1
1
Граф коммитов

2715 Коммитов

Автор SHA1 Сообщение Дата
George Bosilca
50b26ebb6a Allow the ompi_ddt_init and ompi_ddt_finalize to be visible even when
the visibility feature is on.

This commit was SVN r14726.
2007-05-23 14:02:08 +00:00
Sven Stork
88f0845c44 - let the pt2pt component compile with threads enabled
This commit was SVN r14725.
2007-05-23 12:56:34 +00:00
Brian Barrett
38eab3613b * Fix race condition with the pending_{in,out} variables -- if we're going
to do while(...) { } then we can't change the variables in the ... 
    atomically, but should do it while holding the module lock.
  * Fix dumb communicator creation error when we don't create the progress
    stuff (because a window already exists), where we would accidently
    jump to the error case.

This commit was SVN r14715.
2007-05-21 20:53:02 +00:00
Ralph Castain
4fff584a68 Commit the orted-failed-to-start code. This correctly causes the system to detect the failure of an orted to start and allows the system to terminate all procs/orteds that *did* start.
The primary change that underlies all this is in the OOB. Specifically, the problem in the code until now has been that the OOB attempts to resolve an address when we call the "send" to an unknown recipient. The OOB would then wait forever if that recipient never actually started (and hence, never reported back its OOB contact info). In the case of an orted that failed to start, we would correctly detect that the orted hadn't started, but then we would attempt to order all orteds (including the one that failed to start) to die. This would cause the OOB to "hang" the system.

Unfortunately, revising how the OOB resolves addresses introduced a number of additional problems. Specifically, and most troublesome, was the fact that comm_spawn involved the immediate transmission of the rendezvous point from parent-to-child after the child was spawned. The current code used the OOB address resolution as a "barrier" - basically, the parent would attempt to send the info to the child, and then "hold" there until the child's contact info had arrived (meaning the child had started) and the send could be completed.

Note that this also caused comm_spawn to "hang" the entire system if the child never started... The app-failed-to-start helped improve that behavior - this code provides additional relief.

With this change, the OOB will return an ADDRESSEE_UNKNOWN error if you attempt to send to a recipient whose contact info isn't already in the OOB's hash tables. To resolve comm_spawn issues, we also now force the cross-sharing of connection info between parent and child jobs during spawn.

Finally, to aid in setting triggers to the right values, we introduce the "arith" API for the GPR. This function allows you to atomically change the value in a registry location (either divide, multiply, add, or subtract) by the provided operand. It is equivalent to first fetching the value using a "get", then modifying it, and then putting the result back into the registry via a "put".

This commit was SVN r14711.
2007-05-21 18:31:28 +00:00
Brian Barrett
0e9e0c518a Fix a couple more progress thread related issues...
This commit was SVN r14708.
2007-05-21 16:06:14 +00:00
Pavel Shamis
5ceaa605d7 Adding new vendor_part_id for Mellanox Hermon HCA
This commit was SVN r14705.
2007-05-21 13:33:54 +00:00
Brian Barrett
1191677b76 Fix dumb threads-related compile issues
This commit was SVN r14704.
2007-05-21 03:23:58 +00:00
Brian Barrett
2b4b754925 Some much needed cleanup of the point-to-point one-sided component...
* Combine polling of the long requests and buffer requests into
    one type, and in one place
  * Associate the list of requests to poll with the component, not
    the individual modules
  * add progress thread that sits on the OMPI request structure
    and wakes up at the appropriate time to poll the message
    list.  Not the best, but without some asynch notification
    from the PML that a given set of requests has completed, there
    isn't much better
  * Instead of calling opal_progress() all over the place, move
    to using the condition variables like the rest of the project.
    Has the advantage of moving it slightly futher along in the
    becoming thread safe thing
  * Fix a problem with the passive side of unlock where it could
    go recursive and cause all kinds of problems, especially
    when progress threads are used.  Instead, have two parts of
    passive unlock -- one to start the unlock, and another to
    complete the lock and send the ack back.  The data moving
    code trips the second at the right time.

This commit was SVN r14703.
2007-05-21 02:21:25 +00:00
Ralph Castain
fa5a40070d Test the return status code from comm_dyn_start_processes - if we see an error, then let's report it and not continue on with the comm_spawn procedure!
This commit was SVN r14699.
2007-05-18 20:22:32 +00:00
Donald Kerr
23280bd7da remove an assignment which is not required
This commit was SVN r14692.
2007-05-18 01:33:02 +00:00
Donald Kerr
588d5bd6a9 clean up compile warnings
This commit was SVN r14691.
2007-05-17 23:37:47 +00:00
George Bosilca
7738079ab9 Remove unused variable.
This commit was SVN r14689.
2007-05-17 20:01:30 +00:00
Gleb Natapov
b2c8fcdbab Forget to add file in r14681.
This commit was SVN r14682.

The following SVN revision numbers were found above:
  r14681 --> open-mpi/ompi@3ebaff8dfe
2007-05-17 08:41:01 +00:00
Gleb Natapov
3ebaff8dfe Implement new BTL parameters:
We eagerly send data up to btl_*_eager_limit with the match
Upon ACK of the MATCH we start using send/receives of size
btl_*_max_send_size up to the btl_*_rdma_pipeline_offset
After the btl_*_rdma_pipeline_offset we begin using RDMA writes of
size btl_*_rdma_pipeline_frag_size.

Now, on a per message basis we only use the above protocol if the
message is larger than btl_*_min_rdma_pipeline_size

btl_*_eager_limit - > same
btl_*_max_send_size -> same
btl_*_rdma_pipeline_offset -> btl_*_min_rdma_size
btl_*_rdma_pipeline_frag_size -> btl_*_max_rdma_size


btl_*_min_rdma_pipeline_size is new..

This patch also moves all BTL common parameters initialisation into
btl_base_mca.c file.

