1
1
Граф коммитов

5319 Коммитов

Автор SHA1 Сообщение Дата
Edgar Gabriel
52063267df commit of the OMPIO modules and frameworks.
This commit was SVN r25079.
2011-08-25 20:08:17 +00:00
Matthias Jurenz
0409675f90 Changes to VT:
- updated version number to 5.11.2
	- fixed even more Coverity warnings
	- vtunify:
		- replaced std::vector relict by LargeVectorC (fixes segfault during gathering)
	- vtwrapper:
		- do also escape '\', ''', '(', and ')' in arguments

This commit was SVN r25071.
2011-08-23 10:12:43 +00:00
Mike Dubman
98f382ba0e fixes in mxm mtl
This commit was SVN r25066.
2011-08-19 22:18:17 +00:00
Matthias Jurenz
9f154e7060 Changes to VT:
- fixed compiler warnings
	- fixed Coverity warnings
	- vtrun:
		- do preload libvt-fmpi.so only if it is available

This commit was SVN r25056.
2011-08-17 12:23:19 +00:00
Shiqing Fan
20ee92c16e Make the compiler wrappers work correctly for MinGW build.
This commit was SVN r25051.
2011-08-16 12:32:41 +00:00
Shiqing Fan
3af7c9f7bb Complete the MinGW build support on Windows.
This commit was SVN r25048.
2011-08-15 09:47:23 +00:00
Shiqing Fan
627f1dd351 Correct several export declarations.
This commit was SVN r25047.
2011-08-15 09:45:51 +00:00
Jeff Squyres
1cbfb53801 r24976 wasn't quite right -- you now actually get a warning if you
specify btl_tcp_if_include because btl_tcp_if_exclude is defaulted to
the loopback devices.

This commit does a few things:

 * Introduce a new OPAL MCA base function:
   mca_base_param_check_exclusive_string().  It checks to see that the
   ''user'' does not set two MCA parameters that are mutually
   exclusive by checking the source of those MCS param values.
 * Use the above function in many BTLs (and the OOB TCP) to ensure
   that <foo>_if_include and <foo>_if_exclude are not both specified
   ''by the user''.
 * Re-arrange many of these BTLs to move their MCA registration code
   into a separate component_register() function (vs. the
   component_open() function).

This code has been nominally reviewed and checked by Ralph, George,
Terry, and Shiqing.

This commit was SVN r25043.

The following SVN revision numbers were found above:
  r24976 --> open-mpi/ompi@8f4ac54336
2011-08-10 17:24:36 +00:00
Mike Dubman
e3c869d83b fix double free
This commit was SVN r25041.
2011-08-10 05:47:55 +00:00
Mike Dubman
a751cd93d3 improve debug macro availability
This commit was SVN r25022.
2011-08-09 10:54:08 +00:00
Mike Dubman
bfd75de6f9 fix selection logic: if no suitable device found - disqulaify mxm w/o complains.
This commit was SVN r25021.
2011-08-09 07:09:37 +00:00
Wesley Bland
09274cd047 Make sure that the epoch is initialized everywhere so we don't get weird output
during valgrind. This shouldn't have caused any problems with any actual
execution. Just extra warnings in valgrind.

This commit was SVN r25015.
2011-08-08 15:11:55 +00:00
Matthias Jurenz
3a6e9b19ee Fixed several Coverity warnings
This commit was SVN r25014.
2011-08-08 12:53:58 +00:00
Mike Dubman
1d3f5e1314 better mxm selection mechanism, some refactoring
This commit was SVN r25005.
2011-08-07 12:06:49 +00:00
Matthias Jurenz
1b402ecb1a Changes to OTF:
- always check the result of OTF_WStream_get*Buffer since it might be NULL in case OTF_File_open fails

Changes to VT:
	- CUDA Tracing:
		- fixed configure stack for filtered kernels
		- fixed buffer size for CUPTI tracing
		- replaced error message with warning to continue tracing, even if CUDA error occured (VTCUDAsynchronizeEvt)
	- vtunify:
		- enlarged minimum message size for transfering local definitions to rank 0
		- use binary search for searching already created global definitions
		- use binary search for searching already created global marker definitions
		- use LargeVectorC instead of std::vector for pre-allocating elements
	- vtwrapper:
		- added options '-vt:CC' and '-vt:c++' which are synonyms for '-vt:cxx'

This commit was SVN r24997.
2011-08-05 12:46:26 +00:00
Yevgeny Kliteynik
7068dc64eb Dynamic SL rework:
- Added dynamic SL support to xoob
 - Fixed seg fault in finalization
 - All the code has been moved to separate files: connect/btl_openib_connect_sl.{c,h}
 - The new files compilation is conditionalized

This commit was SVN r24991.
2011-08-04 20:26:08 +00:00
Jeff Squyres
d28564aa26 Per http://www.open-mpi.org/community/lists/devel/2011/08/9606.php,
comment out some unused parameter names.  I didn't use
__opal_attribute_unused__ because comm_inln.h is (eventually) included
by <mpi.h>, and therefore we don't have all the OPAL config stuff
available.  And it didn't seem worth it to add the optional
attribute_unused stuff to the top of mpi.h.

