_output(0,...)'s in favor of 
ORTE_ERROR_LOG.

This commit was SVN r22066.
 2009-10-07 18:58:58 +00:00
Jeff Squyres
ec71acf7ca Fix CID 1385: fix an over-aggressive use of close, munmap, etc. in the 
error case.  Also check for MAP_FAILED (instead of -1) from mmap().

This commit was SVN r22065.
 2009-10-07 18:43:37 +00:00
Jeff Squyres
5ec86e5fe5 Fix CID 1386: fd can't be valid here, so don't bother to close/unlink. 
This commit was SVN r22064.
 2009-10-07 18:30:26 +00:00
Jeff Squyres
3b4f695009 MAP_FAILED is more POSIX-ly correct than ((void*)-1). 
This commit was SVN r22063.
 2009-10-07 14:20:18 +00:00
Jeff Squyres
d7db5f4c32 mmap(2) says that you must call mmap() with either MAP_SHARED or 
MAP_PRIVATE.  We didn't catch this because we checked for a NULL
return, not a -1 return.  Doh!  Thanks again to Julian Seward for
continuing to track this down.

This commit was SVN r22062.
 2009-10-07 12:39:01 +00:00
Jeff Squyres
977574bd45 Fix a problem noted by Julian Seward: MAKE_MEM_UNDEFINED is not the 
opposite of MAKE_MEM_DEFINED. Also add in a call to NOACCESS to
(mostly) reverse the effects of MAKE_MEM_DEFINED (technically, page 0
was accessible before this, even though it's a Bad Idea to access it).

This commit was SVN r22056.
 2009-10-06 17:55:49 +00:00
Jeff Squyres
932b43be04 Check to ensure that the mmap succeeded. Thanks to Julia Seward for 
pointing out the problem and suggesting the fix.

This commit was SVN r22055.
 2009-10-06 17:44:14 +00:00
Shiqing Fan
14e6952482 Update two CMake find modules. 
This commit was SVN r22054.
 2009-10-06 08:01:37 +00:00
Shiqing Fan
7dff65cbc9 Clean up a little bit. 
Add an option for setting up the job name.

This commit was SVN r22053.
 2009-10-06 07:52:43 +00:00
George Bosilca
01bb4dafe0 Add a comment. 
This commit was SVN r22052.
 2009-10-05 17:36:11 +00:00
Jeff Squyres
0f8ac9223f Refs trac:2023, #2027. 
This commit does a bunch of things:

 * Address all remaining code review items from CMR #2023:

   * Defer mmap setup to be lazy; only set it up the first time we
     invoke a collective.  In this way, we don't penalize apps that
     make lots of communicators but don't invoke collectives on them
     (per #2027).
   * Remove the extra assignments of mca_coll_sm_one (fixing a
     convertor count setup that was the real problem).
   * Remove another extra/unnecessary assignment.
   * Increase libevent polling frequency when using the RML to
     bootstrap mmap'ed memory.
   * Fix a minor procs-related memory leak in btl_sm.
 * Commit a datatype fix that George and I discovered along the way to
   fixing the coll sm.
 * Improve error messages when mmap fails, potentially trying to
   de-alloc any allocated memory when that happens.
 * Fix a previously-unnoticed confusion between extent and true_extent
   in coll sm reduce.

This commit was SVN r22049.

The following Trac tickets were found above:
  Ticket 2023 --> https://svn.open-mpi.org/trac/ompi/ticket/2023
 2009-10-02 17:13:56 +00:00
Ralph Castain
dcab61ad83 Restore the prior default rank assignment scheme for round-robin mappers. Ensure that each app_context has sequential vpids. 
This commit was SVN r22048.
 2009-10-02 03:16:18 +00:00
Jeff Squyres
c8c3132605 Also check for posix_memalign. 
This commit was SVN r22045.
 2009-10-01 23:51:48 +00:00
George Bosilca
cf9f38eb56 Instead of just complaining about a version mismatch, clearly lists the versions 
available locally.

This commit was SVN r22044.
 2009-10-01 14:06:41 +00:00
George Bosilca
16c6370b73 A little bit of cleanup, the main logic is still the same. 
This commit was SVN r22043.
 2009-10-01 14:05:25 +00:00
Ralph Castain
a15c58c583 Fix the proc assignment into the job data object during assignment of vpids as comm_spawned procs were being overwritten by their parents with the same vpid. 
Add a little debug output when updating proc state

This commit was SVN r22042.
 2009-10-01 13:44:34 +00:00