This commit was SVN r14681.
2007-05-17 07:54:27 +00:00
Brian Barrett
33a5758521 Some IPv6 improvements:
* Move ipv6comat.h code into opal_config_bottom.h and change into some
    more intelligent testing of structures
  * Change opal's if interface to use sockaddr instead of sockaddr_storage,
    as the RFCs suggest we do
  * Move the networking code in opal that isn't directly related to if
    detection into net.h
  * Add quicky function to get the port out of either a sockaddr_in
    or sockaddr_in6, saving a bunch of code in the oob.
  * Update TCP oob and btl with new interface

This commit was SVN r14679.
2007-05-17 01:17:59 +00:00
Donald Kerr
c40307fd27 add user warning message to inform when udapl btl is no longer able to register memory
This commit was SVN r14678.
2007-05-16 21:04:50 +00:00
Brian Barrett
7708c4f887 Don't complain about unsupported protocols. Needs to be made better,
but this will quit the whining from platforms where the kernel doesn't
have IPv6 support.

This commit was SVN r14676.
2007-05-16 20:11:47 +00:00
Sven Stork
22af6d38e6 - UNexport symbols that shouldn't be needed outside the libraries
- replace #if/#endif with BEGIN/END_C_DECLS
- reformating

This commit was SVN r14669.
2007-05-16 15:46:52 +00:00
Sven Stork
bd29eb9bd1 - backout commit r14667, because internal functionality shouldn't be exported.
NOTE: if visibility is enabled "make check" will fail

This commit was SVN r14668.

The following SVN revision numbers were found above:
  r14667 --> open-mpi/ompi@1f526a95e9
2007-05-16 15:43:44 +00:00
Sven Stork
1f526a95e9 - we need to export this internal symbols because the tests in
test/memory need them.

This commit was SVN r14667.
2007-05-16 15:14:31 +00:00
Gleb Natapov
61e889a1d9 Fix breakage of GM by r13921. On receive GM provides only buffer pointer
without any context so we need to save a context somewhere so it can be
retrieved given only buffer pointer. This patch saves context (pointer to
frag) just before start of a buffer so it can be be easily retrieved. 

This commit was SVN r14664.

The following SVN revision numbers were found above:
  r13921 --> open-mpi/ompi@90fb58de4f
2007-05-16 12:20:58 +00:00
Donald Kerr
2ed72bf2e2 break evd_qlen into individual qlens (async,dto,conn); add checks based on udapl limits and number of peers
This commit was SVN r14659.
2007-05-15 17:47:00 +00:00
Pavel Shamis
cd87b05711 Added check for IBV_EVENT_CLIENT_REREGISTER async
event that was not exists in old openib gen2 versions
(Ticket #1025)

This commit was SVN r14658.
2007-05-15 13:53:49 +00:00
Sven Stork
91fa494f0e - another missing symbol
This commit was SVN r14657.
2007-05-15 13:38:50 +00:00
Sven Stork
18a5747799 - this symbol is (at least) used by the basic collective component
This commit was SVN r14654.
2007-05-15 12:48:58 +00:00
Brian Barrett
21e00f6f0c Clean up a couple of configure things:
* Require Autoconf 2.60 or higher and remove some cruft
    required for AC 2.59 or the AC 2.59 / AC 2.60 mix
  * Remove a bunch of now unnecessary AC_SUBST calls
  * Use the libtool-provided variables for the -I and
    library to use when compiling against ltdl

Fixes trac:1000

This commit was SVN r14652.

The following Trac tickets were found above:
  Ticket 1000 --> https://svn.open-mpi.org/trac/ompi/ticket/1000
2007-05-15 04:23:48 +00:00
Jeff Squyres
92090967b1 Add definitions for Hemon/ConnectX Mellanox HCA
This commit was SVN r14639.
2007-05-10 12:27:51 +00:00
Donald Kerr
436d370d51 latency improvements: use ompi_free_list_init_ex, create optimal alignment parameter, remove rdma guarantee path, replace dat_lmt_sync_rdma with use of volatile
This commit was SVN r14634.
2007-05-09 19:41:25 +00:00
Gleb Natapov
2562253678 Do more work at RDMA frag preparation time and less work at RDMA frag sending
time.

This commit was SVN r14627.
2007-05-09 12:11:51 +00:00
Gleb Natapov
78fda79630 Use size_t instead of uint64_t in call to convertor cloning.
This commit was SVN r14626.
2007-05-09 10:02:06 +00:00
Pavel Shamis
e2d0e27111 Adding:
* openib_finalize flow for openib btl
* async event handler for openib btl

This commit was SVN r14623.
2007-05-08 21:47:21 +00:00
Terry Dontje
f864348f97 Put an ifdef to conditionalize the use of memcpy for sparcv9 platforms to
avoid alignmment issues.  This commit fixes trac:1009.

This commit was SVN r14608.

The following Trac tickets were found above:
  Ticket 1009 --> https://svn.open-mpi.org/trac/ompi/ticket/1009
2007-05-08 17:17:34 +00:00
Jeff Squyres
ecf5a3b8dd Fix compiler warning
This commit was SVN r14604.
2007-05-08 13:12:50 +00:00
Sven Stork
d0c936ca85 - export required symbols or library is useless
This commit was SVN r14595.
2007-05-07 13:59:37 +00:00
Sven Stork
32564d73da - the define is always defined independent, therefore we need to check if
it's 1 or not

This commit was SVN r14594.
2007-05-07 13:18:09 +00:00
Jeff Squyres
86f48bf0d7 Fix ordering of dimensions for status arrays. Thanks to Randy Bramley
for noticing the problem.

This commit was SVN r14591.
2007-05-05 05:02:02 +00:00
Sven Stork
a04c8eb39a - Bring over the visibility feature, for a finer symbol export control
via the visibility feature that is provided by some compilers.

