1
1
Граф коммитов

6109 Коммитов

Автор SHA1 Сообщение Дата
Yael Dayan
d6f7e4eb73 openib: modified Mellanox ConnectIB max_inline_data param
This commit was SVN r27457.
2012-10-18 15:59:18 +00:00
Nathan Hjelm
d59034e6ef MCA: remove deprecated mca_base_param functions (mca_base_param_register_int, mca_base_param_register_string, mca_base_param_environ_variable). Remove all uses of deprecated functions.
cmr:v1.7

This commit was SVN r27451.
2012-10-17 20:17:37 +00:00
Vishwanath Venkatesan
6e3c9754a9 Initialization issue + fixing error in determining minimum value.
This commit was SVN r27443.
2012-10-11 23:41:25 +00:00
Vishwanath Venkatesan
95d38fdaf5 # Extracting timing information for the static collective write/read algorithms.
# The processes register their information and continue.
# Actual printing of timing information happens at file close.
# Triggered by MCA parameter at runtime

This commit was SVN r27442.
2012-10-11 21:27:47 +00:00
Vishwanath Venkatesan
240d56feeb # Extracting timing information for the two-phase collective write/read algorithms.
# The processes register their information and continue.
# Actual printing of timing information happens at file close.
# Triggered by MCA parameter at runtime

This commit was SVN r27441.
2012-10-11 21:25:30 +00:00
Vishwanath Venkatesan
7bc35f7862 # Extracting timing information for the dynamic collective write/read algorithms.
# The processes register their information and continue.
# Actual printing of timing information happens at file close.
# Triggered by MCA parameter at runtime

This commit was SVN r27440.
2012-10-11 21:23:24 +00:00
Vishwanath Venkatesan
9eeb3b5d50 # Extracting timing information for individual components of collective algorithm using a generic queue.
# This is triggered based on a mca-paramater and can be used with all collective modules.
# Individual queues maintained for read and write.
# The additional communication to combine data is done at file-close so that the 
  actual timing of collective-operations will not get affected. 
# The queues are initialized in file-open

This commit was SVN r27439.
2012-10-11 21:14:07 +00:00
Jeff Squyres
2ab48b997e Give a better verbose message if we're able to make an RC QP and we
didn't want one.

This commit was SVN r27438.
2012-10-11 19:21:21 +00:00
George Bosilca
9984a7143f Reorder the loop index.
This commit was SVN r27423.
2012-10-08 21:34:26 +00:00
George Bosilca
b46167fc4a Fix some issues with the MPI_IN_PLACE support.
This commit was SVN r27422.
2012-10-08 21:34:04 +00:00
Ralph Castain
95db4b7536 Fix typo
This commit was SVN r27417.
2012-10-05 03:56:31 +00:00
Ralph Castain
dab45fc0f6 Include missing header file in tarball
This commit was SVN r27415.
2012-10-05 01:47:05 +00:00
Vishwanath Venkatesan
c89c9e40be Code to extract neigbhouing offsets information from OMPIO into a file. Driven by an MCA parameter,
turned-off by default.

This commit was SVN r27407.
2012-10-04 21:53:26 +00:00
Vishwanath Venkatesan
c86e5f6263 set status->_ucount correctly for two-phase collective read and write operations in the module
This commit was SVN r27406.
2012-10-04 21:39:37 +00:00
Vishwanath Venkatesan
12363d50bd Setting the default stripe size to 1MB as 64KB is too small for a generic scenario.
This commit was SVN r27405.
2012-10-04 21:26:29 +00:00
Vishwanath Venkatesan
2ddbfbeb26 Setting the default stripe size for lustre to 64KB. This is an MCA parameter, can be overridden at runtime if needed.
This commit was SVN r27404.
2012-10-04 20:56:26 +00:00
Eugene Loh
25ad84b925 Ensure that MPI_Status objects have proper alignment:
- fix the Fortran layer to use new macros to convert Fortran-to-C status
- change the C internals to pull out old OMPI_SET_STATUS* macros

Also, change name of "status" argument in topo_test_f.c to "topo_type".

