ports/openmpi
ports/openmpi
1
1
Форкнуть 0
Код Релизы Активность
19 003 коммитов 3 Ветки 0 Теги 130 MiB
9a65e0d8c6
Граф коммитов

19003 Коммитов

This Branch
This Branch
Все ветки
Автор SHA1 Сообщение Дата
Nathan Hjelm
50b4b92758 hostname may not NULL-terminate the string if the buffer is too small. 
Thanks to Kevin M. Hildebrand for catching this.

cmr=v1.7.3:reviewer=jsquyres

This commit was SVN r29412.
 2013-10-09 15:49:18 +00:00
Jeff Squyres
2d7d7ab731 Absoft caught that bind(c) functions can't have OPTIONAL dummy arguments. 
Also, removed an MPI_Aint snuck through (which is a C type, not a Fortran
type).  Oddly, the Intel compiler complained about neither of these
issues.  :-\

This commit was SVN r29411.
 2013-10-09 14:29:34 +00:00
Jeff Squyres
dc822de80f Fix typo/misspelling in variable name. 
This commit was SVN r29410.
 2013-10-09 14:04:25 +00:00
Jeff Squyres
98d71f555c Use the correct macro: OMPI_HAVE_FORTRAN_REAL16 
Using the "#if defined(ompi_fortran_real16_t)" band-aid that I applied
in r29165 wasn't correct because if the compiler doesn't have a
fortran REAL16 type, then OMPI may well #define ompi_fortran_real16_t
to be empty, which then expands to Badness in the macro.  Hence, use
OMPI_HAVE_FORTRAN_REAL16, which will also be 0 or 1.

This commit was SVN r29408.

The following SVN revision numbers were found above:
  r29165 --> open-mpi/ompi@df7654e8cf
 2013-10-09 13:59:36 +00:00
Alex Mikheev
68b8e562d6 fixed check availability of mxm atomics. No need to run the code. succesful compilation is enough to detect availability. 
reviewed by miked

This commit was SVN r29407.
 2013-10-09 11:52:04 +00:00
Mike Dubman
a9521f3abd Add mellanox platform files 
reviewed by Amir
cmr=v1.7.4:reviewer=ompi-rm1.7

This commit was SVN r29406.
 2013-10-09 10:06:39 +00:00
Ralph Castain
9902748108 ***** THIS INCLUDES A SMALL CHANGE IN THE MPI-RTE INTERFACE ***** 
Fix two problems that surfaced when using direct launch under SLURM:

1. locally store our own data because some BTLs want to retrieve 
   it during add_procs rather than use what they have internally

2. cleanup MPI_Abort so it correctly passes the error status all
   the way down to the actual exit. When someone implemented the
   "abort_peers" API, they left out the error status. So we lost
   it at that point and *always* exited with a status of 1. This 
   forces a change to the API to include the status.

cmr:v1.7.3:reviewer=jsquyres:subject=Fix MPI_Abort and modex_recv for direct launch

This commit was SVN r29405.
 2013-10-08 18:37:59 +00:00
Jeff Squyres
7de2179866 Remove Cisco platform files from Makefile.am so we don't break "make dist" 
Thanks to Mellanox/Jenkins for catching this before the nightly build
tonight!

This commit was SVN r29402.
 2013-10-08 16:24:00 +00:00
Jeff Squyres
0d093176e1 Add missing .so version numbers for libmpi_usempif08.so. 
cmr=v1.7.3:reviewer=rhc:subject=Add missing libmpi_usempif08 shared lib

This commit was SVN r29401.
 2013-10-08 16:20:12 +00:00
Jeff Squyres
66dadbe1e7 Per RFC, remove the udapl BTL. 
This commit was SVN r29400.
 2013-10-08 15:18:59 +00:00
Ralph Castain
9389592e05 Fix --without-hwloc build 
This commit was SVN r29399.
 2013-10-08 15:02:47 +00:00
Jeff Squyres
8b1c0432f9 Remove unused files. 
This commit was SVN r29398.
 2013-10-08 14:56:43 +00:00
Jeff Squyres
a3606c4029 Add weak symbol #defines. 
Fixes compile errors if you compile --without-weak-symbols.

This commit was SVN r29396.
 2013-10-08 14:54:40 +00:00
Jeff Squyres
ad4abdbf53 Avoid using /* in the comment (which starts a C comment) 
Since this header is included in .F90 files (which are preprocessed,
vs. .f90 files, which are *not* preprocessed), we don't want to
accidentally start C-style comments, which are recognized by some
Fortran compiler preprocessors (e.g., Absoft).