  Per default this feature is disabled, to enable it you need to
  configure with --enable-visibility and obviously you need a compiler
  with visibility support. Please refer to the wiki for more information.
  https://svn.open-mpi.org/trac/ompi/wiki/Visibility

This commit was SVN r14582.
2007-05-04 09:03:37 +00:00
Jelena Pjesivac-Grbovic
625c6739ab Removing warning about unsed variable
This commit was SVN r14579.
2007-05-03 20:26:41 +00:00
Gleb Natapov
8029893489 In multithreaded application sending of initial portion of a request may overlap
with RDMAing the rest of it. Also more than one RDMA writes can be performed
simultaneously by different threads. To make this code thread safe this patch
clones original request convertor for each RDMA fragment.

This commit was SVN r14574.
2007-05-03 09:13:17 +00:00
Jelena Pjesivac-Grbovic
9eff74ad4d Modifying generalized reduce "synchronized" behavior:
- Removing "small" message size limit because it really does not relate to the eager size
accross the board.
Now, the leaf nodes in generalized reduce will use blocking send (DEFAULT/ORIGINAL BEHAVIOR) 
either when the maximum number of outstanding requests is 0 or 
when the total number of segments is less than the maximum number of outstanding requests.
Otherwise, it will send messages using non-blocking synchronized send operation.

This commit was SVN r14572.
2007-05-02 21:42:45 +00:00
George Bosilca
69642a9cd4 Remove 2 warnings about ptrdiff_t to unsigned long implicit conversion.
This commit was SVN r14565.
2007-05-01 19:47:33 +00:00
Brian Barrett
a25ce44dc1 Clean up the preconnect code:
* Don't need the 2 process case -- we'll send an extra message, but
    at very little cost and less code is better.
  * Use COMPLETE sends instead of STANDARD sends so that the connection
    is fully established before we move on to the next connection.  The
    previous code was still causing minor connection flooding for huge
    numbers of processes.
  * mpi_preconnect_all now connects both OOB and MPI layers.  There's
    also mpi_preconnect_mpi and mpi_preconnect_oob should you want to
    be more specific.
  * Since we're only using the MCA parameters once at the beginning
    of time, no need for global constants.  Just do the quick param
    lookup right before the parameter is needed.  Save some of that
    global variable space for the next guy.

Fixes trac:963

This commit was SVN r14553.

The following Trac tickets were found above:
  Ticket 963 --> https://svn.open-mpi.org/trac/ompi/ticket/963
2007-05-01 04:49:36 +00:00
Adrian Knoth
d63d125a88 I guess we only need this when IPv6 is enabled.
This commit was SVN r14551.
2007-04-29 16:38:34 +00:00
Adrian Knoth
5765ecc22e This patch reverts r14549 while retaining IPv6 support.
Re #1008

This commit was SVN r14550.

The following SVN revision numbers were found above:
  r14549 --> open-mpi/ompi@386baed55b
2007-04-29 16:23:11 +00:00
Adrian Knoth
386baed55b Hotfix for IPv6 support. Closes trac:1008
This commit was SVN r14549.

The following Trac tickets were found above:
  Ticket 1008 --> https://svn.open-mpi.org/trac/ompi/ticket/1008
2007-04-29 13:46:45 +00:00
George Bosilca
bb481273a6 Typos.
This commit was SVN r14546.
2007-04-28 19:15:53 +00:00
George Bosilca
46265db0a9 Update the TCP BTL in order to bring back some of the functionalities lost
during the IPv6 patch. The most important is the multi BTL support. There
was a quite interesting bug. Instead of setting up the multiple connections
over different physical devices, based on the time when these connections
were created most of the time they were all using the same physical network.
Which, of course, was not the intended goal, as we top at the maximum
bandwidth available over one device instead of gathering all available
bandwidth from all devices.

Second, the IPv6 RFC suggest to use sockaddr_storage as a holder for the
IP information, but use a sockaddr* when we pass it to functions. This is
only partially corrected by this patch.

Some other minor cleanups.

This commit was SVN r14544.
2007-04-28 19:13:47 +00:00
Josh Hursey
4c453caab6 Make the check a bit better
This commit was SVN r14542.
2007-04-27 17:38:36 +00:00
Josh Hursey
486f29eb6b Make sure to use the new metadata flags
This commit was SVN r14541.
2007-04-27 17:18:26 +00:00
Sven Stork
8d92773067 - export required symbol
This commit was SVN r14536.
2007-04-27 11:38:45 +00:00
Rainer Keller
63b904ed1d - Don't segfault, when calling PERUSE_Init before MPI_Init...
This commit was SVN r14535.
2007-04-26 21:06:08 +00:00
Rainer Keller
1aceece03f - Add a few comments for elements for structs, a few spelling fixes.
No functional change.

This commit was SVN r14534.
2007-04-26 21:03:38 +00:00
Rainer Keller
ce32b918da - Fixes for for unlocking the mutex in case of error in functions
mca_btl_openib_post_srr and
     btl_openib_endpoint_post_rr

This commit was SVN r14530.
2007-04-26 13:33:02 +00:00
Sven Stork
fe3b08004e - export symbols that are needed by the fortran libs
This commit was SVN r14527.
2007-04-26 09:34:41 +00:00
Rainer Keller
6f9251ed39 - Small fixes by PGI -Minform=inform
This commit was SVN r14524.
2007-04-26 08:16:07 +00:00
Josh Hursey
af38efd27c Use more of the datatype engine supplied functions
This commit was SVN r14519.
2007-04-26 00:06:22 +00:00
Jelena Pjesivac-Grbovic
3eac49aa59 Adding flow control for leaf nodes in generalized reduce structure.
This "feature" is disabled by default and it should not affect the current performance.

