1
1
Граф коммитов

3400 Коммитов

Автор SHA1 Сообщение Дата
Ralph Castain
e49ca05f11 Remove unused variable
This commit was SVN r32651.
2014-08-31 03:11:50 +00:00
Ralph Castain
5cdbc00136 Re-enable the usock oob component. Ensure the TCP component promotes messages for other procs to the OOB base so that other components have a chance to send the relay. Seems to be passing MTT, so let's see how it works for others.
This commit was SVN r32650.
2014-08-30 19:33:46 +00:00
Ralph Castain
a2085a5916 Fix the PSM transport key generator to match prior releases
This commit was SVN r32649.
2014-08-30 00:48:25 +00:00
Ralph Castain
8faabed2cd Add some further initialization and protection for zero-byte messages
This commit was SVN r32644.
2014-08-29 17:24:55 +00:00
Ralph Castain
2b225e3776 Cleanup a race condition regarding marking that waitpid_fired. We should always mark it as fired when we enter the wait_local_proc routine, and also mark it as no longer alive if iof_complete has also been found. If other places in the code also update those flags, there is no harm done.
This commit was SVN r32643.
2014-08-29 17:03:31 +00:00
Ralph Castain
730e28349e Some minor uninitialized variable cleanups
This commit was SVN r32629.
2014-08-29 02:21:13 +00:00
Ralph Castain
fafdbeec0c Cleanup and enable the new daemon collective modules for more scalable operations. Thanks to Nadezhda Kogteva (Mellanox) for doing them.
This commit was SVN r32624.
2014-08-28 20:35:35 +00:00
Ralph Castain
731a878ff3 Add a bunch of debug to help track down the problem, and eventually find another place where comparison of signatures was incorrectly performed - use the dss compare operation to be consistent and safe
This commit was SVN r32620.
2014-08-27 19:52:20 +00:00
Ralph Castain
b87b69e977 Ensure the nodes get added to the job map on the remote nodes, add some debug to grpcomm daemon array construction
This commit was SVN r32617.
2014-08-27 16:16:46 +00:00
Ralph Castain
842aaf6167 Correctly end mapping oversubscribed nodes round-robin byslot
cmr=v1.8.3:reviewer=rhc

This commit was SVN r32616.
2014-08-27 16:15:18 +00:00
Ralph Castain
5a13cdb739 Fix a race condition caused by a bad attribute flag that created an OR instead of an AND condition check
This commit was SVN r32587.
2014-08-22 22:48:16 +00:00
Ralph Castain
039b7acfb5 Fix the quoting algorithm so only rsh command lines get quoted values
cmr=v1.8.2:reviewer=jsquyres

This commit was SVN r32586.
2014-08-22 22:47:38 +00:00
Ralph Castain
f00af81c1d Little more cleanup under the abort cases cited by Gilles. All seem to be working now
This commit was SVN r32585.
2014-08-22 19:57:57 +00:00
Ralph Castain
8f1b9b463e Fix shared memory operations - need to pass the local topology and cpusets of all local peers so we can properly compute relative locality for them. Also need to set default locality to "on node" in case where cpusets are not passed because procs are not bound.
This commit was SVN r32577.
2014-08-22 05:17:51 +00:00
Ralph Castain
c6f78d6e54 The PMI ess component now gets used for more than direct launch, so only set standalone_operation flag if no daemon uri is available so we aggregate show_help messages
This commit was SVN r32574.
2014-08-22 03:00:56 +00:00
Ralph Castain
aec5cd08bd Per the PMIx RFC:
WHAT:    Merge the PMIx branch into the devel repo, creating a new
               OPAL “lmix” framework to abstract PMI support for all RTEs.
               Replace the ORTE daemon-level collectives with a new PMIx
               server and update the ORTE grpcomm framework to support
               server-to-server collectives

WHY:      We’ve had problems dealing with variations in PMI implementations,
               and need to extend the existing PMI definitions to meet exascale
               requirements.

WHEN:   Mon, Aug 25

WHERE:  https://github.com/rhc54/ompi-svn-mirror.git

Several community members have been working on a refactoring of the current PMI support within OMPI. Although the APIs are common, Slurm and Cray implement a different range of capabilities, and package them differently. For example, Cray provides an integrated PMI-1/2 library, while Slurm separates the two and requires the user to specify the one to be used at runtime. In addition, several bugs in the Slurm implementations have caused problems requiring extra coding.

All this has led to a slew of #if’s in the PMI code and bugs when the corner-case logic for one implementation accidentally traps the other. Extending this support to other implementations would have increased this complexity to an unacceptable level.

