1
1
Граф коммитов

205 Коммитов

Автор SHA1 Сообщение Дата
George Bosilca
15b18cd2cf Make CMA compile and run.
This commit was SVN r27873.
2013-01-19 14:27:54 +00:00
Brian Barrett
579cf4adcd After discussion with Jeff, don't do C++ inline assembly (there is a non-inline
version still avaiable for C++).  This is yet another push to try to make
OPAL a C only interface...

This commit was SVN r27828.
2013-01-15 17:04:42 +00:00
Ralph Castain
ee6c7702d2 Ensure the cma.h file is included in the tarball
This commit was SVN r27235.
2012-09-04 19:34:09 +00:00
Shiqing Fan
42dfbc7d2f Another CMake scripts update for:
correctly generate hwloc library
automatically define OMPI/OPAL/ORTE_OMPORTS for user applications
update the f77 bindings

This commit was SVN r26893.
2012-07-27 11:49:09 +00:00
Shiqing Fan
5d81c27282 Update the CMake files for Fortran 77 bindings, get ready for F90 bindings.
Change several variable names and update the macros.

This commit was SVN r26851.
2012-07-24 08:49:34 +00:00
Shiqing Fan
8c4a3e1269 correct the symbol dllexports for windows build
This commit was SVN r26827.
2012-07-22 08:54:50 +00:00
Christopher Yeoh
9c353cf9d8 This adds a file that was missed in r26134 (adds support for Cross Memory
Attach). It is a header file that contains syscall defs for process_vm_readv
and process_vm_writev. It is only used on systems where glibc does not yet
have support for the new syscalls and where --with-cma has been passed 
to configure. The syscall numbers are hardcoded but have been in a released
kernel and so will not change in the future

Once all linux distros have the new glibc this file can be removed.

This commit was SVN r26615.

The following SVN revision numbers were found above:
  r26134 --> open-mpi/ompi@524de80eaa
2012-06-18 05:01:09 +00:00
Shiqing Fan
aa6cde9886 Change f77 to fortran for the rest of windows build files.
This commit was SVN r26558.
2012-06-06 14:09:51 +00:00
Jeff Squyres
2ba10c37fe Per RFC, bring in the following changes:
* Remove paffinity, maffinity, and carto frameworks -- they've been
   wholly replaced by hwloc.
 * Move ompi_mpi_init() affinity-setting/checking code down to ORTE.
 * Update sm, smcuda, wv, and openib components to no longer use carto.
   Instead, use hwloc data.  There are still optimizations possible in
   the sm/smcuda BTLs (i.e., making multiple mpools).  Also, the old
   carto-based code found out how many NUMA nodes were ''available''
   -- not how many were used ''in this job''.  The new hwloc-using
   code computes the same value -- it was not updated to calculate how
   many NUMA nodes are used ''by this job.''
   * Note that I cannot compile the smcuda and wv BTLs -- I ''think''
     they're right, but they need to be verified by their owners.
 * The openib component now does a bunch of stuff to figure out where
   "near" OpenFabrics devices are.  '''THIS IS A CHANGE IN DEFAULT
   BEHAVIOR!!''' and still needs to be verified by OpenFabrics vendors
   (I do not have a NUMA machine with an OpenFabrics device that is a
   non-uniform distance from multiple different NUMA nodes).
 * Completely rewrite the OMPI_Affinity_str() routine from the
   "affinity" mpiext extension.  This extension now understands
   hyperthreads; the output format of it has changed a bit to reflect
   this new information.
 * Bunches of minor changes around the code base to update names/types
   from maffinity/paffinity-based names to hwloc-based names.
 * Add some helper functions into the hwloc base, mainly having to do
   with the fact that we have the hwloc data reporting ''all''
   topology information, but sometimes you really only want the
   (online | available) data.

This commit was SVN r26391.
2012-05-07 14:52:54 +00:00
Jeff Squyres
c30d1ef0df Patch from Evan Clinton, reviewed by Leif Lindholm, for supporting
ARM5 and ARM6.