In case when the message size is large and segment size is smaller than eager size for particular interface,
the leaf nodes in generalized reduce function can overflood parent nodes by sending all segments without 
any synchronization.  This can cause the parent to have HIGH number of unexpected messages (think 16MB 
message with 1KB segments for example).  In case of binomial algorithm root node always has at least one
child which is leaf, so this can potentially affect the root's performance significantly [Especially in 
large communicators where root may have quite a few children (binomial tree for example)].
When the segment size is bigger than the eager size, rendezvous protocol ensures that this does 
not happen so it is not necessary.
Originally, the problem was exposed in "infinite" bucket allocator clean up time for "small" segment sizes
(which may explain some "deadlocks" on Thunderbird tests).

To prevent this, we allow user to specify mca parameter "--mca coll_tuned_reduce_algorithm_max_requests NUM"
this limits number of outstanding messages from a leaf node in generalized reduce to the parent to NUM.
Messages are sent as non-blocking synchrnous messages, so syncronization happens at "wait" time.
The synchronization actually improved performance of pipeline and binomial algorithm for large message sizes
with 1KB segments over MX, but I need to test it some more to make sure it is consistent.

Since there is no easy way to find out what is "the eager" size for particular btl, I set the limit to 4000B.
If message/individual segment size is greater than 4000B - we will not use this feature.  This variable may
or may not be exposed as mca parameter later...

I did not have any problems running it and both "default" and "synchronous" tests passed Intel Reduce* tests 
up to 80 processes (over MX).

This commit was SVN r14518.
2007-04-25 20:39:53 +00:00
Adrian Knoth
e3d35258b4 Cosmetics. Brian fixes my crappy code and I fix the curly braces.
That's teamwork, right? ;)

This commit was SVN r14517.
2007-04-25 20:17:19 +00:00
Brian Barrett
4b8bb70afb A couple cleanups for the IPv6 support:
- make opal_sockaddr2str() take a sockaddr_storage instead of a sockaddr_in6
    so that it works for IPv4 and IPv6 addresses, and remove a whole bunch
    of #ifs in the OOOB code.
  - Fix a compiler warning in the TCP BTL due to run-time determined
    array size by making it a dynamicly allocated array.
  - Fix the unpacking code of IPv4 addresses when using IPv6 support, so
    that the address is in the correct location (instead of in an IPv6
    structure, use an IPv4 structure).  Refs trac:1005.

This commit was SVN r14514.

The following Trac tickets were found above:
  Ticket 1005 --> https://svn.open-mpi.org/trac/ompi/ticket/1005
2007-04-25 19:08:07 +00:00
Adrian Knoth
d1ce39de4f Move mca_btl_tcp_addr_isipv4public to opal_addr_isipv4public
This commit was SVN r14512.
2007-04-25 18:06:06 +00:00
Donald Kerr
80d984441f change so that we only check connection queue when expecting a connection; create a mca parameter that controls frequency at which the async queue is checked
This commit was SVN r14511.
2007-04-25 17:46:25 +00:00
Ralph Castain
7d0f51e6b9 Begin setting up for a change to the OOB information passing functionality - this is totally transparent at the moment (need to change computers).
This commit was SVN r14510.
2007-04-25 17:36:26 +00:00
Jeff Squyres
c4c68e666a Merge in the ipv6 work from /tmp/ipv6-merge.
This commit was SVN r14503.
2007-04-25 01:55:40 +00:00
Donald Kerr
cae24fcde1 move mca parameter registration into own .c and .h files
This commit was SVN r14493.
2007-04-24 18:34:16 +00:00
Josh Hursey
8c2385416f Per a developer request -
Make sure that the wrapper selection is compiled out if not enabling FT. Before the 
logic would skip over it since the conditional if statements would not be satisfied, 
now there are no additional if statements when compiled out.

With this modification the selection logic looks nearly identical to pre-r14051
with the exception of the non-FT related improvements.

This commit was SVN r14491.

The following SVN revision numbers were found above:
  r14051 --> open-mpi/ompi@dadca7da88
2007-04-24 17:08:48 +00:00
Ralph Castain
18b2dca51c Bring in the code for routing xcast stage gate messages via the local orteds. This code is inactive unless you specifically request it via an mca param oob_xcast_mode (can be set to "linear" or "direct"). Direct mode is the old standard method where we send messages directly to each MPI process. Linear mode sends the xcast message via the orteds, with the HNP sending the message to each orted directly.
There is a binomial algorithm in the code (i.e., the HNP would send to a subset of the orteds, which then relay it on according to the typical log-2 algo), but that has a bug in it so the code won't let you select it even if you tried (and the mca param doesn't show, so you'd *really* have to try).

This also involved a slight change to the oob.xcast API, so propagated that as required.

Note: this has *only* been tested on rsh, SLURM, and Bproc environments (now that it has been transferred to the OMPI trunk, I'll need to re-test it [only done rsh so far]). It should work fine on any environment that uses the ORTE daemons - anywhere else, you are on your own... :-)

Also, correct a mistake where the orte_debug_flag was declared an int, but the mca param was set as a bool. Move the storage for that flag to the orte/runtime/params.c and orte/runtime/params.h files appropriately.

This commit was SVN r14475.
2007-04-23 18:41:04 +00:00
Donald Kerr
3f428af7b8 couple of minor changes to fix #973 and seperated eager rdma fragments into structure only and data only area
This commit was SVN r14470.
2007-04-23 17:41:34 +00:00
Jelena Pjesivac-Grbovic
53cbec7a09 Make coll/tuned dynamic rules more verbose (when promted with --mca coll_base_verbose 1)
This commit was SVN r14469.
2007-04-23 16:34:52 +00:00
Rich Graham
ce35761683 make sure not to go out of bounds. element i+1 of bml_btls
is referenced, which for i-arr_size-1 is beyond the array dimentions.