Accordingly, we have:

* created a new OPAL “pmix” framework to abstract the PMI support, with separate components for Cray, Slurm PMI-1, and Slurm PMI-2 implementations.

* Replaced the current ORTE grpcomm daemon-based collective operation with an integrated PMIx server, and updated the grpcomm APIs to provide more flexible, multi-algorithm support for collective operations. At this time, only the xcast and allgather operations are supported.

* Replaced the current global collective id with a signature based on the names of the participating procs. The allows an unlimited number of collectives to be executed by any group of processes, subject to the requirement that only one collective can be active at a time for a unique combination of procs. Note that a proc can be involved in any number of simultaneous collectives - it is the specific combination of procs that is subject to the constraint

* removed the prior OMPI/OPAL modex code

* added new macros for executing modex send/recv to simplify use of the new APIs. The send macros allow the caller to specify whether or not the BTL supports async modex operations - if so, then the non-blocking “fence” operation is used, if the active PMIx component supports it. Otherwise, the default is a full blocking modex exchange as we currently perform.

* retained the current flag that directs us to use a blocking fence operation, but only to retrieve data upon demand

This commit was SVN r32570.
2014-08-21 18:56:47 +00:00
Ralph Castain
b4511913f6 Remove an unnecessary optimization that can cause more trouble than it's worth - just try all the addresses that are given to us.
Refs trac:4870

This commit was SVN r32558.

The following Trac tickets were found above:
  Ticket 4870 --> https://svn.open-mpi.org/trac/ompi/ticket/4870
2014-08-20 20:58:07 +00:00
Ralph Castain
fa28710d53 Track down the last piece of the connection problem. It appears that
providing a netmask of 0 to opal_net_samenetwork results in everything
looking like it is on the same network. Hence, we were not retaining any
of the alternative addresses, so we had no other way to check them.

Refs trac:4870

This commit was SVN r32556.

The following Trac tickets were found above:
  Ticket 4870 --> https://svn.open-mpi.org/trac/ompi/ticket/4870
2014-08-20 16:55:36 +00:00
Ralph Castain
343038af7b Frazzle-frump! Missed that we reset the peer state just before the new check.
Refs trac:4870

This commit was SVN r32554.

The following Trac tickets were found above:
  Ticket 4870 --> https://svn.open-mpi.org/trac/ompi/ticket/4870
2014-08-19 22:34:49 +00:00
Ralph Castain
0a91fdf85f If an initial address fails to connect, record that fact and attempt the next address for that proc. If nothing succeeds, then declare failure.
cmr=v1.8.2:reviewer=edgar

This commit was SVN r32553.
2014-08-19 19:48:24 +00:00
Ralph Castain
024572cb6c Sigh - I promised to remove these deprecation warnings back in June. My apologies to Dave Goodell and others who requested it.
cmr=v1.8.2:reviewer=dgoodell:subject=remove deprecation warnings for pernode, npernode, and npersocket

This commit was SVN r32552.
2014-08-19 19:40:20 +00:00
Jeff Squyres
1551339eba rsh: revert part of r32517: keep the quoting
As part of reviewing CMR #4860, I talked through r32517 with Ralph.

In attempt to fix various rsh quoting problems, r32517 removed all the
quoting from the main code path and then only added it back in at the
end in some cases.

This commit puts back the quoting parts that were removed in r32517
(r32517 fixed 2 other important bugs: a) change "--<foo>" to "--mca
<foo_equivalent> 1" so that de-duplication works, and b) change a !=
to ==).

refs trac:4860

This commit was SVN r32524.

The following SVN revision numbers were found above:
  r32517 --> open-mpi/ompi@7342bce58f

The following Trac tickets were found above:
  Ticket 4860 --> https://svn.open-mpi.org/trac/ompi/ticket/4860
2014-08-13 19:27:10 +00:00
Ralph Castain
7342bce58f Cleanup the over-aggressive quoting of params on the orted cmd line. Remove duplicates caused by passing on both cmd line shortcuts and the mca param version of the same thing.
Fixes trac:4857

cmr=v1.8.2:reviewer=jsquyres

This commit was SVN r32517.

The following Trac tickets were found above:
  Ticket 4857 --> https://svn.open-mpi.org/trac/ompi/ticket/4857
2014-08-13 03:51:04 +00:00
George Bosilca
de7191132d Remove few warnings.
This commit was SVN r32506.
2014-08-11 13:34:44 +00:00
Gilles Gouaillardet
a873f45a90 Fix r32460 race condition resolution when procs call MPI_Abort.
do not invoke orte_session_dir_finalize(...) so
orte_ess_base_app_abort(...) can successfully createi
<orte_process_info.proc_session_dir>/aborted

cmr=v1.8.2:reviewer=rhc

This commit was SVN r32498.