This commit was SVN r27403.
2012-10-04 14:39:51 +00:00
Yael Dayan
0122cf6cbb openib: added Mellanox ConnectIB device ID and params to the device parameters ini file
This commit was SVN r27402.
2012-10-04 13:20:47 +00:00
Pavel Shamis
e9df33c10b Fixing a coupe of issues in the iboffload code. It should fix the XRC compilation issue.
This commit was SVN r27400.
2012-10-03 15:44:51 +00:00
Jeff Squyres
8c369224bf More common/verbs improvements:
* Add OMPI_COMMON_VERBS_FLAGS_NOT_RC, which looks for a device	that
   does ''not'' support RC
 * Add ompi_common_verbs_find_max_inline(), and	remove that code from
   the openib BTL component

This commit was SVN r27393.
2012-10-03 00:57:39 +00:00
Jeff Squyres
82171eef9d Fixes from Ake Sandgren for MPI_SIZEOF for the TKR mpi module
implementation.  This was already fixed in the ignore TKR mpi module.

This commit was SVN r27392.
2012-10-02 16:29:06 +00:00
Nathan Hjelm
9200473a59 mtl/psm: do not let psm set processor affinity
This commit was SVN r27389.
2012-10-02 15:56:51 +00:00
Vishwanath Venkatesan
76dc8a7c55 Fix for two_phase_read_all for multiple cycles.
This commit was SVN r27388.
2012-10-01 19:14:14 +00:00
Eugene Loh
11f8dcb538 Update Mprobe/Improbe man page for MPI_PROC_NULL case and
mprobe.c/improbe.c for MPI_STATUS_IGNORE/MPI_PROC_NULL case.

This commit was SVN r27386.
2012-10-01 05:18:21 +00:00
Ralph Castain
54db4c35eb Get the trunk to build again when --without-hwloc is specified. Move a couple of key type definitions and utilities out from under the HAVE_HWLOC test so they are always available as they don't really depend on hwloc's presence. Tell two compnents not to build if hwloc is disabled:
ompi/mca/sbgp/basesmsocket
orte/mca/rmaps/lama

Remove stale configure.params files from the sbgp framework as the OMPI build system no longer looks at those files.

This commit was SVN r27377.
2012-09-26 23:24:27 +00:00
George Bosilca
48f528f142 icc complains about the missing prototype.
This commit was SVN r27373.
2012-09-26 09:56:14 +00:00
George Bosilca
890dedf13f Cleanup.
This commit was SVN r27372.
2012-09-26 09:44:46 +00:00
Vishwanath Venkatesan
c6751eaf70 Fixing all two_phase read_all bugs,
1. Multiple aggregator with non-contiguous datatype,
2. Memory corruption bugs.

Cleaned version, with proper initialization and memory management.

This commit was SVN r27370.
2012-09-26 00:16:08 +00:00
Vishwanath Venkatesan
2e2a46b2be Fixing two-phase write all bug for non-contiguous file-type.
This commit was SVN r27369.
2012-09-26 00:10:40 +00:00
Jeff Squyres
0b45d4ba86 Per discussion with Ralph and Nathan, disable UDCM for now. It's
borked and needs some surgery to get back on its feet.

This commit was SVN r27335.
2012-09-13 18:25:10 +00:00
Jeff Squyres
30d9c36275 FreeBSD detection improvement. Thanks to Brooks Davis for the patch.
This commit was SVN r27334.
2012-09-13 13:25:04 +00:00
Jeff Squyres
3cc8b0461a More updates to common verbs infrastructure:
* Moved "check basics" sanity check from openib BTL to common/verbs
   (which also allows us to have openib ''not'' include
   <infiniband/driver.h>, which is a Very Good Thing)
 * Add new ompi_common_verbs_qp_test() function, which tests to see
   whether a device supports RC and/or UD QPs.  The openib BTL now
   uses this function to ensure that the device supports RC QPs.
 * Rename ompi_common_verbs_find_ibv_ports() to be
   ompi_common_verbs_find_ports() -- the "ibv" was redundant.
 * Re-work ompi_common_verbs_find_ports() to use
   ompi_common_verbs_qp_test() instead of testing for RC/UD QPs itself
 * Add bunches of opal_output_verbose() to the find_ports() routine
   (to help diagnosing connectivity problems -- imaging running with
   --mca btl_base_verbose 10; you'll see all the find_ports() test
   results)
 * Make ompi_common_verbs_qp_test() warn if devices/ports are supplied
   in the if_include/if_exclude strings that do not exists (quite
   similar to what the openib BTL does today).
 * Add ompi_common_verbs_mca_register() function, which registers
   common verbs MCA params.  It will also register MCA param synonyms
   for thse MCA params to upper-level components (e.g.,
   btl_<upper-level-component>_<the-mca-param>). 
   * common_verbs_warn_nonexistent_if: warn if
     if_include/if_exclude-specified devices or ports do not exist.  

