1
1
Граф коммитов

78 Коммитов

Автор SHA1 Сообщение Дата
Jeff Squyres
6557d74e01 Make sure we get the entire hwloc tree, including IO devices.
This commit was SVN r25887.
2012-02-09 16:59:38 +00:00
Ralph Castain
a3ab70c53f Correctly parse socket:core syntax in rankfile
This commit was SVN r25848.
2012-02-01 01:50:05 +00:00
Ralph Castain
3f31feee6f Handle the case where a user's rankfile specifies only cpus, and not socket:cpu pairs.
This commit was SVN r25803.
2012-01-27 12:21:45 +00:00
Jeff Squyres
ecf6ba910c Silence a few icc warnings and about mixing enums with other types.
This commit was SVN r25560.
2011-12-02 13:18:54 +00:00
Jeff Squyres
6fbbfd0f7a Gah! r25545 acidentally included ''waaaay'' more stuff than it was
supposed to.  I.e., half-baked/not complete stuff.

This commit backs out all of r25545.  Sorry folks!

This commit was SVN r25546.

The following SVN revision numbers were found above:
  r25545 --> open-mpi/ompi@7f9ae11faf
2011-11-29 23:24:52 +00:00
Jeff Squyres
7f9ae11faf Per http://www.open-mpi.org/community/lists/users/2011/11/17862.php,
to make MPI_IN_PLACE (and other sentinel Fortran constants) work on OS
X, we need to use the following compiler (linker) flag:

    -Wl,-commons,use_dylibs 

So if we're compiling on OS X, test to see if that flag works with the
compiler.  If so, add it to the wrapper FFLAGS and FCFLAGS (note that
per a future update, we'll only have one Fortran compiler anyway).

Fixes trac:1982.  

This commit was SVN r25545.

The following Trac tickets were found above:
  Ticket 1982 --> https://svn.open-mpi.org/trac/ompi/ticket/1982
2011-11-29 23:05:54 +00:00
George Bosilca
61f273b987 Do not tolerate uninitialized variables.
This commit was SVN r25489.
2011-11-18 10:19:24 +00:00
Ralph Castain
6310361532 At long last, the fabled revision to the affinity system has arrived. A more detailed explanation of how this all works will be presented here:
https://svn.open-mpi.org/trac/ompi/wiki/ProcessPlacement

The wiki page is incomplete at the moment, but I hope to complete it over the next few days. I will provide updates on the devel list. As the wiki page states, the default and most commonly used options remain unchanged (except as noted below). New, esoteric and complex options have been added, but unless you are a true masochist, you are unlikely to use many of them beyond perhaps an initial curiosity-motivated experimentation.

In a nutshell, this commit revamps the map/rank/bind procedure to take into account topology info on the compute nodes. I have, for the most part, preserved the default behaviors, with three notable exceptions:

1. I have at long last bowed my head in submission to the system admin's of managed clusters. For years, they have complained about our default of allowing users to oversubscribe nodes - i.e., to run more processes on a node than allocated slots. Accordingly, I have modified the default behavior: if you are running off of hostfile/dash-host allocated nodes, then the default is to allow oversubscription. If you are running off of RM-allocated nodes, then the default is to NOT allow oversubscription. Flags to override these behaviors are provided, so this only affects the default behavior.

2. both cpus/rank and stride have been removed. The latter was demanded by those who didn't understand the purpose behind it - and I agreed as the users who requested it are no longer using it. The former was removed temporarily pending implementation.

3. vm launch is now the sole method for starting OMPI. It was just too darned hard to maintain multiple launch procedures - maybe someday, provided someone can demonstrate a reason to do so.

As Jeff stated, it is impossible to fully test a change of this size. I have tested it on Linux and Mac, covering all the default and simple options, singletons, and comm_spawn. That said, I'm sure others will find problems, so I'll be watching MTT results until this stabilizes.

