ports/openmpi
ports/openmpi
1
1
Форкнуть 0
Код Релизы Активность
16 083 коммитов 3 Ветки 0 Теги
Граф коммитов

5405 Коммитов

Автор SHA1 Сообщение Дата
Nathan Hjelm
414677a082 default to no xpmem support 
This commit was SVN r25345.
 2011-10-20 22:13:45 +00:00
Nathan Hjelm
808a73a5c5 removed erroneous add of .deps 
This commit was SVN r25343.
 2011-10-20 21:41:51 +00:00
Nathan Hjelm
3dbaaf6879 initial commit of vader (xpmem) btl 
This commit was SVN r25342.
 2011-10-20 21:39:44 +00:00
Nathan Hjelm
586403f052 more pmi return code wtf 
This commit was SVN r25337.
 2011-10-20 17:53:04 +00:00
Ralph Castain
43e35486a4 Correct flag type - thanks George! 
This commit was SVN r25335.
 2011-10-20 04:00:13 +00:00
Nathan Hjelm
e1e8837992 add a uintptr_t to the seg_key union 
This commit was SVN r25334.
 2011-10-19 21:48:52 +00:00
George Bosilca
78751b3b2d Put back the OPI errors after the ORTE one. 
This commit was SVN r25333.
 2011-10-19 20:57:13 +00:00
Ralph Castain
b44f8d4b28 Complete implementation of the ess.proc_get_locality API. Up to this point, the API was only capable of telling if the specified proc was sharing a node with you. However, the returned value was capable of telling you much more detailed info - e.g., if the proc shares a socket, a cache, or numa node. We just didn't have the data to provide that detail. 
Use hwloc to obtain the cpuset for each process during mpi_init, and share that info in the modex. As it arrives, use a new opal_hwloc_base utility function to parse the value against the local proc's cpuset and determine where they overlap. Cache the value in the pmap object as it may be referenced multiple times.

Thus, the return value from orte_ess.proc_get_locality is a 16-bit bitmask that describes the resources being shared with you. This bitmask can be tested using the macros in opal/mca/paffinity/paffinity.h

Locality is available for all procs, whether launched via mpirun or directly with an external launcher such as slurm or aprun.

This commit was SVN r25331.
 2011-10-19 20:18:14 +00:00
George Bosilca
1bc5da0911 These are supposed to be OPAL level errors. 
This commit was SVN r25326.
 2011-10-19 14:22:09 +00:00
Ralph Castain
72a4b0bd8a Fix constants 
This commit was SVN r25325.
 2011-10-19 14:14:58 +00:00
George Bosilca
a5f24bcdcf The error here is meaningless. 
This commit was SVN r25324.
 2011-10-19 13:04:46 +00:00
George Bosilca
efd88e10d7 Cleanup the error codes. Get rid of all the useless ones, and 
mark the distinction between ORTE and OMPI errors.

This commit was SVN r25323.
 2011-10-19 03:51:53 +00:00
George Bosilca
c453614f8b A more meaningful name for this function (mpi_proc_complete_init 
instead of ompi_proc_set_arch). Change the comment to reflect the
real behavior of the function.

This commit was SVN r25312.
 2011-10-18 02:54:38 +00:00
Ralph Castain
0bf4f48aa3 Don't need priority in this framework 
This commit was SVN r25308.
 2011-10-17 22:39:15 +00:00
Ralph Castain
8f0ef54130 Complete implementation of pmi support. Ensure we support both mpirun and direct launch within same configuration to avoid requiring separate builds. Add support for generic pmi, not just under slurm. Add publish/subscribe support, although slurm's pmi implementation will just return an error as it hasn't been done yet. 
This commit was SVN r25303.
 2011-10-17 20:51:22 +00:00
Ralph Castain
e7f6be5385 Unused variable 
This commit was SVN r25301.
 2011-10-17 18:59:22 +00:00
Ralph Castain
2eaadcfab9 Remove unused variable 
This commit was SVN r25284.
 2011-10-14 15:32:18 +00:00
Vishwanath Venkatesan
8dd07bdceb Removing .ompi_ignore and .ompi_unignore from fs/pvfs2 and fbtl/pvfs2 
This commit was SVN r25283.
 2011-10-14 00:40:11 +00:00
Vishwanath Venkatesan
8f6b29e95b Fixing the default file view issue and merging contiguous lengths and offsets 
for explicit offset case.

This commit was SVN r25281.
 2011-10-13 19:50:45 +00:00
Jeff Squyres
2c6254b70d Second change from Intel. 
This commit was SVN r25279.
 2011-10-12 23:26:34 +00:00
Jeff Squyres
28118d0611 Updte the parameters for the Intel iWARP devices, per request from 
Faisal Latif <faisal.latif@intel.com>.

This commit was SVN r25278.
 2011-10-12 22:58:30 +00:00
Brian Barrett
d8b5b544ad Update list name to match change in spec 
This commit was SVN r25273.
 2011-10-12 20:09:39 +00:00
Rainer Keller
4e6a6fc146 - Check, whether the compiler supports __builtin_clz (count leading 
zeroes);
   if so, use it for bit-operations like opal_cube_dim and opal_hibit.
   Implement two versions of power-of-two.
   In case of opal_next_poweroftwo, this reduces the average execution
   time from 83 cycles to 4 cycles (Intel Nehalem, icc, -O2, inlining,
   measured rdtsc, with loop over 2^27 values).
   Numbers for other functions are similar (but of course heavily depend
   on the usage, e.g. opal_hibit() with a start of 4 does not save
   much).  The bsr instruction on AMD Opteron is also not as fast.

 - Replace various places where the next power-of-two is computed.
   
   Tested on Intel Nehalem Cluster with openib, compilers GNU-4.6.1 and
   Intel-12.0.4 using mpi_testsuite -t "Collective" with 128 processes.

