ports/openmpi
ports/openmpi
1
1
Форкнуть 0
Код Релизы Активность
17 915 коммитов 3 Ветки 0 Теги 130 MiB
55cb00f8a3
Граф коммитов

4096 Коммитов

Автор SHA1 Сообщение Дата
Mike Dubman
55cb00f8a3 Remove references to unexisting files: 
ompi/mca/common/netpatterns/
    ompi/mca/common/commpatterns/

This commit was SVN r28044.
 2013-02-12 13:21:47 +00:00
Pavel Shamis
a31bc57849 Moving mca/common/netpatterns and commpaterns to ompi/patterns. 
This commit was SVN r28035.
 2013-02-05 21:52:55 +00:00
Brian Barrett
d80218996f Rather than setting up the direct call stuff in ompi_mca (which requires 
modifying ompi_mca for every interface that is direct called), do it in
the framework's .m4 file.

This commit was SVN r28031.
 2013-02-04 23:26:42 +00:00
Vasily Filipov
21b170b43b MTL MXM: push commit r27987 back, now with right user. 
r27987 - MTL MXM: ver. 2.0 interface changes.

This commit was SVN r28026.

The following SVN revision numbers were found above:
  r27987 --> open-mpi/ompi@2735658d81
 2013-02-04 06:59:24 +00:00
Vasily Filipov
aa5e436479 Revert revesion -r27986, the reason is - it was submitted with wrong user name. 
This commit was SVN r28025.

The following SVN revision numbers were found above:
  r27986 --> open-mpi/ompi@729caaf0cd
 2013-02-04 06:54:24 +00:00
Jeff Squyres
c8dc1905f0 Fixes trac:3494: If we get 0 bytes back for the ACK, it doesn't 
necessarily mean an error -- it could (and usually does) mean that the
peer realized that we both initiated a connect at the same time, and
therefore it decided to hang up.

I also added a friendly show_help error message for other cases where
recv_blocking() fails (i.e., "Something went wrong. Kaboom! Your job
will abort...").

This commit was SVN r28023.

The following Trac tickets were found above:
  Ticket 3494 --> https://svn.open-mpi.org/trac/ompi/ticket/3494
 2013-02-02 01:19:03 +00:00
Jeff Squyres
f05b7aa6d8 As the help message states, it's not an ''error'' if the specified 
interface is not found.  It should just be skipped.

This commit was SVN r28016.
 2013-02-01 20:17:43 +00:00
Ralph Castain
afb0db5b6f Okay, Jeff - just for you...flow the show help thru the orte functions so help messages will be aggregated 
This commit was SVN r28007.
 2013-02-01 00:35:48 +00:00
Ralph Castain
e6555408f4 When we say abort, we mean ABORT!! Actually implement the ompi_rte_abort and ompi_rte_show_help functions in the ORTE module. 
This commit was SVN r28004.
 2013-01-31 23:12:11 +00:00
Igor Usarov
8d80af6c10 Support FCA v3.0 
This commit was SVN r27988.
 2013-01-31 11:14:27 +00:00
Pavel Shamis
2735658d81 MTL MXM: ver. 2.0 interface changes. 
This commit was SVN r27987.
 2013-01-31 08:38:08 +00:00
Rolf vandeVaart
729caaf0cd Remove any dependency on libcuda.so in opal layer. All changes are within OMPI_CUDA_SUPPORT code. 
This commit was SVN r27986.
 2013-01-30 23:07:32 +00:00
Rolf vandeVaart
aa04de4f1e Add run-time parameter to enable and disable CUDA GPU support. 
This commit was SVN r27970.
 2013-01-29 20:24:04 +00:00
Rolf vandeVaart
de5b7f5c6a Add mpool_base_verbose parameter. All the other base components appear to have this and it can help with debug. 
This commit was SVN r27968.
 2013-01-29 17:52:18 +00:00
Brian Barrett
49b2b5bf4f Fix double-install issue when --with-devel-headers is used 
This commit was SVN r27967.
 2013-01-29 17:23:18 +00:00
Brian Barrett
b8442ba505 Revamp the handling of wrapper compiler flags. The user flags, main configure 
flags, and mca flags are kept seperate until the very end.  The main configure
wrapper flags should now be modified by using the OPAL_WRAPPER_FLAGS_ADD
macro.  MCA components should either let <framework>_<component>_{LIBS,LDFLAGS}
be copied over OR set <framework>_<component>_WRAPPER_EXTRA_{LIBS,LDFLAGS}.
The situations in which WRAPPER CPPFLAGS can be set by MCA components was
made very small to match the one use case where it makes sense.

