ports/openmpi
ports/openmpi
1
1
Форкнуть 0
Код Релизы Активность
22 676 коммитов 3 Ветки 0 Теги
Граф коммитов

546 Коммитов

Автор SHA1 Сообщение Дата
Ralph Castain
f55c587a74 Per patch from Tetsuya Mishima, ensure the rank_file mapper accurately tracks number of nodes in the map 
Refs trac:4594

This commit was SVN r31725.

The following Trac tickets were found above:
  Ticket 4594 --> https://svn.open-mpi.org/trac/ompi/ticket/4594
 2014-05-13 14:36:25 +00:00
Ralph Castain
5602156a1c Use the correct abstraction layer name for the data dirs 
This commit was SVN r31684.
 2014-05-08 14:32:24 +00:00
Ralph Castain
11faab1091 The final step of the RFC: convert the <foo>libdir and friends to fit their respective code areas, and equate them all at the top. Note that we can't entirely separate things as the opal_install_dirs framework can't handle separated locations for the various trees. 
This commit was SVN r31679.
 2014-05-08 02:01:35 +00:00
Ralph Castain
6545e6e9a8 Add one more check for failed mapping that rarely occurs, but results in a hang when it does 
cmr=v1.8.2:reviewer=rhc

This commit was SVN r31598.
 2014-05-02 10:35:14 +00:00
Ralph Castain
61d94fcee2 Fix the sequential mapper - it was out-of-sync with the hostfile changes, and we missed the "seq" policy when parsing the --map-by option. Thanks to Bill Chen for reporting it 
cmr=v1.8.1:reviewer=jsquyres

This commit was SVN r31333.
 2014-04-08 03:38:25 +00:00
Jeff Squyres
82e104719a hwloc/rmaps base: Add missing help message. 
Also, add missing ORTE_ERROR_LOG in the other case where this error
message is used (i.e., ORTE_ERROR_LOG was used in the one place, so
let's also use it in the other place).

This commit was SVN r31321.
 2014-04-07 15:39:54 +00:00
Ralph Castain
3fdcaeab97 Fix a problem where we need to abort due to a mapping failure, but we are in a managed environment and thus the orteds have not wired up. Thus, if we send the exit message across the routed network, the remote daemons won't have a way to relay the message along - and we won't exit. 
If we are aborting, then set the flags so the HNP directly sends an exit command to each daemon. Make it the halt_vm command so the remote daemon doesn't try to relay it, but instead just exits without waiting for its routed children to exit first.

cmr=v1.8.1:reviewer=jsquyres:subject=fix hangs due to abort prior to daemon wireup

This commit was SVN r31304.
 2014-04-02 04:17:55 +00:00
Ralph Castain
714cb8f573 Silence warnings 
cmr=v1.8:reviewer=rhc

This commit was SVN r31248.
 2014-03-27 14:16:54 +00:00
Ralph Castain
390645ac2a Per patch from Tetsuya Mishima, do a nicer job of warning the user that we need to map to a higher level to get the number of requested cpus/rank. Also, change the mapping policy to "byslot" when falling back to that option. 
cmr=v1.8:reviewer=rhc

This commit was SVN r31196.
 2014-03-24 15:47:29 +00:00
Ralph Castain
081669b440 When pretty-printing binding info, we need to pass the topology down to the routine as the mapper isn't always working with the local topology - otherwise, we get an erroneous help message. Thanks to Tetsuya Mishima for reporting it 
cmr=v1.7.5:reviewer=rhc:subject=fix pretty-print of bindings

This commit was SVN r30968.
 2014-03-10 15:53:07 +00:00
Ralph Castain
fc2dd6ac48 Per Jeff's request, add a more detailed comment as to why we are turning off the warning at this time. 
Refs trac:4339

This commit was SVN r30948.