The following SVN revision numbers were found above:
  r32460 --> open-mpi/ompi@abedb97be4
2014-08-11 05:50:32 +00:00
Gilles Gouaillardet
e184733ef6 check-help-strings cleanup
This commit was SVN r32496.
2014-08-11 03:26:21 +00:00
Gilles Gouaillardet
c3c364a262 check-help-strings cleanup
This commit was SVN r32494.
2014-08-11 03:22:05 +00:00
Gilles Gouaillardet
d9e0212e0e check-help-strings cleanup
This commit was SVN r32493.
2014-08-11 03:21:08 +00:00
Gilles Gouaillardet
d139f75db4 check-help-strings cleanup
This commit was SVN r32492.
2014-08-11 03:20:37 +00:00
Ralph Castain
abedb97be4 Resolve race condition when procs call MPI_Abort. Since we go thru the errmgr instead of the normal proc termination routines, we need to ensure we mark that the proc has fired its waitpid and is no longer alive. Otherwise, the local daemon won't terminate because it thinks there is still a local proc alive and we hang.
Thanks to Gilles for tracking it down.

cmr=v1.8.2:reviewer=rhc

This commit was SVN r32460.
2014-08-08 15:58:49 +00:00
Ralph Castain
8ea576c870 I have no idea how they did it, but someone managed to write a test that circled around and around and eventually reached this point with a NULL pointer. So protect against that possibility.
This commit was SVN r32434.
2014-08-05 16:20:46 +00:00
Ralph Castain
d29a5ab69d Okay, now handle the non-MPI apps
This commit was SVN r32399.
2014-08-01 14:49:25 +00:00
Ralph Castain
daeb9b6c4f Some more cleanups. Remove direct references to ORTE by changing OMPI_CAST_ORTE_NAME -> OMPI_CAST_RTE_NAME. Ensure that ORTE tools (mpirun, orted, tools) set the OPAL proc structure fields so OPAL knows what is going on and uses the correct print functions (still need to fix the problem for non-MPI apps). Properly return uint32_t from the opal utilities instead of int32_t as that is what the ORTE process name fields contain.
Thanks to Gilles for pointing out some of the discrepancies.

This commit was SVN r32398.
2014-08-01 14:44:11 +00:00
Ralph Castain
8cfadd1842 Per Paul Hargrove, add missing include to fix build under FreeBSD
RM-approved

cmr=v1.8.2:reviewer=ompi-gk1.8

This commit was SVN r32397.
2014-08-01 13:37:41 +00:00
George Bosilca
daa076995a orte_rmaps_numa_node_t -> opal_rmaps_numa_node_t
This commit was SVN r32380.
2014-07-31 19:58:47 +00:00
Ralph Castain
5db717f090 Some small leak cleanups
cmr=v1.8.3:reviewer=artpol

This commit was SVN r32358.
2014-07-30 15:46:02 +00:00
Ralph Castain
98b5a86a58 Now that we are using the radix routed module, teach it how to behave nicely with singletons
This commit was SVN r32312.
2014-07-24 22:46:17 +00:00
Ralph Castain
0cad281a92 Single-word cmd line values for orted are dealt with in orte_plm_base_orted_append_basic_args, so protect against special characters there. Have the rsh module only deal with multi-word arguments as those were skipped by orte_plm_base_orted_append_basic_args.
Refs trac:4802

This commit was SVN r32293.

The following Trac tickets were found above:
  Ticket 4802 --> https://svn.open-mpi.org/trac/ompi/ticket/4802
2014-07-23 17:06:51 +00:00
Ralph Castain
a94a97bd50 Cleanup the passing of MCA params on the orted cmd line in ssh by ensuring that we quote all values since they could be multi-word and/or contain special characters. Thanks to Dirk Schubert for pointing it out.
cmr=v1.8.2:reviewer=jsquyres

This commit was SVN r32280.
2014-07-22 18:22:06 +00:00
Ralph Castain
2f579806ae Make the radix routed component the default pending repair/completion of debruijn option
This commit was SVN r32276.
2014-07-22 16:48:51 +00:00
Ralph Castain
6c5e592785 Revert r32222, r32210, and r32203 as they created a problem when daemon collectives did not involve app procs on every node. Instead, modify the ompi/mca/rte/orte/rte_orte.h to add a new function that allows apps to request new daemon collective ids for use in barrier and modex operations. This will only appear in ORTE-based installations, but it is only being used by a couple of researchers at the moment.
Update the orte/test/mpi/coll_test.c test to show the revised example.