This commit was SVN r27332.
2012-09-12 20:47:47 +00:00
Pavel Shamis
1e7b958c2a Cleaning warning in collectives code
This commit was SVN r27331.
2012-09-12 19:47:23 +00:00
Jeff Squyres
171f6efd70 Don't free heap objects!
This commit was SVN r27326.
2012-09-12 15:11:56 +00:00
Matthias Jurenz
e643c09dee Changes to OTF:
- otfaux:
		- fixed build error on Solaris and NetBSD (removed -lm from library dependencies)
Changes to VT:
	- vtunify:
		- disable OpenMP parallelization if PGI compiler version < 9 is used (threadprivate not supported)

This commit was SVN r27316.
2012-09-12 09:03:10 +00:00
Ralph Castain
b22fc54d9b Remove unused variable
This commit was SVN r27312.
2012-09-12 08:41:48 +00:00
Jeff Squyres
a7ea880d0a Refs trac:3309
* Minor man page tweaks
 * Use existing ompi_mpi_thread_requested global

This commit was SVN r27308.

The following Trac tickets were found above:
  Ticket 3309 --> https://svn.open-mpi.org/trac/ompi/ticket/3309
2012-09-11 21:12:06 +00:00
Ralph Castain
bc1300f5cc Remove debug
This commit was SVN r27307.
2012-09-11 20:54:04 +00:00
Jeff Squyres
fb2e543a57 Refs trac:3275.
We ran into a case where the OMPI SVN trunk grew a new acceptable MCA
parameter value, but this new value was not accepted on the v1.6
branch (hwloc_base_mem_bind_failure_action -- on the trunk it accepts
the value "silent", but on the older v1.6 branch, it doesn't).  If you
set "hwloc_base_mem_bind_failure_action=silent" in the default MCA
params file and then accidentally ran with the v1.6 branch, every OMPI
executable (including ompi_info) just failed because hwloc_base_open()
would say "hey, 'silent' is not a valid value for
hwloc_base_mem_bind_failure_action!".  Kaboom.

The only problem is that it didn't give you any indication of where
this value was being set.  Quite maddening, from a user perspective.

So we changed the ompi_info handles this case.  If any framework open
function return OMPI_ERR_BAD_PARAM (either because its base MCA params
got a bad value or because one of its component register/open
functions return OMPI_ERR_BAD_PARAM), ompi_info will stop, print out
a warning that it received and error, and then dump out the parameters
that it has received so far in the framework that had a problem.

At a minimum, this will show the user the MCA param that had an error
(it's usually the last one), and ''where it was set from'' (so that
they can go fix it).  

We updated ompi_info to check for O???_ERR_BAD_PARAM from each from
the framework opens.  Also updated the doxygen docs in mca.h for this
O???_BAD_PARAM behavior.  And we noticed that mca.h had MCA_SUCCESS
and MCA_ERR_??? codes.  Why?  I think we used them in exactly one
place in the code base (mca_base_components_open.c).  So we deleted
those and just used the normal OPAL_* codes instead.

While we were doing this, we also cleaned up a little memory
management during ompi_info/orte-info/opal-info finalization.
Valgrind still reports a truckload of memory still in use at ompi_info
termination, but they mostly look to be components not freeing
memory/resources properly (and outside the scope of this fix).

This commit was SVN r27306.