The following Trac tickets were found above:
  Ticket 4339 --> https://svn.open-mpi.org/trac/ompi/ticket/4339
 2014-03-06 02:17:25 +00:00
Ralph Castain
a2b539c763 Per the telecon, silence the warning for 1.7.5 to give us time to consider a better permanent solution 
Refs trac:4339

This commit was SVN r30941.

The following Trac tickets were found above:
  Ticket 4339 --> https://svn.open-mpi.org/trac/ompi/ticket/4339
 2014-03-05 03:02:29 +00:00
Ralph Castain
50c30d62ca Repair builds without hwloc 
cmr=v1.7.5:reviewer=jsquyres

This commit was SVN r30940.
 2014-03-05 02:48:15 +00:00
Ralph Castain
0ac97761cc Now that we are binding by default, the issue of #slots and what to do when oversubscribed has become a bit more complicated. This isn't a problem in managed environments as we are always provided an accurate assignment for the #slots, or when -host is used to define the allocation since we automatically assume one slot for every time a node is named. 
The problem arises when a hostfile is used, and the user provides host names without specifying the slots= paramater. In these cases, we assign slots=1, but automatically allow oversubscription since that number isn't confirmed. We then provide a separate parameter by which the user can direct that we assign the number of slots based on the sensed hardware - e.g., by telling us to set the #slots equal to the #cores on each node. However, this has been set to "off" by default.

In order to make this a little less complex for the user, set the default such that we automatically set #slots equal to #cores (or #hwt's if use_hwthreads_as_cpus has been set) only for those cases where the user provides names in a hostfile but does not provide slot information.

Also cleanup some a couple of issues in the mapping/binding system:

* ensure we only override the binding directive if we are oversubscribed *and* overload is not allowed

* ensure that the MPI procs don't attempt to bind themselves if they are launched by an orted as any binding directive (no matter what it was) would have been serviced by the orted on launch

* minor cleanup to the warning message when oversubscribed and binding was requested

cmr=v1.7.5:reviewer=rhc:subject=update mapping/binding system

This commit was SVN r30909.
 2014-03-03 16:46:37 +00:00
Ralph Castain
88b0e0cc6d Allow the user to turn off the oversubscribed-binding warning if overload-allowed has been provided 
Refs trac:4317

This commit was SVN r30892.

The following Trac tickets were found above:
  Ticket 4317 --> https://svn.open-mpi.org/trac/ompi/ticket/4317
 2014-02-28 17:55:53 +00:00
Ralph Castain
4a645f0342 Add detection of oversubscription with binding requested - if binding requested to core or hwt, warn and do not bind or else we will hurt performance. Also, if no binding directive was given, turn off the default binding 
Refs trac:4317

This commit was SVN r30888.

The following Trac tickets were found above:
  Ticket 4317 --> https://svn.open-mpi.org/trac/ompi/ticket/4317
 2014-02-28 16:08:52 +00:00
Ralph Castain
8500247c7b Fix the by-obj mapper in the case where slots are not specified, and so we are in a perpetual oversubscribed state 
cmr=v1.7.5:reviewer=rhc

This commit was SVN r30887.
 2014-02-28 05:21:46 +00:00
Ralph Castain
a4c3d0a5a0 Add some more debug to the by-obj mapper 
This commit was SVN r30884.
 2014-02-28 02:52:53 +00:00
Ralph Castain
d109c523b9 Per patch from Tetsuya Mishima, complete the overhaul of the round-robin mappers 
Refs trac:4296

This commit was SVN r30861.

The following Trac tickets were found above:
  Ticket 4296 --> https://svn.open-mpi.org/trac/ompi/ticket/4296
 2014-02-27 00:43:53 +00:00
Ralph Castain
61a21e4f31 Based on Tetsuya's patch, with some changes, correct the case of map-by node where multiple cpus/rank are requested and result in a non-integer match with num slots. Also correct tests for binding policy given to use the proper macro. 
Refs trac:4296

This commit was SVN r30857.