This commit was SVN r26361.
2012-04-30 20:49:55 +00:00
Nathan Hjelm
e84f9ec8c3 don't define OPAL_HAVE_ATOMIC_SWAP_64/32 in amd/atomic.h unless we have inlined assembly. fixes pgi complilation on XE/XK-6
This commit was SVN r26343.
2012-04-26 20:43:30 +00:00
Jeff Squyres
b6a90434e4 Fix some include file header ordering issues for some BSDs, suggested
by Paul Hargrove.

This commit was SVN r25984.
2012-02-21 13:32:14 +00:00
Jeff Squyres
63a96e92b5 In a recent v1.5 branch issue, it took a while to figure out that
paffinity hwloc was returning "NOT_SUPPORTED" when the real problem
was that the underlying hwloc simply hadn't been initialized yet.  So
let's clearly delineate this case: return OPAL_ERR_NOT_INITIALIZED if
the underlying hwloc is not initialized.

This commit was SVN r25902.
2012-02-10 18:29:52 +00:00
Brian Barrett
2bb447c804 * Shouldn't have a timer header for sync_builtin, since it doesn't actually
have timer support
* Default timer size should be a long, not an int.  Int will roll over way
  too fast, with no performance benifit on 64 bit machines...

This commit was SVN r25501.
2011-11-23 17:05:01 +00:00
Brian Barrett
5cd5ef623d Fix compatibility implementation of swap. Turns out that you shouldn't test
the compatibility code on a platform which has a native swap.  Sorry to all!

This commit was SVN r25500.
2011-11-23 16:28:00 +00:00
Brian Barrett
f971a541f1 Implement swap in terms of compare and swap if it isn't implemented directly
This commit was SVN r25499.
2011-11-23 05:57:52 +00:00
Brian Barrett
86f555121c Add (optional/last ditch effort) support for GCC/Intel __sync_ builtin atomic
operations.  Much easier than adding support for a new architecture.

This commit was SVN r25498.
2011-11-23 04:25:41 +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
Nathan Hjelm
ad9005820f fixed typo in last commit
This commit was SVN r25306.
2011-10-17 21:35:22 +00:00
Nathan Hjelm
e6ead53eef add opal_atomic_swap_xx for amd64
This commit was SVN r25305.
2011-10-17 21:33:44 +00:00
Brian Barrett
431f1b6f8c Remove long-dead sparc support. Sparc (v8) has not been a supported platform
since the 1.4 release (and configure would abort when run with sparc v8), but
the code was left in place.  Sparc v9 (32 or 64 bit) are still supported
targets.

This commit was SVN r25258.
2011-10-11 18:46:06 +00:00
Brian Barrett
98e98ce2c5 * opal_atomic_trylock is documented to return 0 if the lock was acquired,
1 otherwise.  It was doing the opposite, so this patch fixes the
  return values.  All uses (all in ORTE) used the actual return values,
  not the documented values, so fix them as well.

This commit was SVN r25257.
2011-10-11 18:43:45 +00:00
Jeff Squyres
f539b20a8f Patch from ARM for assembly:
http://www.open-mpi.org/community/lists/devel/2011/08/9586.php

This commit was SVN r24979.
2011-08-02 19:15:24 +00:00
Jeff Squyres
ceabe91484 Yow; we forgot to include the ARM stuff in the tarball. :-(
This commit was SVN r24875.
2011-07-11 23:52:07 +00:00
Ralph Castain
f3cae3d6f3 Cleanup the handling of if_include and if_exclude arguments based on CIDR notation.
Fix a bug in the new code that prevented the system from correctly matching addresses.

Remove comments in the show-help text indicating that we would continue in the face of incorrect specifications - leave that to the calling layer to decide.

Modify the new opal_ifmatches so it returns error codes letting the caller better understand the result.

Modify the oob to ensure we abort if we don't find interfaces matching specified constraints, and that we do so without multiple error messages.

NOTE: we have a conflict in our standards. We have been using comma-delimited lists of interfaces for all our params. However, one param - opal_net_private_ipv4 - now uses semicolons instead of comma separators. No idea why, but it is confusing.