The following Trac tickets were found above:
  Ticket 3275 --> https://svn.open-mpi.org/trac/ompi/ticket/3275
2012-09-11 20:47:24 +00:00
Ralph Castain
ffb8c2a2ba Add the MPI_INFO_ENV man page
This commit was SVN r27293.
2012-09-11 17:35:32 +00:00
Ralph Castain
fb4af5e29c Implement the rest of MPI-3 ticket #313 based on side-bar agreement with MPICH2 folks. Fix a bug in the original ompi_info code that put the NULL terminator one position too far if the returned string exceeded MPI_MAX_INFO_VAL in length in ompi_info_get.
This commit was SVN r27292.
2012-09-11 17:03:49 +00:00
Shiqing Fan
ec4cf39925 Windows doesn't need to exclude any interface by default. This will avoid tcp warnings.
This commit was SVN r27291.
2012-09-11 15:39:37 +00:00
Shiqing Fan
0c4c2a5f5d Revert r27283. A better solution is found. Thanks to Ralph anyway.
This commit was SVN r27290.

The following SVN revision numbers were found above:
  r27283 --> open-mpi/ompi@38bcd86ae4
2012-09-11 15:37:22 +00:00
Ralph Castain
38bcd86ae4 Per request by Shiqing, specifically exclude the "lo" interface from the TCP btl. Apparently, Windows sometimes fails to resolve the 127.0.0.1 to "lo", causing subsequent failures.
This commit was SVN r27283.
2012-09-10 16:22:46 +00:00
Matthias Jurenz
f5ffb4783c Changes to VT:
- configure: do not build OpenMP support if CFLAGS contains a compiler flag for disabling handling OpenMP directives (e.g. -fno-openmp, -nomp, -hnoomp)
	  Fixes trac:3117

This commit was SVN r27282.

The following Trac tickets were found above:
  Ticket 3117 --> https://svn.open-mpi.org/trac/ompi/ticket/3117
2012-09-10 14:29:47 +00:00
Matthias Jurenz
ef0e8f859a Changes to OTF:
- general:
		- incremented version number ro 1.11.3openmpi
	- otfaux:
		- fixed build error when using the Oracle compiler on Solaris. (removed usage of too recent rint() function)
		  Fixes trac:3257

This commit was SVN r27279.

The following Trac tickets were found above:
  Ticket 3257 --> https://svn.open-mpi.org/trac/ompi/ticket/3257
2012-09-10 12:59:07 +00:00
Jeff Squyres
b793e4ebc6 Now that Absoft has fixed the problem in their compiler, revert the
patch that helped them test (r27184).  Thanks Abosoft!

This commit was SVN r27277.

The following SVN revision numbers were found above:
  r27184 --> open-mpi/ompi@a951a5ee99
2012-09-10 12:54:24 +00:00
Vishwanath Venkatesan
6d9d0f2968 Initialize the iov_count, this crashes static write/read in certain platforms while decoding datatype
This commit was SVN r27273.
2012-09-08 00:40:21 +00:00
Jeff Squyres
dd254cc202 OMPI_HAVE_IBV_LINK_LAYER does not exist. Instead, check defined(HAVE_IBV_LINK_LAYER_ETHERNET).
This commit was SVN r27251.
2012-09-06 18:25:36 +00:00
Jeff Squyres
0d2962ebf0 Fixes trac:3294: space for the periods has already been allocated by
ompi_comm_split(), and the entire set of periods from the old
communicator have already been copied to the new communicator.  But up
here in mca_topo_base_cart_sub(), we need to subset the periods that
are actually stored on the new communicator according to remain_dims
(just like we did for the set of dimensions).

This commit renames a few variables to be a little less misleading,
and then adds a loop to copy over the periods information.  I could
have added this into the first loop (that subset-copies the
dimensions), but this code is already confusing enough and this is not
a performance-critical section: so I made it a new loop.

Note that all the topo code will be revamped a bit when the new
MPI-2.2 topo stuff (currently off in a mercurial branch) finally makes
it back to the SVN trunk.  But that new stuff will only get to v1.7 --
this commit will need to be CMR'ed to v1.6.x.

cmr:v1.7
cmr:v1.6.2

This commit was SVN r27248.

The following Trac tickets were found above:
  Ticket 3294 --> https://svn.open-mpi.org/trac/ompi/ticket/3294
2012-09-06 14:16:29 +00:00