This commit was SVN r14464.
2007-04-22 21:43:34 +00:00
Sharon Melamed
cf3f41288b Add pkey value MCA parameter. if this param is used,
only ports with the actual pkey value will be initiate.

This commit was SVN r14463.
2007-04-22 10:22:12 +00:00
Adrian Knoth
339dbf6cd5 Cosmetics. Enforcing style guide.
This commit was SVN r14459.
2007-04-21 21:47:25 +00:00
Josh Hursey
4159b72a60 Some minor updates to go along with commit r14457
This commit was SVN r14458.

The following SVN revision numbers were found above:
  r14457 --> open-mpi/ompi@2af38229c1
2007-04-21 21:24:44 +00:00
Josh Hursey
2af38229c1 Re-worked the implementation of the LAM-like coord component.
It's a bit longer, but much more clear in it's implementation I believe.

Fundamentally it is the same, but is much more solid in the implementation.
I created quite a few directed tests that this version of the implementation 
now passes.

This commit was SVN r14457.
2007-04-21 20:35:01 +00:00
Jeff Squyres
401a072888 Revert r14435 -- it breaks compiling on Linux with at least the PGI
compiler suite.  The rule is that ompi_info.h is supposed to be the
''first'' file included so that it can affect system header files if
necessary (and it is sometimes necessary, such as with the PGI
compiler suite).

If this breaks VC on Windows, we'll have to find another fix.  More on
the mailing list...

This commit was SVN r14453.

The following SVN revision numbers were found above:
  r14435 --> open-mpi/ompi@a20b43ace9
2007-04-21 00:51:31 +00:00
Jeff Squyres
5bebd24250 Bring over Brian's installdirs fixes from this afternoon (r14445).
This commit was SVN r14450.

The following SVN revision numbers were found above:
  r14445 --> open-mpi/ompi@13d366b827
2007-04-21 00:16:31 +00:00
Jeff Squyres
0ba47105ed Merge the /tmp/jms-installdirs-trunk branch into the trunk. This
finally brings in functionality that is already on the 1.2 branch, and
was developed and tested in the v1.2ofed branch (and other places).

Short version of new features:

 * Support for ibv_fork_init() 
 * Automatically fill in the openib BTL bandwidth value by 
   querying the HCA port 
 * Installdirs functionality 
 * Fixes to always use -I in the Fortran wrapper compilers (#924) 
 * Gleb's mpool updates 
 * Remove some kruft in btl/openib/configure.m4, therefore 
   fixing the harmless warnings noted in #665 
 * Bunches of updates to the Linux RPM spec file 

I.e., effectively the same thing that r14411 brought to the v1.2
branch.

Also effectively brought in r14432 and r14433 (some fixes on top of
the original r14411 commit to v1.2).  Still need to bring in the moral
equivalent of r14445 after this commit (fixes to installdirs).

This commit was SVN r14449.

The following SVN revision numbers were found above:
  r14411 --> open-mpi/ompi@83b31314ae
  r14432 --> open-mpi/ompi@a48f160595
  r14433 --> open-mpi/ompi@68f346d2bc
  r14445 --> open-mpi/ompi@13d366b827
2007-04-21 00:15:05 +00:00
Josh Hursey
eef364546c Check for NULL before trying to use the variable.
This commit was SVN r14444.
2007-04-20 17:17:11 +00:00
Jeff Squyres
40f4b60a2a Use #ifdef, not #if
This commit was SVN r14439.
2007-04-20 14:26:20 +00:00
Shiqing Fan
a20b43ace9 The header file ompi_config.h should be included in ompi_info.h file but not in the .cc files. This make it could be compiled with VC compiler, and it works no difference under linux systems.
This commit was SVN r14435.
2007-04-20 09:03:16 +00:00
Josh Hursey
12e5d0e817 ft_event Commit:
- Move the PML Modex stuff out of the BML -- Abstraction violation.
- Also fix the location of the add_procs with respect to the stage gates.

This commit was SVN r14422.
2007-04-19 03:05:12 +00:00
Josh Hursey
d12ddcdb7a Protect the free since if we never send any messages this could be NULL.
This commit was SVN r14421.
2007-04-19 02:17:50 +00:00
George Bosilca
51fc2474f1 Don't keep the data attached to a fragment segmented when we have
to move it into the unexpected queue. Instead pack the data in
only one buffer. Now the code look more optimized and clear, but
I have a doubt about who's using this functionality. I think that
all BTLs always return only one memory segment attached to the
matching fragment (i.e. there is no unexpected iov type receive).

This commit was SVN r14416.
2007-04-18 15:52:11 +00:00
Sven Stork
037b01ce9e - more symbols that need to be exported
This commit was SVN r14415.
2007-04-18 14:53:56 +00:00
George Bosilca
66a110e115 Add some comments on the internals of the bucket structure. Alter the cleanup
function to make it more scalable. The memory fragmentation is still high, but
at least in most of the cases (where all ressources are correctly released
before the cleanup) the code is now highly efficient. Before the code execute
in (N * (N-1))!, which take a while when the number of allocated ressources
increase (which is the case when a lot of unexpected messages are created).

The fix consist of checking if all items are freed and if it's the case
then do not recreate the free items list (as we know that everything will
be released). If this condition is not true, we fall back on the
original execution path (which is still sub-sub-sub ... optimal).

This commit was SVN r14406.
2007-04-17 20:43:30 +00:00
Jeff Squyres
82caceda08 A minor change to ROMIO's configure script: make it use exactly the
same "restrict" check as the top-level OMPI configure.ac script so
that it will guarantee to always get the same result.  Therefore, the
#define for restrict will always have the same value in both
opal_config.h and romioconf.h, and we get 7 less warnings (6 in the IO
ROMIO component, 1 in ROMIO itself) when compiling with icc on Linux
(because PAC_C_RESTRICT and AC_C_RESTRICT would get different values
for the "restrict" #define in this case).