This commit was SVN r27950.
 2013-01-29 00:00:43 +00:00
Rolf vandeVaart
b5672927f2 Fix build issue when building with --disable-dlopen. 
This commit was SVN r27945.
 2013-01-28 20:14:59 +00:00
Rolf vandeVaart
c6412f6dff Add new rte headers in files that need them. 
This commit was SVN r27943.
 2013-01-28 19:32:33 +00:00
Pavel Shamis
1f1e1efb7b Removing leftovers of old infrastructure. 
cmr:v1.7

This commit was SVN r27942.
 2013-01-28 19:11:42 +00:00
Vishwanath Venkatesan
5be992f445 The pointer to the structure was also never allocated before retrieving 
the stripe size. Fixing that too.

This commit was SVN r27941.
 2013-01-28 07:21:22 +00:00
Vishwanath Venkatesan
817f6cd868 To remove the warning due to uninitialized variable. 
This commit was SVN r27940.
 2013-01-28 06:55:46 +00:00
George Bosilca
4defdea9f2 The shortest lifespan for a BTL. 
This commit was SVN r27939.
 2013-01-28 03:43:23 +00:00
George Bosilca
1b7dff3f2f A copy for posterity of the Open MPI Sicortex BTL. 
This commit was SVN r27938.
 2013-01-28 03:42:52 +00:00
Brian Barrett
f42783ae1a Move the RTE framework change into the trunk. With this change, all non-CR 
runtime code goes through one of the rte, dpm, or pubsub frameworks.

This commit was SVN r27934.
 2013-01-27 23:25:10 +00:00
Brian Barrett
14f4aa1198 Fix memory leak in nbc init 
This commit was SVN r27884.
 2013-01-21 22:45:59 +00:00
Brian Barrett
407714a85a Fix a memory leak in the RDMA one-sided component. Thanks to Victor Vysotskiy 
for letting us know about this one.

This commit was SVN r27883.
 2013-01-21 22:45:37 +00:00
George Bosilca
42753b4690 Make the TCP BTL really fail-safe. It now trigger the error callback on 
all pending fragments when the destination goes down. This allows the PML
to recalibrate its behavior, either find an alternate route or just give up.

This commit was SVN r27881.
 2013-01-21 11:41:08 +00:00
George Bosilca
d2281cc672 Remove the CMA related warnings. 
This commit was SVN r27872.
 2013-01-19 14:26:43 +00:00
Rolf vandeVaart
f63c88701f Improve CUDA GPU transfers over openib BTL. Use aynchronous copies. 
This is RFC that was submitted in July and December of 2012.

This commit was SVN r27862.
 2013-01-17 22:34:43 +00:00
Rolf vandeVaart
a07a4bb3f7 Update smcuda to match recent changes in sm BTL. 
This commit was SVN r27803.
 2013-01-14 14:42:19 +00:00
Rolf vandeVaart
34d1f0a585 Add some comments to the #ifdefs for clarity. No functional changes. 
This commit was SVN r27802.
 2013-01-13 16:08:48 +00:00
Alex Mikheev
344d407ed4 fixed compilation warning 
always send signalled when BTL_OPENIB_FAILOVER is defined

This commit was SVN r27801.
 2013-01-13 10:11:03 +00:00
Jeff Squyres
b2d5d1e348 Along with the Automake 1.13.x changes in r27790, rename these third 
party configure.in scripts to be configure.ac so that Automake stops
complaining about them.

This commit was SVN r27791.