The following Trac tickets were found above:
  Ticket 4296 --> https://svn.open-mpi.org/trac/ompi/ticket/4296
 2014-02-26 18:12:23 +00:00
Ralph Castain
b880aa46bd Update the map-by obj and map-by obj:span mappers to correct for errors in computing carryover across the nodes. Be a little less complex in the algorithm so it is easier to follow and debug. 
Refs trac:4296

This commit was SVN r30826.

The following Trac tickets were found above:
  Ticket 4296 --> https://svn.open-mpi.org/trac/ompi/ticket/4296
 2014-02-25 23:32:43 +00:00
Ralph Castain
c8112c1086 Loadbalancing across nodes (i.e., map-by node) wasn't working correctly - the algorithm relied on the nodes being defined in descending order of slots, or the numbe 
r of slots remaing to be assigned being only one/node. Regardless, it didn't work for the case where nodes were defined in ascending order of slots.

Tetsuya's proposed patch didn't solve the problem for me, but it did correct the case where cpus/proc > 1. The final patch requires that we loop over the assignment
 algo until all procs are assigned or all nodes are filled - any remaining procs are then handled in the cleanup loop.

cmr=v1.7.5:reviewer=rhc:subject=fix map-by node for different cases

This commit was SVN r30798.
 2014-02-22 16:39:41 +00:00
Mike Dubman
8d4592a94b rmaps/mindist: better error message 
better error message when there is only one socket available

fixed by Elena, reviewed by Miked
cmr=v1.7.5:reviewer=ompi-rm1.7

This commit was SVN r30787.
 2014-02-21 11:38:35 +00:00
Ralph Castain
91f90058ce Add missing options and cleanup the code a bit. Default to by-slot ranking if a non-hardware option isn't given. Thanks to Tetsuya Mishima for the assist. 
cmr=v1.7.5:reviewer=ompi-gk1.7

This commit was SVN r30725.
 2014-02-14 10:23:16 +00:00
Ralph Castain
fd9b301a8b Check equality instead of bit-mask - thanks to Tetsuya Mishima for reporting it 
cmr=v1.7.5:reviewer=ompi-gk1.7

This commit was SVN r30722.
 2014-02-14 02:34:42 +00:00
Ralph Castain
1473dde6ea Okay, once again be caught by the blasted hwloc inability to cleanly handle caches. Protect the calls to get_depth by first checking to see if it is a "cache", then use a cache-specific function to get the stupid data. Very, very irritating. 
cmr=v1.7.5:reviewer=jsquyres:subject=treat caches as something different yet again

This commit was SVN r30693.
 2014-02-12 01:45:06 +00:00
Ralph Castain
b566cd5e30 Protect against no modifiers 
Refs trac:4117

This commit was SVN r30672.

The following Trac tickets were found above:
  Ticket 4117 --> https://svn.open-mpi.org/trac/ompi/ticket/4117
 2014-02-11 17:34:37 +00:00
Ralph Castain
6fa34407bf Handle modifiers to the --map-by dist option 
Refs trac:4117

This commit was SVN r30671.

The following Trac tickets were found above:
  Ticket 4117 --> https://svn.open-mpi.org/trac/ompi/ticket/4117
 2014-02-11 17:19:05 +00:00
Ralph Castain
4781ea71b6 Correct the handling of various map/bind combinations when pe=N is given. Thanks to Elena Elkina for reporting it. 
Refs trac:4117

This commit was SVN r30663.

The following Trac tickets were found above:
  Ticket 4117 --> https://svn.open-mpi.org/trac/ompi/ticket/4117
 2014-02-11 03:05:26 +00:00
Ralph Castain
707e51d786 Check for --cpus-per-proc earlier, before the correct option can be processed. Thanks to Tetsuya Mishima for reporting it. 
Refs trac:4117

This commit was SVN r30662.