This commit was SVN r14387.
2007-04-17 03:10:06 +00:00
Adrian Knoth
e3178fd39f Cosmetics. PTLs are now called BTLs.
This commit was SVN r14382.
2007-04-16 10:12:27 +00:00
Josh Hursey
8f119d9063 Closes trac:977
Fix for memory corruption in the restarted process stack. This stemed from 
the brute force method we were previously using. This commit fixes this by
using a lighter weight solution focused in the r2 BML instead of above the PML.
This is a more efficient and flexible solution, and it solves the original
problem.

In the process I pulled out the ft_event function in the tcp BTL and r2 BML
into a set of *_ft.[c|h] files just to keep any updates to these code paths
as isolated as possible to make merging easier on everyone.

This commit was SVN r14371.

The following SVN revision numbers were found above:
  r2 --> open-mpi/ompi@58fdc18855

The following Trac tickets were found above:
  Ticket 977 --> https://svn.open-mpi.org/trac/ompi/ticket/977
2007-04-14 02:06:05 +00:00
Jeff Squyres
51f286d737 Just like r14289 on the ORTE trunk:
Per discussions with Brian and Ralph, make a slight correction in
where components are installed. Use $pkglibdir, not $libdir/openmpi,
so that when compiled in the orte trunk, components are installed to
the right directory (because the component search patch is checking
$pkglibdir).

This commit was SVN r14345.

The following SVN revisions from the original message are invalid or
inconsistent and therefore were not cross-referenced:
  r14289
2007-04-12 11:19:42 +00:00
Gleb Natapov
d41ca417e8 Delete declaration of non-existent functions and no longer relevant comment.
This commit was SVN r14341.
2007-04-12 08:12:31 +00:00
George Bosilca
20f0ec584a A tricky optimization. On my test machine it improve the bandwidth by about 3Mb/s out of 580Mb/s. But
the real interest is for small to middle size unexpected messages. The unexpected messages are copied
by the PML in it's own unexpected buffers. Therefore, there is no reason to make a first copy in the
TCP BTL. The BTL can handle to the PML it's own buffer, and can be sure that once the callback
completed it can reuse the buffer, no matter what happened with the fragment.

This commit was SVN r14320.
2007-04-12 04:52:29 +00:00
George Bosilca
88365518aa Small cleanup.
This commit was SVN r14319.
2007-04-12 04:34:53 +00:00
George Bosilca
6b217d31e1 Add OPAL_LIKELY where necessary.
This commit was SVN r14318.
2007-04-12 04:32:07 +00:00
Galen Shipman
ebca0bb34e fix for aggregated writes
This commit was SVN r14314.
2007-04-11 22:07:19 +00:00
Galen Shipman
d7e428909e two fixes, one mine, the other gleb's, I'm committing for gleb due to
time difference...  

1) The PML makes an assumption on local/remote completion semantics of the BTL
which Self BTL does not obey, nor should it, so we fix the PML
2) The Get protocol must handle the case when sender and reciever do not agree
on wheter the data is contiguous 

This commit was SVN r14313.
2007-04-11 22:03:06 +00:00
Josh Hursey
fbc59f668c fix typo
This commit was SVN r14301.
2007-04-11 15:39:42 +00:00
Josh Hursey
5efae25390 No functionality changes (yet).
Just fix the indentation to meet the coding standard.

This commit was SVN r14300.
2007-04-11 15:19:51 +00:00
Jeff Squyres
85d7678350 Revert r14286; it worked for icc, but not for gcc. #$%@#$%
Sorry for configure changes during the day; I totally forgot about
that.  :-(

This commit was SVN r14288.

The following SVN revision numbers were found above:
  r14286 --> open-mpi/ompi@0083eba18e
2007-04-10 15:42:59 +00:00
Jeff Squyres
0083eba18e Comment out the PAC_C_RESTRICT test from ROMIO's configure.in script.
The top-level OMPI configure script already checks for "restrict" and
will issue a #define for it.  PAC_C_RESTRICT would also check for
restrict, but sometimes come up with a different answer than the
top-level OMPI configure script, thereby resulting in conflicting
#define's for "restrict" (e.g., icc 9.0/9.1 on linux x86-64).

So it's easiest just to remove this test from ROMIO's configure.in
script.

This commit was SVN r14286.
2007-04-10 14:50:47 +00:00
Rich Graham
f481722bdf move the code that sets the thread level information before the btl are
initialized, so that the btl's have this information for correct setup.

This commit was SVN r14258.
2007-04-07 05:06:47 +00:00
Tim Prins
f0e6a28a1f pedantic indentation...
This commit was SVN r14251.
2007-04-06 19:18:31 +00:00
Josh Hursey
38547459ae Improve the cleanup process in ob1
Remove a redundant statement in the r2 BML.

This commit was SVN r14228.

The following SVN revision numbers were found above:
  r2 --> open-mpi/ompi@58fdc18855
2007-04-05 17:37:29 +00:00
Josh Hursey
98fb9f26ef Some cleanup.
- Remove an old comment from crcp_base_fns.c
- Let ob1 have its very own ft_event function (which I'll fill in shortly)
- Make sure ob1 finalizes the bsend stuff so we don't leave a bunch of memory sitting around
- PML base - destruct the array upon finalize. Shrink the include search so it stops after finding a match

This commit was SVN r14222.
2007-04-05 13:52:05 +00:00
Josh Hursey
a8918fe3d5 pedantic cleanup. Switch loop to lowest rank sends first
This commit was SVN r14215.
2007-04-04 14:23:45 +00:00
Edgar Gabriel
4d2b3e859d fix the indenting from tabs to spaces :-)
This commit was SVN r14211.
2007-04-03 21:33:44 +00:00
Edgar Gabriel
188f770d94 ok, increase the reference count on ompi_mpi_group_null twice when
creating ompi_mpi_comm_null, since the destructor of ompi_mpi_comm_null will
decrease the reference counter of ompi_mpi_group_null twice according to the
last fix of Mohamad.