Thanks to Júlio Hoffimann for reporting the issue.

This commit was SVN r24989.
2011-08-04 15:39:12 +00:00
Rolf vandeVaart
3d3b3d4dad Add support for CUDA registering sm and openib buffers. Feature is disabled by default.
This commit was SVN r24987.
2011-08-04 10:15:45 +00:00
Mike Dubman
7b18ab2fa9 remove unused includes
This commit was SVN r24980.
2011-08-03 07:07:29 +00:00
Mike Dubman
45ea375531 code and readme updates, some refactoring
This commit was SVN r24977.
2011-08-02 14:30:11 +00:00
Jeff Squyres
8f4ac54336 Fixes trac:2838: add a warning message and disqualify the TCP BTL if both
btl_tcp_if_include and btl_tcp_if_exclude are specified. 

This commit was SVN r24976.

The following Trac tickets were found above:
  Ticket 2838 --> https://svn.open-mpi.org/trac/ompi/ticket/2838
2011-08-01 23:30:33 +00:00
Wesley Bland
5fde3e0e00 Move the resilient orte errmgr code into a seperate errmgr for now while it's
still unstable. Reverted errmgr modules back to the original errmgr (with the
updates since the resilient code was brought into the trunk).

This commit was SVN r24958.
2011-07-28 21:24:34 +00:00
Matthias Jurenz
4ca70e5c91 Changes to OTF:
- improved zlib compression
   - otfprofile-mpi:
      - fixed progress

Changes to VT:
   - fixed C++ linker issue for manual instrumentation of multiple files
   - fixed CUDA kernel launch configuration
   - process and thread buffer size can be explicitly specified by the user via the environment variables VT_BUFFER_SIZE and VT_THREAD_BUFFER_SIZE
   - fixed CUDA buffer management
   - vtfilter:
      - fixed progress
   - vtwrapper:
      - link CUPTI library, if available
   - vtsetup:
      - removed fixed path to *.dtd file in vtsetup-data.xml[.in] (fixes 'java.net.MalformedURLException')

This commit was SVN r24950.
2011-07-26 12:47:05 +00:00
Yevgeny Kliteynik
c1ab24c687 openib: added Mellanox ConnectX3 device ID to the device parameters ini file
This commit was SVN r24947.
2011-07-26 12:06:43 +00:00
Mike Dubman
aefffa073d initial implementation of MXM MTL layer
This commit was SVN r24946.
2011-07-26 04:36:21 +00:00
Mike Dubman
96ef2fc0e4 fix handling datatypes which have a gap in the beginning
This commit was SVN r24936.
2011-07-25 06:30:09 +00:00
Jeff Squyres
d6bc78920e Add a few manual cleanups that were missed (i.e., this is the
''other'' direction, so to speak, compared to r24921).

This commit was SVN r24924.

The following SVN revision numbers were found above:
  r24921 --> open-mpi/ompi@bd96d028de
2011-07-22 21:05:39 +00:00
Jeff Squyres
bd96d028de George identified some memory leaks and inconsistencies in the F77 API
when sizeof(int) != sizeof(MPI_Fint).  This commit should fix those
problems. 

This commit was SVN r24921.
2011-07-22 19:49:27 +00:00
Yevgeny Kliteynik
78ea8bcea2 Always defining OMPI_ENABLE_DYNAMIC_SL, not only when the feature enabled.
Also, refactoring some code - make all the checks only when relevant.

This commit was SVN r24900.
2011-07-13 23:19:58 +00:00
Terry Dontje
fbda6aaf89 Fixes trac:2532 issues with 32-bit binaries
This commit was SVN r24891.

The following Trac tickets were found above:
  Ticket 2532 --> https://svn.open-mpi.org/trac/ompi/ticket/2532
2011-07-13 16:38:03 +00:00
Jeff Squyres
51ac69b05f Remove a now-nonexistent file
This commit was SVN r24874.
2011-07-11 23:51:41 +00:00
Abhishek Kulkarni
5501f83fb5 shmem fixes to make the trunk build with C/R flags on.
This commit was SVN r24871.
2011-07-10 23:32:23 +00:00
Jeff Squyres
08a05a1e35 Minor additions to make OMPI trunk compatible with the latest GNU
Autotools:

 * Autoconf 2.68
 * Automake 1.11.1
 * Libtool 2.4
 * m4 1.4.16

This commit was SVN r24867.
2011-07-10 12:11:47 +00:00
Jeff Squyres
b2b781e537 Fix a few miscelaneous memory leaks.
This commit was SVN r24865.
2011-07-08 16:39:58 +00:00
Mike Dubman
fd17f20ed5 Currently MTLs do no handle communicator contexts in any special way,
they only add the context id to the tag selection of the underlying 
messaging meachinsm. 
 
 We would like to enable an MTL to maintain its own context data
per-communicator. This way an MTL will be able to queue incoming eager 
messages and rendezvous requests per-communicator basis.