The following SVN revision numbers were found above:
  r27790 --> open-mpi/ompi@675a2f5c48
 2013-01-11 20:26:19 +00:00
Jeff Squyres
675a2f5c48 Updates for Automake 1.13.x. Without these changes, Automake 1.13.x 
will error out, due to use of the
previously-deprecated-and-now-removed AM_CONFIG_HEADER macro.

This commit was SVN r27790.
 2013-01-11 20:20:02 +00:00
Samuel Gutierrez
4c28c8cbd0 New sm BTL initialization take two. This approach is pretty simple. Instead of 
using the modex or RML to share sm initialization information, have node rank 0
create a file containing initialization information in a well-known place. Then
during add_procs, the rest of the node processes requiring sm BTL initialization
will just read from that file to complete their initialization.

This commit was SVN r27789.
 2013-01-11 16:24:56 +00:00
Brian Barrett
b817166072 Use a process name instead of a name list in bcol_basesmuma 
This commit was SVN r27779.
 2013-01-09 16:43:49 +00:00
Joshua Ladd
77df51c516 Fixes the definition of the first fragment and does not assume that first frag has offset_into_user_buff equal to zero. This fix should be added to cmr:v1.7.1:reviewer=pasha 
This commit was SVN r27775.
 2013-01-08 20:24:58 +00:00
Alex Mikheev
fe672f255f request signal when sending over SRQ and number of SRQ sd_credits is 0 
This commit was SVN r27767.
 2013-01-08 14:00:29 +00:00
Samuel Gutierrez
c4acd20eb9 Backout r27739. 
This commit was SVN r27745.

The following SVN revision numbers were found above:
  r27739 --> open-mpi/ompi@a159bfaf25
 2013-01-05 01:54:23 +00:00
Nathan Hjelm
84e34ee0d7 Fix a bug in the uGNI btl that could cause certain descriptor callbacks to be called twice. 
There was a race condition in the eager get protocol where the RDMA complete message could be received before the local completion of the SMSG message that started the eager get protocol.

cmr:v1.7

This commit was SVN r27740.
 2013-01-03 23:11:13 +00:00
Samuel Gutierrez
a159bfaf25 sm BTL initialization via modex, as discussed at last year's meeting. 
This commit was SVN r27739.
 2013-01-03 21:52:20 +00:00
Mike Dubman
889d46e966 support for FCA v3.0 and up 
This commit was SVN r27731.
 2012-12-31 05:49:22 +00:00
Mike Dubman
b6d50a5733 Performance optimizations by alexm: 
* btl sendi(): if message can be send inline try to avoid signal
* signal is requested one per 64 or when
    there are no send wqes 
    when message can not be send inline 
    any other btl method then sendi()

This commit was SVN r27724.
 2012-12-26 10:19:12 +00:00
George Bosilca
ed77868984 No need for event.h in the SM BTL. 
This commit was SVN r27718.
 2012-12-23 19:33:53 +00:00
Nathan Hjelm
ef49fcea25 Remove debug printfs. 
cmr:v1.7

This commit was SVN r27680.
 2012-12-17 16:34:07 +00:00
Nathan Hjelm
ba5b2b0540 btl/vader: fix bug in single copy code that could cause ob1 sends to not get marked complete. 
cmr:v1.7

This commit was SVN r27671.
 2012-12-13 23:18:53 +00:00
Mike Dubman
a454341e2b add support for mxm 2.0 
This commit was SVN r27661.
 2012-12-09 22:58:37 +00:00
Nathan Hjelm
3e1b13b13a Re-add support for old flex (2.5.4a and earlier) while still cleaning up properly in new flex. 
This commit was SVN r27657.
 2012-12-07 00:12:43 +00:00
Brian Barrett
702451111b Remove Portals 3.3 support 
This commit was SVN r27656.
 2012-12-06 20:11:27 +00:00
Jeff Squyres
c00e6a7abf Remove the OFUD BTL. It doesn't work, and isn't included in 1.7. 
An upcoming BTL from Cisco used ofud as a starting point, and should
probably be used as a starting point for any future UD-based BTL.

And this OFUD BTL is obviously still in history if anyone ever wants
to resurrect it.

This commit was SVN r27655.
 2012-12-06 17:43:28 +00:00