Added also a lengthy comment in ompi_comm_finalize about why we do 
not decrease the reference counters for ompi_mpi_comm_null,
ompi_mpi_group_null etc. for the parent 
communicator, although we do increase it in ompi_comm_init

This commit was SVN r14210.
2007-04-03 21:16:26 +00:00
Li-Ta Lo
ec8a859a44 fixed typo
This commit was SVN r14207.
2007-04-03 17:21:54 +00:00
George Bosilca
667bda0fef Rework the code a little bit to make things simpler.
This commit was SVN r14203.
2007-04-03 16:05:51 +00:00
George Bosilca
cb1b976486 Big update. Correct the behavior for true_lb and true_ub computation
when the size of the data is zero. Now they are not updated, which leave
us with the correct memory layout in all situations (so far). Update all
the comments to reflect exactly the supported behavior of the DDT engine.

This commit was SVN r14202.
2007-04-03 16:05:15 +00:00
Josh Hursey
51daa15f9c play a bit nicer with references.
This commit was SVN r14201.
2007-04-02 22:27:52 +00:00
Josh Hursey
5ff1c10e70 minor cleanup
This commit was SVN r14199.
2007-04-02 20:39:36 +00:00
Josh Hursey
b0b91a5fde A couple more fixes for async case.
Mostly working again, 1 small bug I'm still tracking.

This commit was SVN r14198.
2007-04-02 20:00:58 +00:00
Josh Hursey
71937c3eaf A bit of cleanup for async case... Still one bug in there.
This commit was SVN r14197.
2007-04-02 19:25:22 +00:00
George Bosilca
120cf76ad8 Remove some warnings.
This commit was SVN r14196.
2007-04-02 19:11:06 +00:00
Mohamad Chaarawi
0e98bf2ac6 quick fix for the cart create problem caused by the previous memory leak
fix

This commit was SVN r14195.
2007-04-02 19:06:52 +00:00
George Bosilca
8273c5eeba Correct an error introduced by commit r14180.
This commit was SVN r14191.

The following SVN revision numbers were found above:
  r14180 --> open-mpi/ompi@1cb26e3b9c
2007-04-02 02:59:23 +00:00
George Bosilca
f2a6b9394f Deal with the include spree. Protect "environ" on Windows.
Some others minors modifications in order to make it
compile [again] on Windows.

This commit was SVN r14188.
2007-04-01 16:16:54 +00:00
Tim Prins
80e047b843 make the mx btl compile again...
This commit was SVN r14183.
2007-04-01 02:49:23 +00:00
George Bosilca
f518a9c1f6 Remove some warnings from the data-type engine.
This commit was SVN r14181.
2007-03-31 04:14:47 +00:00
George Bosilca
1cb26e3b9c Finally the convertor export a convenience function to allow a consistent
computation of the current location on the pack/unpack process. This can
be used both for retrieving the pointer to the first byte (in the special
case of the cached RDMA protocol) and for getting the current
position (for the pipelined protocol).

I modified all BTLs, but most of them are still untested.

This commit was SVN r14180.
2007-03-30 22:02:45 +00:00
Mohamad Chaarawi
8f4f992bfc fixed the memory leak problem by decrementing the ref count on the
remote group in case of Intra communicators. This needs to go in V1.2.
We will file a move request on monday..

This commit was SVN r14179.
2007-03-30 19:30:40 +00:00
Galen Shipman
a78672be2b fix mpi_leave_pinned case for arbitrary datatypes
George will be streamlining this with a new convertor function soon... 

This commit was SVN r14174.
2007-03-30 02:06:08 +00:00
Galen Shipman
db63458495 bring disable_sbrk back online, there was a change to properly support AIX
some time ago (last summer) that included checking for M_TRIM_THRESHOLD and
M_MMAP_MAX, unfortunately we didn't include <malloc.h> which is where these
are define, so disabling sbrk for the registration cache has been busted for
some time. 

This commit was SVN r14169.
2007-03-29 16:11:00 +00:00
George Bosilca
cb93b1d40d Deal with compiler warnings and size_t in same time ... It's getting more
and more tricky !!!

This commit was SVN r14162.
2007-03-28 22:02:13 +00:00
George Bosilca
4bc69447b4 Setting a size_t to -1 leads to unexpected results ...
This commit was SVN r14160.
2007-03-28 18:23:42 +00:00
George Bosilca
cc65814969 And set the message size before the first use too.
This commit was SVN r14159.
2007-03-28 18:01:13 +00:00
George Bosilca
b540545fa7 Set the communicator size before using it.
This commit was SVN r14158.
2007-03-28 17:59:21 +00:00
George Bosilca
78f362d0d6 Be consistent about the definitions of mca_mpool_base_page_size and
mca_mpool_base_page_size_log. They are exported by the mpool/base/base.h,
if some other code need them, then it should include this file
instead of having it's own redefinition of these externals.

This commit was SVN r14156.
2007-03-28 14:14:05 +00:00
Shiqing Fan
fb50a72e92 Unnecessary header removed.
This commit was SVN r14152.
2007-03-27 14:32:30 +00:00
Shiqing Fan
91cfb2f149 A few mismatched declearations are fixed, and several header files are added for Cygwin...
This commit was SVN r14151.
2007-03-27 14:17:25 +00:00
Mohamad Chaarawi
bfaf9d4a12 Added new module for intercomm collectives. This will require an
autogen.

