1
1
Граф коммитов

22613 Коммитов

Автор SHA1 Сообщение Дата
Rolf vandeVaart
91a8ec52ca Fix possible unintialized warnings 2015-04-28 16:25:35 -04:00
Rolf vandeVaart
b260dc4228 Cleanup interface that handles events. No functional changes 2015-04-28 15:15:24 -04:00
Todd Kordenbrock
7fd54dcda1 Merge pull request #560 from tkordenbrock/topic/update.portals4.description
README: update the Portals4 description
2015-04-28 11:16:54 -05:00
Todd Kordenbrock
04c6bc1fc8 README: update the Portals4 description
After a simple search-replace, the Portals4 description actually
described Portals3.  This commit replaces the Portals3 description
with a Portals4 description.

Thanks to Paul Hargrove for spotting this and supplying the patch.
2015-04-28 11:12:53 -05:00
Gilles Gouaillardet
18b75bd40d io/base: check the MCA version matches 2015-04-28 17:48:23 +09:00
Rolf vandeVaart
2b99b44a16 Fix error only seen when running with CUDA 5.5 or less. Introduced with BTL 3.0 2015-04-27 17:09:53 -04:00
Jeff Squyres
876662f538 Merge pull request #554 from jsquyres/pr/verbs-fork-support
oob ud: put call to ibv_fork_init() before *all* ibv calls
2015-04-27 14:35:51 -04:00
Jeff Squyres
5114f7f37a Merge pull request #556 from jsquyres/pr/libfabric-1.0rc6
Update embedded libfabric to 1.0rc6
2015-04-27 13:40:58 -04:00
Jeff Squyres
df5043bc3f usnic: adjust for libfabric API return value change 2015-04-27 10:18:49 -07:00
Jeff Squyres
84f5c2bfb6 libfabric: update to upstream libfabric 1.0rc6 2015-04-27 10:18:49 -07:00
Jeff Squyres
60127b1caa distscript: hugely simplify
Based on initial suggestions from Paul Hargrove
(http://www.open-mpi.org/community/lists/devel/2015/04/17354.php),
simplify the distscript.sh; it really only has to do one thing:
substitute the repo_rev in VERSION.
2015-04-27 06:50:40 -07:00
Jeff Squyres
3627c82774 INSTALL: whitespace cleanup
No other changes.
2015-04-27 06:50:40 -07:00
Jeff Squyres
b6ec0348e4 INSTALL: minor refresh of text 2015-04-27 06:33:01 -07:00
Ralph Castain
3d46850c4d Per patch from Marco Atzeri, have the fortran wrapper links go directly to opal_wrapper to avoid breaks in the chain in some environments. 2015-04-25 17:09:06 -07:00
George Bosilca
424b8f170d Another uninitialized common symbol removed. 2015-04-25 14:04:56 -04:00
George Bosilca
98cca1b6b3 Remove all uninitialized common symbols in the test directory. 2015-04-25 13:54:45 -04:00
Jeff Squyres
7d870c8b9e Merge pull request #550 from jsquyres/pr/turn-down-for-WHAT
RFC: opal_config_asm.m4: enable inline assembly on little-endian POWER
2015-04-25 06:25:11 -04:00
Yohann Burette
1be185ed87 mtl/ofi: Remove use of MR. 2015-04-24 15:55:21 -07:00
Jeff Squyres
8fbf34b196 oob ud: put call to ibv_fork_init() before *all* ibv calls
Move the call to opal_common_verbs_fork_test() to up before the call
to ibv_get_device_list() (just curious -- why not use
opal_ibv_get_device_list()?).  This ensures that the call to
ibv_fork_init() is before *all* other ibv_* calls.
2015-04-24 14:19:06 -07:00
Nathan Hjelm
2716b8b1da osc/pt2pt: correct flush expected counts
Signed-off-by: Nathan Hjelm <hjelmn@me.com>
2015-04-24 13:34:21 -06:00
Nathan Hjelm
efdc1c37af Merge pull request #553 from hjelmn/osc_pt2pt_fixes
osc/pt2pt: fix bugs that caused incorrect fragment counting
2015-04-24 08:37:59 -06:00
Jeff Squyres
0afda878a2 opal_config_asm: remove support for OS X Leopard
We haven't supported OS X Leopard (10.5) for a long time.  So remove
this dead code.
2015-04-24 03:43:56 -07:00
Ralph Castain
9104e81958 When --map-by node, we should be unbound. Also remove dead code due to copy/paste error. 2015-04-23 20:35:54 -07:00
Ralph Castain
df2f002657 Update the name of the distscript.[c]sh file 2015-04-23 18:21:30 -07:00
Nathan Hjelm
f1d09e55ec osc/pt2pt: silence warnings
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-04-23 15:35:47 -06:00
Jeff Squyres
d0f4b1a755 READ: Typos/fixes from Paul Hargrove 2015-04-23 16:58:30 -04:00
Ralph Castain
5003be5c5c If the user specifies a --map-by <foo> option, then default to bind-to <foo> unless they specify a bind-to option. If they map-by slot/node, then use the default policy based on num_procs. 2015-04-23 13:30:21 -07:00
Jeff Squyres
9c1c746309 Merge pull request #552 from hjelmn/no_csh
Convert csh script distscript.csh into a POSIX sh script
2015-04-23 16:28:21 -04:00
Nathan Hjelm
b64cf39c3e Remove unnecessary quoting from variables in distscript.sh
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-04-23 13:27:31 -06:00
Nathan Hjelm
29b435a5a4 osc/pt2pt: fix bugs that caused incorrect fragment counting
This commit fixes a bug identified by MTT that occurred when mixing
passive and active target synchronization. The bugs fixed in this
commit are:

 - Do not update incoming fragment counts for any type of unbuffered
   control message. These messages are out-of-band and should not be
   considered towards the signal counts.