This commit was SVN r24755.
2011-06-07 02:09:11 +00:00
Shiqing Fan
4490fdbd34 Add the initial support for MinGW and MSYS.
Correctly check the dependencies of MSYS env.
Set up configure include and lib path for building the package.
update a few more CMake scripts.

This commit was SVN r24663.
2011-04-29 14:42:07 +00:00
Jeff Squyres
58a13f87e6 Oops -- forgot to add opal_config_top.h to Makefile.am (so that it'll
be included in the tarball).

This commit was SVN r24572.
2011-03-25 01:21:11 +00:00
Jeff Squyres
5ae1b15b6e Ensure that other packages defining PACKAGE_ macros don't hurt us, and protect others from our PACKAGE_ macros.
This commit was SVN r24571.
2011-03-24 22:39:56 +00:00
Eugene Loh
2770a12beb Continue clean up of thread options started in r22841, 22842, and 22849.
No need for any CMRs to 1.5... that was already done in CMR 2728.

This commit was SVN r24545.

The following SVN revision numbers were found above:
  r22841 --> open-mpi/ompi@b400b84162
2011-03-18 21:36:35 +00:00
Ralph Castain
7eede54b39 Solve a problem when cross-compiling for PPC32 - in this case, OPAL_HAVE_ATOMIC_CMPSET_64 is not set, but the code requires that the ADD_64 and SUB_64 values at least be defined.
This commit was SVN r24528.
2011-03-15 15:50:49 +00:00
George Bosilca
f981e02b4a Fix a typo and correct the usage of the defines.
This commit was SVN r24454.
2011-02-24 06:34:30 +00:00
George Bosilca
f79c87f0c3 Correct the assembly using xaddl for IA32.
Add atomic functions for add and sub 32 and 64 bits for AMD64.

This commit was SVN r24453.
2011-02-24 06:31:47 +00:00
Jeff Squyres
3f4d4886f2 Minor update for something that has been bugging me for quite a while:
OMPI supports multiple different repository systems (SVN, hg, git).
But the VERSION file has listed "want_svn" and "svn_r" as fields, even
though the actual repo system and version may not be SVN.

So search/replace those fields (and derrivative values that come from
those fields) with "want_repo_rev" and "repo_rev", respectively.

This commit was SVN r24405.
2011-02-16 22:53:23 +00:00
Jeff Squyres
511f87665b Fixes trac:2680: Add ARM support.
This commit was SVN r24308.

The following Trac tickets were found above:
  Ticket 2680 --> https://svn.open-mpi.org/trac/ompi/ticket/2680
2011-01-26 17:22:44 +00:00
George Bosilca
b4355408f5 Fix the Sparc and Sparcv9 atomics based on Nicolai Stange
patch.

CMR:v1.5
CMR:v1.4

This commit was SVN r24150.
2010-12-03 19:16:53 +00:00
Shiqing Fan
f43862420c Convert the bad dos line endings to unix style for all windows related files.
This commit was SVN r24137.
2010-12-02 12:08:08 +00:00
Shiqing Fan
39c9f7468e Add support for managing priorities of windows mca components.
Correct the generated strings in mpi.h.

This commit was SVN r24082.
2010-11-23 19:09:06 +00:00
George Bosilca
96abaf2e17 Pushing the Debian patch (based on Manuel Prinz modifications).
This commit was SVN r24061.
2010-11-17 02:36:03 +00:00
Rolf vandeVaart
37d5267895 The fix for ticket #2560 was somehow removed in the
great autogen update.  Therefore, put them back.

This commit was SVN r24053.
2010-11-15 21:41:56 +00:00
Shiqing Fan
482a621e31 Change the behavior of exporting/importing symbols on Windows, so that to fit the new build procedure, i.e. import statically linked opal/orte libraries for other libraries/binaries. There are several use cases when creating dll libraries:
1. create DLL A, export symbols of A, import nothing  (A normally is OPAL)
   should define _USRDLL , A_EXPORT 

2. create DLL B, export symbols of B, import A.lib    (B could be ORTE, OMPI or other ompi tools)
   should define _USRDLL, B_EXPORT

3. create DLL C, import B.dll    (C could be external libs or apps)
   should define B_IMPORT