This commit was SVN r32234.

The following SVN revision numbers were found above:
  r32203 --> open-mpi/ompi@a523dba41d
  r32210 --> open-mpi/ompi@2ce11ed5c4
  r32222 --> open-mpi/ompi@d55f16db50
2014-07-15 03:48:00 +00:00
Ralph Castain
1feaffbb15 Get the blasted singleton comm_spawn working again. There remain problems with the Slurm interaction in this use-case as the PMI components (if configured to build) try to run even when a Slurm allocation hasn't been made, but I leave that to someone else to resolve. I did, however, tell the Slurm ess to quit interfering with applications launched in this use-case by ORTE daemons, so things do work when inside a Slurm allocation.
Also discovered that the rsh launcher is not picking up --enable-orterun-prefix-by-default when invoked during singleton comm_spawn, but I was unable to see why that was happening and ran out of time.

cmr=v1.8.2:reviewer=rhc

This commit was SVN r32229.
2014-07-13 14:47:22 +00:00
Ralph Castain
d55f16db50 Fix a hang in daemon collectives when run on multinode systems
This commit was SVN r32222.
2014-07-12 00:43:12 +00:00
Ralph Castain
2ce11ed5c4 Fix one spot missed by recent commit
This commit was SVN r32210.
2014-07-10 20:08:57 +00:00
Ralph Castain
a523dba41d NOTE: this modifies the MPI-RTE interface
We have been getting several requests for new collectives that need to be inserted in various places of the MPI layer, all in support of either checkpoint/restart or various research efforts. Until now, this would require that the collective id's be generated at launch. which required modification
s to ORTE and other places. We chose not to make collectives reusable as the race conditions associated with resetting collective counters are daunti
ng.

This commit extends the collective system to allow self-generation of collective id's that the daemons need to support, thereby allowing developers to request any number of collectives for their work. There is one restriction: RTE collectives must occur at the process level - i.e., we don't curren
tly have a way of tagging the collective to a specific thread. From the comment in the code:

 * In order to allow scalable
 * generation of collective id's, they are formed as:
 *
 * top 32-bits are the jobid of the procs involved in
 * the collective. For collectives across multiple jobs
 * (e.g., in a connect_accept), the daemon jobid will
 * be used as the id will be issued by mpirun. This
 * won't cause problems because daemons don't use the
 * collective_id
 *
 * bottom 32-bits are a rolling counter that recycles
 * when the max is hit. The daemon will cleanup each
 * collective upon completion, so this means a job can
 * never have more than 2**32 collectives going on at
 * a time. If someone needs more than that - they've got
 * a problem.
 *
 * Note that this means (for now) that RTE-level collectives
 * cannot be done by individual threads - they must be
 * done at the overall process level. This is required as
 * there is no guaranteed ordering for the collective id's,
 * and all the participants must agree on the id of the
 * collective they are executing. So if thread A on one
 * process asks for a collective id before thread B does,
 * but B asks before A on another process, the collectives will
 * be mixed and not result in the expected behavior. We may
 * find a way to relax this requirement in the future by
 * adding a thread context id to the jobid field (maybe taking the
 * lower 16-bits of that field).

This commit includes a test program (orte/test/mpi/coll_test.c) that cycles 100 times across barrier and modex collectives.

This commit was SVN r32203.
2014-07-10 18:53:12 +00:00
Ralph Castain
8c85ca350e Remove debug
This commit was SVN r32200.
2014-07-10 18:28:24 +00:00
Ralph Castain
7beb8f6799 Silence used-before-set var warning
This commit was SVN r32188.
2014-07-09 22:37:47 +00:00
Ralph Castain
5bb5b22573 When a user asks for cpus/rank > 1 and only has one slot, we need to ensure we always map at least one process when they don't tell us -np
cmr=v1.8.2:reviewer=rhc:subject=correct num_procs in corner case

This commit was SVN r32142.
2014-07-04 17:00:35 +00:00
Ralph Castain
9f97c74ba3 Silence warning
This commit was SVN r32136.
2014-07-03 17:29:04 +00:00
Ralph Castain
356e7ea904 Move all collective id's into the attributes and let the job pack/unpack take care of them instead of singling them out. Add the envars just prior to forking the children instead of into the launch message itself. Remove a few #if CR as the attributes functionality can handle this condition now.
This commit was SVN r32133.
2014-07-03 15:58:13 +00:00