The following Trac tickets were found above:
  Ticket 4117 --> https://svn.open-mpi.org/trac/ompi/ticket/4117
 2014-02-11 02:53:53 +00:00
Ralph Castain
d66d2f5fb3 It is just fine to map by node or slot and bind, so ensure the switch statement includes those options. Thanks to Tatsuya Mishima for point it out. 
Refs trac:4240

This commit was SVN r30661.

The following Trac tickets were found above:
  Ticket 4240 --> https://svn.open-mpi.org/trac/ompi/ticket/4240
 2014-02-11 02:52:01 +00:00
Ralph Castain
1a12325094 Rats - need to include bydist in the mapping list 
Refs trac:4117

This commit was SVN r30649.

The following Trac tickets were found above:
  Ticket 4117 --> https://svn.open-mpi.org/trac/ompi/ticket/4117
 2014-02-09 16:17:05 +00:00
Ralph Castain
ca0c806662 Resolve the problem of binding in inverted topologies - check the relative depth of the map and bind objects in the topology, and let that determine whether we bind downward or upwards. 
cmr=v1.7.5:reviewer=jsquyres:subject=Resolve the problem of binding in inverted topologies

This commit was SVN r30643.
 2014-02-09 05:30:17 +00:00
Ralph Castain
bc7cc09749 After a lot of pain, I've managed to resolve the problem of conflicting mapping directives caused by mismatched MCA params - i.e., where someone has one variant of an MCA param (e.g., rmaps_base_mapping_policy) in their default MCA param file, and then specifies another variant (e.g., --npernode) on the command line. I can't fully resolve the problem as there is no way to know precisely what the user meant - we can only guess which param was really intended since the MCA param system 
can't apply its normal precedence rules.

So...print a big "deprecated" warning for the old params and error out if a conflict is detected. I know that isn't what people really wanted, but it's the best we
 can do. If only the old style param is given, then process it after the warning.

Extend the current map-by param to add support for ppr and cpus-per-proc, adding the latter to the list of allowed modifiers using "pe=n" for processing elements/proc. Thus, you can map-by socket:pe=2,oversubscribe to map by socket, binding 2 processing elements/process, with oversubscription allowed. Or you can map-by ppr:2:socket:pe=4 to map two processes to every socket in the allocation, binding each process to 4 processing elements.

For those wondering, a processing element is defined as a hwthread if --use-hwthreads-as-cpus is given, or else as a core.

Refs trac:4117

This commit was SVN r30620.

The following Trac tickets were found above:
  Ticket 4117 --> https://svn.open-mpi.org/trac/ompi/ticket/4117
 2014-02-07 21:25:40 +00:00
Ralph Castain
e43589ed84 Fix warning - thanks to Paul Hargrove for reporting it 
cmr=v1.7.4:reviewer=ompi-gk1.7

This commit was SVN r30548.
 2014-02-03 23:51:45 +00:00
Ralph Castain
410a3afa7b Fix --without-hwloc operations - must default to map-by slot in that scenario 
cmr=v1.7.4:reviewer=jsquyres

This commit was SVN r30474.
 2014-01-29 16:54:05 +00:00
Ralph Castain
42eb0bbe1b Fix --without-hwloc builds 
cmr=v1.7.4:reviewer=jsquyres

This commit was SVN r30462.
 2014-01-28 17:10:32 +00:00
Ralph Castain
84a0ab3a75 Ah @$#!$#% - missed one last help message that needs to be corrected. 
cmr=v1.7.4:reviewer=jsquyres:subject=correct help message

This commit was SVN r30449.
 2014-01-28 04:03:24 +00:00
Ralph Castain
941bfd4604 Final cleanup of cpus-per-proc for 1.7.4 - provide better checking for cpus-per-proc and mismatched mapping/binding directives, and provide error messages telling the user what to do to get it right. 
cmr=v1.7.4:reviewer=jsquyres

This commit was SVN r30438.
 2014-01-27 22:40:51 +00:00
Ralph Castain
886fee9367 Properly set num_procs when np is not given, but cpus-per-proc is used. Thanks to Tetsuya Mishima for pointing it out 
cmr=v1.7.4:reviewer=jsquyres

This commit was SVN r30389.
 2014-01-23 05:01:07 +00:00
Jeff Squyres
7768828d2d Addendum to r30298: tweak the wording of the help messages a bit. 
Refs trac:4117.  Please use this commit rather than the patch attached to
the ticket; the patch had a few mistakes in the tweaked wording.