This commit was SVN r25476.
2011-11-15 03:40:11 +00:00
Jeff Squyres
4fe26b0392 Fix some minor memory leaks
This commit was SVN r25410.
2011-11-01 20:22:26 +00:00
Ralph Castain
4368199c86 Missing include
This commit was SVN r25402.
2011-10-31 13:39:57 +00:00
Ralph Castain
96332a2859 Fix typo
This commit was SVN r25400.
2011-10-30 13:23:42 +00:00
Ralph Castain
71ed8e3cd3 Bring back the local node's binding capabilities along with its topology. Clean up indentation.
This commit was SVN r25399.
2011-10-30 13:20:16 +00:00
Ralph Castain
7ba4675adf Bring over some useful utilities and definitions for working with hwloc inside ORTE/OMPI. Cache frequently computed info to save processing time when handling multiple nodes with the same topology. Deal with available cpus as defined by online vs allowed vs user-specified limits. Help deal with hwloc's unfortunate decision to lump all caches in the same object type.
This commit was SVN r25393.
2011-10-29 14:58:58 +00:00
Ralph Castain
21d45b0807 Just some cleanup in case of error
This commit was SVN r25387.
2011-10-29 01:55:19 +00:00
Ralph Castain
a7cbc25658 Minor cleanups - check hwloc returns everywhere. Thanks to Chris Yeoh for pointing this out.
This commit was SVN r25360.
2011-10-24 14:05:26 +00:00
Shiqing Fan
5711414eb7 Fix Windows build
This commit was SVN r25351.
2011-10-21 14:46:58 +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
Ralph Castain
69a0882207 Correctly setup hwloc when passing a topology from an external source
This commit was SVN r25277.
2011-10-12 21:34:46 +00:00
Jeff Squyres
ff97b57c90 Change the names to be slightly more descriptive.
This commit was SVN r25271.
2011-10-12 16:07:09 +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
Shiqing Fan
4caed984ed Need to exclude another file for windows build.
This commit was SVN r25168.
2011-09-20 16:09:03 +00:00
Ralph Castain
51129cc2a8 If built without hwloc xml support, we cannot currently pass the local topology from the daemon to an MPI app. This makes it impossible to set affinity, for example. In this case, have the app get its own copy of the topology at startup.
For safety sake, protect hwloc-based affinity modules from NULL topology

This commit was SVN r25164.
2011-09-20 14:46:55 +00:00
Jeff Squyres
9db4542c2b Move maffinity_base_alloc_policy and
maffinity_base_bind_failure_action MCA params to the hwloc base
(hwloc_base_alloc_polocy and hwloc_base_bind_failure_action).  Since
these MCA parameters were never on a release branch, I'm just
moving/renaming them outright and not leaving aliases to the old
names.

Note that some upper layer needs to call
opal_hwloc_base_set_process_membind_policy() to set the
set-by-MCA-param process-wide memory affinity policy.  We can't do
this automatically during hwloc_base_open() because, for reasons
described elsewhere, opal_hwloc_topology is not automatically filled
during hwloc_base_open() (in short: potential scalability issues when
launching many MPI processes simultaneously on a single machine, for
example).

This commit was SVN r25156.
2011-09-19 16:10:37 +00:00
Jeff Squyres
ecd603256a * Rename opal_hwloc_components to opal_hwloc_base_components
* Fix some comments

This commit was SVN r25150.
2011-09-17 11:54:36 +00:00
Jeff Squyres
d6682523f6 Put in proper basename so that "make dist" can find it.
This commit was SVN r25135.
2011-09-13 11:09:56 +00:00
Shiqing Fan
b61eed801f Fix the problem of building hwloc on Windows. Temporarily not using it for Windows.
This commit was SVN r25128.
2011-09-12 13:55:34 +00:00
Ralph Castain
6460fe5480 Silence warning
This commit was SVN r25127.
2011-09-12 13:32:21 +00:00
Ralph Castain
92c7372e20 Per the RFC from Jeff, move hwloc from opal/mca/common to its own static framework ala libevent. Have ORTE daemons collect the topology info at startup and, if --enable-hwloc-xml is set, send that info back to the HNP for later use. The HNP only retains unique topology "templates" to reduce memory footprint. Have the daemon include the local topology info in the nidmap buffer sent to each app so the apps don't all hammer the local system to discover it for themselves.
Remove the sysinfo framework as hwloc replaces that functionality.

This commit was SVN r25124.
2011-09-11 19:02:24 +00:00