This commit was SVN r29394.
 2013-10-08 14:22:55 +00:00
Jeff Squyres
ceecc04bfe Some Fortran compilers warn about ! comments after #endif. 
So just move the comment to the prior line, and it's all good.  This
is obviosuly not *necessary*, but it helps cut down on warning noise.

This commit was SVN r29393.
 2013-10-08 13:47:44 +00:00
Ralph Castain
2f9374e2b4 Update README to reflect removal of support for native Windows and reduced support for Solaris 
cmr:v1.7.3:reviewer=jsquyres

This commit was SVN r29391.
 2013-10-07 19:20:26 +00:00
Ralph Castain
6951976bc4 Update struct member name - this is why we put such things in the trunk before moving them to a branch, especially when coming from outside :-) 
Refs trac:3830

This commit was SVN r29390.

The following Trac tickets were found above:
  Ticket 3830 --> https://svn.open-mpi.org/trac/ompi/ticket/3830
 2013-10-07 15:43:43 +00:00
Ralph Castain
3d1fdf5528 Ensure we add --no-undefined to libtool for cygwin 
This commit was SVN r29389.
 2013-10-06 23:56:54 +00:00
Ralph Castain
13cd112fb4 Avoid use of interface in struct because cygwin compilers apparently object (go figure) 
This commit was SVN r29388.
 2013-10-06 23:55:38 +00:00
Ralph Castain
2d2307b6eb Modify libevent to support cygwin - patch will be pushed upstream 
This commit was SVN r29387.
 2013-10-06 23:53:31 +00:00
Mike Dubman
d2d533cf6c enable ikrit for np>=0 with mxm2 
reviewed by Amir
Refs: #3763

This commit was SVN r29386.
 2013-10-06 12:43:47 +00:00
Jeff Squyres
9f9fb5ce38 Add weak symbols for the MPI-3.1 MPI_Alloc_mem_cptr fortran subroutine. 
This subroutine was added in an errata to MPI-3.0; here's the MPI
Forum ticket about it:
    
    https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/390

This commit was SVN r29385.
 2013-10-04 22:43:40 +00:00
Jeff Squyres
c08f97b030 Fix problems with calling back-end BIND(C) ompi_*_f() functions. 
BIND(C) doesn't let us have LOGICAL parameters, so we have to be
creative in how we invoke back-end ompi_*_f() C functions.
Additionally, the mpi_f08 type for MPI_Status presented some
difficulties, too.
    
See the large comment in
ompi/mpi/fortran/use-mpi-f08/mpi-f-interfaces-bind.h that explains
this in much more detail.

This commit was SVN r29384.
 2013-10-04 22:43:07 +00:00
Jeff Squyres
f23d3bca64 Remove BIND(C) from all mpi_f08 interfaces. 
This commit was SVN r29383.
 2013-10-04 22:41:59 +00:00
Jeff Squyres
e81e3ccee0 Add missing implementation of MPI_FREE_MEM in the TKR mpi module. 
This commit was SVN r29382.
 2013-10-04 22:38:57 +00:00
Jeff Squyres
886e2cbf0f Remove and eliminate this extra redundant phrase. 
This commit was SVN r29381.
 2013-10-04 22:12:04 +00:00
Jeff Squyres
4899c531d3 Add PMPI versions for all ignore-TKR Fortran mpi module interfaces. 
This commit was SVN r29380.
 2013-10-04 21:58:56 +00:00
Rolf vandeVaart
3bd02fbaf5 Add one more verbose debug output that prints when we are out of memory. 
This commit was SVN r29378.
 2013-10-04 18:56:06 +00:00
Nathan Hjelm
b025babfec Correctly set the state for communicator requests. Refs trac:3796 
This commit was SVN r29377.

The following Trac tickets were found above:
  Ticket 3796 --> https://svn.open-mpi.org/trac/ompi/ticket/3796
 2013-10-04 17:34:54 +00:00
Nathan Hjelm
ba4b0f9235 Remove meaningless checks. Refs trac:3828 
This commit was SVN r29375.

The following Trac tickets were found above:
  Ticket 3828 --> https://svn.open-mpi.org/trac/ompi/ticket/3828
 2013-10-04 17:24:59 +00:00
Nathan Hjelm
70878764d7 Silence compiler warnings on 32-bit platforms. Refs trac:3828 
This commit was SVN r29374.