 The MTL will be allowed to override comm->c_pml_comm member, 
since it's unused in pml_cm anyway. 

This commit was SVN r24858.
2011-07-06 18:25:49 +00:00
Shiqing Fan
1ed0f40d35 Fix a few type casts on Windows.
This commit was SVN r24857.
2011-07-06 08:08:53 +00:00
Yevgeny Kliteynik
4fbe68dd86 Removing trailing white spaces in all the openib btl code.
This commit was SVN r24855.
2011-07-04 14:00:41 +00:00
Yevgeny Kliteynik
5cae33503d Changing the weird non-ASCII sign with '*'
This commit was SVN r24854.
2011-07-04 13:39:38 +00:00
Yevgeny Kliteynik
3e43b21d41 As suggested by Jeff, checking if dynamic SL support
was explicitly requested. If it was, but opensm-devel
package is not found, warn and abort.
Otherwise, doing the best effort: if opensm-devel found,
enable dynamic SL. If it's not found, disable dynamic
SL and build OMPI w/o it.

This commit was SVN r24852.
2011-07-03 14:02:50 +00:00
Matthias Jurenz
5e6919b4e1 Changes to OTF:
otfprofile-mpi:
      - added progress display
      - added verbose messages
      - added functions to sychronize the error indicator to all worker ranks
        (enforces that all ranks will be terminated by calling MPI_Abort if anyone fails)
      - wrap def. comments after 80 characters
      - use pdf[la]tex instead of latex/dvipdf to convert TeX output to PDF
      - added configure checks for pdf[la]tex and PGFPLOTS v1.4
      - fixed function invocation statistics generated from summarized information (--stat)
      - fixed memory leak

Changes to VT:
   MPI wrappers:
      - fixed wrapper generation for MPI implementations which don't support the MPI-2 standard (e.g. MVAPICH, MPICH)
      - corrected IN_PLACE denotation for MPI_Alltoall* and MPI_Scatter*
   vtwrapper:
      - corrected detection of IBM XL's OpenMP flag -qsmp=*:omp:*
   vtunify:
      - fixed faulty cleanup of temporary files which occurred if VT is configured without trace compression support

This commit was SVN r24851.
2011-07-01 07:17:15 +00:00
Brian Barrett
a4b2bd903b * Implement long-ago discussed RFC to add a callback data pointer in the
request completion callback
* Use the completion callback pointer to remove all need for opal_progress
  calls in the one-sided layer

This commit was SVN r24848.
2011-06-30 20:05:16 +00:00
Rolf vandeVaart
e6295159ae Fix compilation of file due to some changes in btl structure.
This commit was SVN r24847.
2011-06-30 19:22:41 +00:00
Yevgeny Kliteynik
a09245e6c8 Changed default behavior when opensm-devel pachege not found - warn, not exit
This commit was SVN r24836.
2011-06-29 14:52:11 +00:00
Yevgeny Kliteynik
b05211148d Supporting dynamic SL (#2674)
- Added enable/disable configuration parameter for dynamic SL
 - All the dynamic SL code is conditionalized
 - Removed libibmad dependency
 - Using only one include - ib_types.h (part of opensm-devel package)
 - Removed all the macro and data types definitions, using the
   existing definitions from ib_types.h instead
 - general cleaning here and there

The async mode is not implemented yet - stay tuned...

This commit was SVN r24830.
2011-06-28 14:28:29 +00:00
Wesley Bland
84be81df95 Standardize the initialization of the EPOCH's.
Everyone will be starting at MIN anyway (until we implement restart of course)
so there's no reason to set the epoch to INVALID and then immediately reset them
to MIN. This way there's less room to make mistakes later.

This commit was SVN r24829.
2011-06-28 14:20:33 +00:00
Jeff Squyres
b05ca7c76b Only build the f90 ext stuff if we're also building the f90 bindings
(e.g., cover the case where the f90 bindings are not being built
because you don't have an f90 compiler)

This commit was SVN r24827.
2011-06-27 20:39:11 +00:00
Jeff Squyres
e32b9a8547 Fix minor typo -- f90, not f77
This commit was SVN r24826.
2011-06-27 20:38:30 +00:00
Jeff Squyres
2b2f6fc580 Fix typo noted by Zbigniew Koza.
This commit was SVN r24818.
2011-06-25 10:35:47 +00:00
Wesley Bland
e1ba09ad51 Add a resilience to ORTE. Allows the runtime to continue after a process (or
ORTED) failure. Note that more work will be necessary to allow the MPI layer to
take advantage of this.

Per RFC:
http://www.open-mpi.org/community/lists/devel/2011/06/9299.php

This commit was SVN r24815.
2011-06-23 20:38:02 +00:00
Brian Barrett
e8817f3f63 * Don't send acks for expected triggered messages; still need to get the rest of the data
* Don't ask for UNLINK events for persistent long unexpected ME or the get MEs.

This commit was SVN r24814.
2011-06-23 16:21:10 +00:00