This commit was SVN r14149.
2007-03-27 02:06:42 +00:00
Brian Barrett
e283e6f9d9 Retry of r14142, without the one-sided code...
Back out r14073 - it speeds up TCP latency / bandwidth but at the same time 
it kills ROMIO and one-sided performance when using only TCP. The problem 
is that it only allows those two to be progressed every couple of seconds, 
leading to what looks like hangs in the one-sided tests (and the ROMIO stuff, 
although people seem to not notice that at this point). 

This commit was SVN r14144.

The following SVN revision numbers were found above:
  r14073 --> open-mpi/ompi@64fbbc20b8
  r14142 --> open-mpi/ompi@241545a098
2007-03-26 16:01:27 +00:00
Brian Barrett
62e5e81e99 revert r14142, as the onesided change should *not* have come over
This commit was SVN r14143.

The following SVN revision numbers were found above:
  r14142 --> open-mpi/ompi@241545a098
2007-03-26 15:58:41 +00:00
Brian Barrett
241545a098 Back out r14073 - it speeds up TCP latency / bandwidth but at the same time
it kills ROMIO and one-sided performance when using only TCP.  The problem
is that it only allows those two to be progressed every couple of seconds,
leading to what looks like hangs in the one-sided tests (and the ROMIO stuff,
although people seem to not notice that at this point).

This commit was SVN r14142.

The following SVN revision numbers were found above:
  r14073 --> open-mpi/ompi@64fbbc20b8
2007-03-26 15:56:23 +00:00
Josh Hursey
7c4ca3c420 remove some stale code
This commit was SVN r14134.
2007-03-23 14:11:12 +00:00
Gleb Natapov
e5450613b5 Add new SM BTL parameter btl_sm_cb_max_num. If set to value greater then zero
it limits the number of circular buffers allocated between each pair of peers.
This allows for more tight memory usage control.

This commit was SVN r14120.
2007-03-22 12:21:42 +00:00
Gleb Natapov
efe0323d35 Initialize fifos at SM BTL init time instead of waiting for first send. This
waist slightly more memory, but prevents problem when fifo cannot be allocated
later during a job run when memory resource is exhausted.

This commit was SVN r14119.
2007-03-22 12:18:44 +00:00
Galen Shipman
ace68b1883 Change the way we handle unexpected messages,
if less than or equal  pml_ob1_unexpected_limit just buffer in the PML level recv
fragment else allocate a buffer via the bucket allocator 

This commit was SVN r14117.
2007-03-22 01:00:34 +00:00
Gleb Natapov
c389c47d79 Fix SM connectivity calculations.
This commit was SVN r14109.
2007-03-21 13:29:19 +00:00
Jeff Squyres
3e2031e0e3 Finally commit something that has been sitting around in one of my
development trees since last year (had to wait for some intel tests to
run yesterday, so I finally took the time to finish this work):

 * Improve MPI API argument checking by also checking for NULL values
   (especially helps when invalid Fortran MPI handles are passed,
   because the various MPI_*f2c functions are supposed to return an
   "invalid" MPI handle [meaning NULL] when this happens).  So now
   OMPI will generate an MPI exception rather than a segv.
 * Removed a few redundant DATATYPE_NULL checks.
 * Also check for some other forms of "invalid" handles (e.g., already
   been freed, etc.) in some cases.  We could probably be a bit more
   stringent in this regard if we really wanted to.
 * Change MPI_Get_processor_name to zero out the string up to
   MPI_MAX_PROCESSOR_NAME characters, because the MPI spec says that
   the string must be at least that long.  We were already passing
   that length to gethostname(), anyway.

This commit was SVN r14100.
2007-03-21 11:10:42 +00:00
Gleb Natapov
a1a14aa4c3 Add memory barriers during SM btl initialization.
This commit was SVN r14099.
2007-03-21 10:25:10 +00:00
Gleb Natapov
435565590f Don't relay on opcode to decide how to progress pending message.
This commit was SVN r14098.
2007-03-21 07:59:59 +00:00
Josh Hursey
299332ecac fix small compiler warning
This commit was SVN r14097.
2007-03-21 04:44:54 +00:00
Brian Barrett
464d536928 remove debugging printf
This commit was SVN r14088.
2007-03-20 21:28:28 +00:00
Josh Hursey
3492fdeae3 Fix a couple of compiler warnings (errors?) caught by ICC testing at Cisco.
This commit was SVN r14080.
2007-03-20 14:12:13 +00:00
George Bosilca
8c9e4baa47 Add multi-link capabilities to the TCP BTL. This is useful for systems where the
latency is high and the network relatively fast. This will allow for more kernel
level buffering, which allow overlap between system calls and communications.
Somehow, even on fast clusters there is an improvement (non significant).

This patch create multiple modules for the same device, which in turn will
create multiple sockets between the peers. By default the number of BTL by
device is set to 1, so there is no fundamental difference with the current
version. Change the value of btl_tcp_links to enable multiple links between
peers.

This commit was SVN r14076.
2007-03-20 11:50:17 +00:00
George Bosilca
0edd770644 Nothing really relevant.
This commit was SVN r14075.
2007-03-20 11:21:23 +00:00
George Bosilca
4332295b32 Typos.
This commit was SVN r14074.
2007-03-20 11:18:05 +00:00
George Bosilca
64fbbc20b8 Switch the event engine to a blocking mode if there is no high performance
networks available.

This commit was SVN r14073.
2007-03-20 11:15:08 +00:00
Rainer Keller
249abd29c2 - Mark some deprecated functions (two still commented) and fix to
not use opal_cmd_line_make_opt anymore.

This commit was SVN r14072.
2007-03-20 10:08:58 +00:00
Gleb Natapov
e551c5f1a3 Get rid of separate sm BTL for different shared memory base addresses. Now,
when we precalculate most of the addresses there is no point to have separate
BTL for this. The sm_progress() code become much more simple as a result.

This commit was SVN r14071.
2007-03-20 08:15:58 +00:00