This commit was SVN r30362.

The following SVN revision numbers were found above:
  r30298 --> open-mpi/ompi@58479399c3

The following Trac tickets were found above:
  Ticket 4117 --> https://svn.open-mpi.org/trac/ompi/ticket/4117
 2014-01-22 12:17:14 +00:00
Ralph Castain
58479399c3 As per RFC and telecon, deprecate cmd line options and their corresponding MCA params for old-style mapping and binding directives 
cmr=v1.7.5:reviewer=jsquyres:subject=deprecate old-style mapping and binding directives

This commit was SVN r30298.
 2014-01-15 14:48:39 +00:00
Ralph Castain
fb9e427320 One last corner case - when encountering an overload condition (e.g., by comm_spawning more procs than we have cores) and we are using the default binding policy, do *not* bind the new procs to anything as this can cause major problems. Instead, let the spawn succeed since the user didn't specifically ask to be bound, and leave the new procs as unbound. 
Refs trac:4077

This commit was SVN r30200.

The following Trac tickets were found above:
  Ticket 4077 --> https://svn.open-mpi.org/trac/ompi/ticket/4077
 2014-01-09 22:39:34 +00:00
Ralph Castain
24e990e747 Fix comm_spawn for oversubscribed systems by correctly computing the number of available slots 
cmr=v1.7.4:reviewer=jsquyres:subject=Fix comm_spawn for oversubscribed systems

This commit was SVN r30197.
 2014-01-09 20:33:48 +00:00
Ralph Castain
9fcb46d85a Correctly detect and handle oversubscription for comm_spawn 
cmr=v1.7.4:reviewer=jsquyres:subject=Correctly detect and handle oversubscription for comm_spawn

This commit was SVN r30186.
 2014-01-09 18:27:51 +00:00
Ralph Castain
6e5fedeb04 Oops - add verbose output to inform that cannot default bind due to no cores detected 
Refs trac:4074

This commit was SVN r30185.

The following Trac tickets were found above:
  Ticket 4074 --> https://svn.open-mpi.org/trac/ompi/ticket/4074
 2014-01-09 18:17:14 +00:00
Ralph Castain
7e4748a0f1 Handle the case of nodes that do not report cores, and thus our default binding policy will fail even though binding is supported by defaulting to not binding on those nodes. 
Thanks to Paul Hargrove for reporting the problem on NetBSD.

cmr=v1.7.4:reviewer=jsquyres:subject=Handle the case of nodes that do not report cores

This commit was SVN r30180.
 2014-01-09 16:27:58 +00:00
Ralph Castain
bf453a2575 Reference the correct variable...sigh 
Refs trac:4059

This commit was SVN r30163.

The following Trac tickets were found above:
  Ticket 4059 --> https://svn.open-mpi.org/trac/ompi/ticket/4059
 2014-01-08 22:36:39 +00:00
Ralph Castain
e724d0d12d Ensure comm_spawn'd jobs get treated the same wrt setting default mapping directives 
Refs trac:4059

This commit was SVN r30158.

The following Trac tickets were found above:
  Ticket 4059 --> https://svn.open-mpi.org/trac/ompi/ticket/4059
 2014-01-08 15:16:22 +00:00
Ralph Castain
fb650aed0c Fix how we transfer mapping directives to the job, ensuring that directives that can be given outside of a mapping policy (e.g., oversubscribe and no-use-local) are retained. 
cmr=v1.7.4:reviewer=jsquyres:subject=Fix how we transfer mapping directives to the job

This commit was SVN r30155.
 2014-01-08 04:25:43 +00:00