 - Complete a change from using received counts to expected counts for
   lock, unlock, and flush acks. Part of the change made it into
   master before the rest was ready. This was preventing wakeups in
   some cases.

 - Turn the passive_target_access_epoch module member into a
   counter. As long as at least one peer is locked we are in a
   passive-target epoch and not an active target one. This fix will
   ensure that fragment flags are set appropriately.

fixes #538

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-04-23 13:22:24 -06:00
Nathan Hjelm
deb1542d0d Convert csh script distscript.csh into a POSIX sh script
This commit removes the only cshell script in the Open MPI
project.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-04-23 11:36:12 -06:00
Ralph Castain
d5e4fd059f Ensure the binding and locale strings are always defined 2015-04-23 07:43:37 -07:00
Ralph Castain
cb7330a543 Get the output to lineup properly 2015-04-23 07:38:51 -07:00
rhc54
9258919c3b Merge pull request #551 from jsquyres/pr/minor-display-devel-map-output-tweak
display-devel-map: minor output tweak
2015-04-23 07:37:17 -07:00
Jeff Squyres
79243aca4e display-devel-map: minor output tweak
hwloc output can get fairly long, especially on machines with lots of
cores and/or hyperthreads.  So put the Locale and Binding output on
separate lines.
2015-04-23 06:14:57 -07:00
Jeff Squyres
f7b8c11027 opal_config_asm.m4: enable inline assembly on little endian POWER
Suggested by Paul Hargrove,
http://www.open-mpi.org/community/lists/devel/2015/04/17323.php,
amended by Nysal Jan for 32 bit, too.
2015-04-23 05:43:24 -07:00
Ralph Castain
58e646ccfd Reduce confusion by having the devel-map display in the same format as report-bindings 2015-04-23 04:30:00 -07:00
Jeff Squyres
7da50433f7 opal_check_package: make the var name match
The cache variable now only needs the function name, not the lib name.
2015-04-23 03:22:38 -07:00
Jeff Squyres
3481d4ef1f README: update --with-libltdl docs 2015-04-22 15:00:29 -07:00
Jeff Squyres
63c7520273 use-mpi-f08/Makefile.am: also link in libmpi_mpifh.la
Per mail from Macro Atzeri, we also need to link in libmpi_mpifh.la,
lest we exhaust relative offset addressing (e.g., in 32 bit builds).

See http://www.open-mpi.org/community/lists/devel/2015/04/17304.php.
2015-04-22 14:22:36 -07:00
Ryan Grant
1436417488 Merge pull request #547 from tkordenbrock/topic/mtl.add.logical.mode
mtl-portals4: add the option to use the Portals4 logical to physical mapping
2015-04-22 15:06:13 -06:00
Jeff Squyres
4b8fa24682 configure.ac: fix the flex error message
Per Paul Hargrove's email
(http://www.open-mpi.org/community/lists/devel/2015/04/17310.php),
there were a few typos in the message displayed when flex is not
found.  Additionally, somehow it didn't show the list of *.l files on
his system.  This is not a super-important message -- so just take
that "find" out (and force developers to install flex).
2015-04-22 16:59:59 -04:00
Jeff Squyres
46aa20a919 timer_linux_component.c: fix the operator precedence
Paul Hargrove noted in
http://www.open-mpi.org/community/lists/devel/2015/04/17312.php that
we need () to ensure that we get the correct mathemtical result.
2015-04-22 13:40:09 -07:00
Nathan Hjelm
b06b8584f7 btl/openib/udcm: OBJ_CONSTUCT UDCM module member before attempting to initialize
This commit fixes an assert when trying to cleanup a module we failed
to initialize. There is no protection around the OBJ_DESTRUCT calls so
they will always be called so similarly we should always call
OBJ_CONSTRUCT at init.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-04-22 13:25:12 -06:00
Ryan Grant
9ccb932880 Merge pull request #546 from tkordenbrock/topic/btl.logical.mode.init.patch
btl-portals4: delay interface initialization until add_procs()
2015-04-22 11:22:46 -06:00
Rolf vandeVaart
030dddde23 Remove unneeded copy and minor clean up 2015-04-22 13:17:16 -04:00
Ryan Grant
0e528be2ac Merge pull request #545 from tkordenbrock/topic/btl30.large.message.patch
btl-portals4: fix bugs in the Portals4 BTL after the port to BTL3 interface
2015-04-22 11:05:35 -06:00
Nathan Hjelm
47ef12b15e Merge pull request #549 from hjelmn/topo_fix
mca/base: provide functions to determine if a framework is registered/open
2015-04-21 14:32:20 -06:00
Nathan Hjelm
7c95ecf859 mca/base: provide functions to determine if a framework is registered/open
This commit also fixes a problem with the lazy opening of topo
components. The topo framework incorrectly: 1) checked if the topo
framework was open by checking the length of the components list, and
2) called the framework open directly instead of using
mca_base_framework_open.

fixes #544

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-04-21 13:54:25 -06:00
Todd Kordenbrock
8e56002ec7 mtl-portals4: add missing return to portals4_init_interface() 2015-04-21 11:30:33 -05:00