This commit was SVN r25270.
 2011-10-11 22:49:01 +00:00
George Bosilca
74c88a9e48 This was never used (sm_ctl_header). 
This commit was SVN r25267.
 2011-10-11 20:37:00 +00:00
George Bosilca
ca6c282f23 Small cleanups in the SM BTL. 
This commit was SVN r25266.
 2011-10-11 20:32:10 +00:00
George Bosilca
3241bea696 Apply a patch provided by Sébastien Boisvert fixing an issue 
with the probe fairness.

This commit was SVN r25265.
 2011-10-11 20:28:33 +00:00
George Bosilca
4fd78c4683 Keep track of the last probe on each communicator, so we can probe all 
peers in a round-robin fashion. A little bit more fair ...

This commit was SVN r25264.
 2011-10-11 20:24:54 +00:00
Jeff Squyres
951c745590 We always have hwloc xml support (now that it's built into to hwloc 
without needing libxml2).  So OPAL_HAVE_HWLOC_XML is no longer
necessary.  

This commit was SVN r25263.
 2011-10-11 20:20:59 +00:00
George Bosilca
2fefd3a928 Don't forget to move the pointer back by the true_lb. 
This commit was SVN r25262.
 2011-10-11 20:15:49 +00:00
Yevgeny Kliteynik
a503bd53b7 Fixing the librdmacm dependency for build process 
If a user specifically asks for rdmacm support in configure script and
librdmacm (usual and devel) libraries are not found, configure script
would abort.

If a user didn't specify anything, and rdmacm libraries are not found,
configure script will continue after issuing warning message:
"Please install librdmacm and librdmacm-devel or disable rdmacm support"

-- YK

This commit was SVN r25253.
 2011-10-11 09:10:06 +00:00
George Bosilca
649af6c925 Enumerated mixed with another type (int) is tolerated but 
easily fixable.

This commit was SVN r25241.
 2011-10-09 03:54:52 +00:00
Rainer Keller
61ffabed44 - Fix typo in _OMPI_CXX_CHECK_BUILTIN_BACKEND: 
frontend checks have_cxx-variant, yet the backend (if run)
   defined have_-variant; therefore these were never set.

This commit was SVN r25233.
 2011-10-05 08:21:59 +00:00
George Bosilca
07f6ce235f Return an OMPI_ error not an ORTE_. 
This commit was SVN r25232.
 2011-10-04 14:57:24 +00:00
George Bosilca
ce7935c8fa Obviously these were not needed. 
This commit was SVN r25231.
 2011-10-04 14:56:34 +00:00
George Bosilca
80c02647c8 Each level (OPAL/ORTE/OMPI) should only return it's own constants, 
instead of the current mismatch.

This commit was SVN r25230.
 2011-10-04 14:50:31 +00:00
Mike Dubman
7a9ae43276 added support for shared memory transport in mxm 
This commit was SVN r25220.
 2011-10-03 12:59:55 +00:00
Brian Barrett
fc29ffebdb * remove two aborts that aren't necessary 
This commit was SVN r25214.
 2011-09-29 22:27:23 +00:00
Brian Barrett
14f32a1a54 * Clean up progress function 
* Only print returnable errors when verbose=1.  Still print errors when
  we're going to abort, since those obviously aren't returnable

This commit was SVN r25213.
 2011-09-29 22:26:33 +00:00
Brian Barrett
758f8a4d87 * More debugging output 
* Make recv short block events use the callback mechanism so that can
  add overflow debugging

This commit was SVN r25212.
 2011-09-29 21:59:48 +00:00
Brian Barrett
c08ea5c0f5 Set options correctly for the two pts 
This commit was SVN r25211.
 2011-09-29 21:56:37 +00:00
Brian Barrett
05f800abae Properly unpack data for long unexpected 
This commit was SVN r25210.
 2011-09-29 17:25:45 +00:00
Rolf vandeVaart
3d8c6b83a9 Make some error messages more helpful 
This commit was SVN r25209.
 2011-09-29 16:32:46 +00:00
Brian Barrett
bb9e73232a * Leverage hdr_data and opcount to improve debugging 
* Clean up handling of short synchronous messages

This commit was SVN r25208.
 2011-09-28 21:18:47 +00:00
Brian Barrett
71d8300607 * Fix name clash with macros in mtl_portals4.h 
* hdr_data now includes opcount and length for all messages, which is the match
  bits for long and rndv messages
* Re-add probe implementation 

This commit was SVN r25207.
 2011-09-28 16:53:01 +00:00
Brian Barrett
2fb8045fad clean up printfs 
This commit was SVN r25206.
 2011-09-28 15:28:46 +00:00
Brian Barrett
26e781f002 * Remove triggered code for now 
* Move from per-endpoint send/recv count to just send side op count

This commit was SVN r25205.
 2011-09-28 15:25:39 +00:00
Brian Barrett
592c1ab6db * revert probe and size information changes, since it seems to break everything 
This commit was SVN r25204.
 2011-09-28 14:57:19 +00:00
Brian Barrett
211b5c7824 * Make triggered protocol only work for non-wildcard receives 
* Always encode length in header data to make probe work
* General send/receive cleanups
* Implement iprobe

This commit was SVN r25197.
 2011-09-27 22:45:00 +00:00
Brian Barrett
77c560be42 updates to match new api changes 
This commit was SVN r25196.
 2011-09-27 20:38:22 +00:00
Brad Benton
0f2475c554 Modified set_remote_info() to use memmove() instead of memcpy() when 
copying rem_qp info.  This avoids potential errors when src & dest overlap.
This is a workaround for the issue in #2871

This commit was SVN r25180.
 2011-09-26 20:07:36 +00:00