The following Trac tickets were found above:
  Ticket 3828 --> https://svn.open-mpi.org/trac/ompi/ticket/3828
 2013-10-04 15:53:14 +00:00
Rolf vandeVaart
dfd3883416 Update some CUDA-aware NEWS items. 
This commit was SVN r29372.
 2013-10-04 15:32:44 +00:00
Jeff Squyres
be191619a9 Remove erroneous use of "MPI_Aint". 
There is no "MPI_Aint" in the Fortran interface.  Surprisingly, the
Intel compiler didn't choke on this, but the Absoft compiler did.

This commit was SVN r29371.
 2013-10-04 15:27:29 +00:00
Ralph Castain
2bd2284b93 Add a useful test and update another 
This commit was SVN r29370.
 2013-10-04 15:21:40 +00:00
Jeff Squyres
587f21b18b Fix typo; thanks to Tim Mattox for noticing. 
This commit was SVN r29369.
 2013-10-04 15:08:20 +00:00
Jeff Squyres
3fb4401dee Remove an unused configure option, and comment that another 
seemingly-unused configure option is used by downstream projects.

This commit was SVN r29367.
 2013-10-04 14:16:09 +00:00
Ralph Castain
2b74e04394 We still have other projects using sensors, and we do still enable static ports for OOB 
This commit was SVN r29366.
 2013-10-04 14:11:43 +00:00
Rolf vandeVaart
66725f6973 Enable some CUDA-aware support on tcp btl. Only when configured in. 
This commit was SVN r29364.
 2013-10-04 12:50:16 +00:00
Jeff Squyres
676d48d384 Sync with changes to the README on the v1.7 branch. 
This commit was SVN r29363.
 2013-10-04 12:12:18 +00:00
Jeff Squyres
d7df5aeafe Sync with README on v1.7 branch 
This commit was SVN r29361.
 2013-10-04 11:01:50 +00:00
Jeff Squyres
88fd3bd9b3 Sync with NEWS bullets on the 1.7 branch (for 1.7.3). 
Remove ''some'' of the bullets from the "not on any branch" section;
there may be some others that are now eligible for removal, but I
wasn't 100% sure, so I left them.

This commit was SVN r29360.
 2013-10-04 10:57:05 +00:00
Ralph Castain
697fb253fa Minor modification to test code 
This commit was SVN r29357.
 2013-10-04 03:11:31 +00:00
Ralph Castain
f4f2287958 Singletons currently start out by spawning an HNP - this is required solely in the cases where the singleton subsequently calls MPI_Comm_spawn or publishes port info without support from an external orte-server. In all other cases, the HNP is of no value and can actually be a detriment by creating additional overhead on the node. This is particularly concerning for async operations where processes may begin as singletons and then dynamically wireup to perform pt2pt communications. 
So we now allow singletons to start on their own, only spawning an HNP when initiating an operation that actually requires it.

cmr:v1.7.4:reviewer=jsquyres

This commit was SVN r29354.
 2013-10-04 02:58:26 +00:00
Rolf vandeVaart
4dd1c86b36 Add a few support functions for future features. 
This commit was SVN r29353.
 2013-10-03 21:06:17 +00:00
Ralph Castain
61c4baefe0 Silence compiler warning 
Refs trac:3796

This commit was SVN r29352.

The following Trac tickets were found above:
  Ticket 3796 --> https://svn.open-mpi.org/trac/ompi/ticket/3796
 2013-10-03 19:50:10 +00:00
Mike Dubman
4c1c8d6c38 fix oshmem static linking 
fixed by Roman, reviewed by miked

This commit was SVN r29351.
 2013-10-03 15:54:09 +00:00
Alex Mikheev
7f221f7e1d oshmem & mxm2.0 yet another compilation fix 
reviewed by miked

This commit was SVN r29350.
 2013-10-03 15:50:04 +00:00
Alex Mikheev
78b860dc1d compilation fix for oshmem and mxm 
reviewed by miked

This commit was SVN r29349.
 2013-10-03 15:04:36 +00:00
Nathan Hjelm
11722457ce Fix typo in grpcomm_pmi_module.c that was giving the wrong locality for direct launched jobs. Refs trac:3824 
This commit was SVN r29348.

The following Trac tickets were found above:
  Ticket 3824 --> https://svn.open-mpi.org/trac/ompi/ticket/3824
 2013-10-03 14:38:45 +00:00
Rolf vandeVaart
ce61985503 Move registration function inside initial initialization function. 
This commit was SVN r29347.
 2013-10-03 14:14:42 +00:00