This commit was SVN r24016.
2010-11-09 16:13:30 +00:00
Shiqing Fan
7bac326920 Fix Windows build, add custom command to generate static libraries (opal and orte) for shared build.
This commit was SVN r24012.
2010-11-09 08:32:45 +00:00
Ralph Castain
fceabb2498 Update libevent to the 2.0 series, currently at 2.0.7rc. We will update to their final release when it becomes available. Currently known errors exist in unused portions of the libevent code. This revision passes the IBM test suite on a Linux machine and on a standalone Mac.
This is a fairly intrusive change, but outside of the moving of opal/event to opal/mca/event, the only changes involved (a) changing all calls to opal_event functions to reflect the new framework instead, and (b) ensuring that all opal_event_t objects are properly constructed since they are now true opal_objects.

Note: Shiqing has just returned from vacation and has not yet had a chance to complete the Windows integration. Thus, this commit almost certainly breaks Windows support on the trunk. However, I want this to have a chance to soak for as long as possible before I become less available a week from today (going to be at a class for 5 days, and thus will only be sparingly available) so we can find and fix any problems.

Biggest change is moving the libevent code from opal/event to a new opal/mca/event framework. This was done to make it much easier to update libevent in the future. New versions can be inserted as a new component and tested in parallel with the current version until validated, then we can remove the earlier version if we so choose. This is a statically built framework ala installdirs, so only one component will build at a time. There is no selection logic - the sole compiled component simply loads its function pointers into the opal_event struct.

I have gone thru the code base and converted all the libevent calls I could find. However, I cannot compile nor test every environment. It is therefore quite likely that errors remain in the system. Please keep an eye open for two things:

1. compile-time errors: these will be obvious as calls to the old functions (e.g., opal_evtimer_new) must be replaced by the new framework APIs (e.g., opal_event.evtimer_new)

2. run-time errors: these will likely show up as segfaults due to missing constructors on opal_event_t objects. It appears that it became a typical practice for people to "init" an opal_event_t by simply using memset to zero it out. This will no longer work - you must either OBJ_NEW or OBJ_CONSTRUCT an opal_event_t. I tried to catch these cases, but may have missed some. Believe me, you'll know when you hit it.

There is also the issue of the new libevent "no recursion" behavior. As I described on a recent email, we will have to discuss this and figure out what, if anything, we need to do.

This commit was SVN r23925.
2010-10-24 18:35:54 +00:00
Ralph Castain
29b16cc800 Add missing include
This commit was SVN r23892.
2010-10-15 04:04:20 +00:00
Jeff Squyres
66d15035ab Replace some intentional-segv's with abort(). Seems safer and doesn't
cause all kinds of compiler warnings.

This commit was SVN r23889.
2010-10-14 22:01:14 +00:00
Jeff Squyres
73bcc4a36b Fix mistake that came in via the ompi-agen tree in r23764. The mistake wasn't part of the core autogen upgrade; it was an additional 'bonus' cleanup. Oops. The mistake will always create a set of directories under installdir, even if you do not --with-devel-headers. The set of directories will be empty, but still -- they should not be there at all. This commit fixes that -- the directories are not created at all if you do not --with-devel-headers
This commit was SVN r23801.

The following SVN revision numbers were found above:
  r23764 --> open-mpi/ompi@40a2bfa238
2010-09-24 22:53:28 +00:00
Ralph Castain
40a2bfa238 WARNING: Work on the temp branch being merged here encountered problems with bugs in subversion. Considerable effort has gone into validating the branch. However, not all conditions can be checked, so users are cautioned that it may be advisable to not update from the trunk for a few days to allow MTT to identify platform-specific issues.
This merges the branch containing the revamped build system based around converting autogen from a bash script to a Perl program. Jeff has provided emails explaining the features contained in the change.

Please note that configure requirements on components HAVE CHANGED. For example. a configure.params file is no longer required in each component directory. See Jeff's emails for an explanation.

This commit was SVN r23764.
2010-09-17 23:04:06 +00:00
Rolf vandeVaart
ef8090ec71 Fix the ia32 atomic add and subtract functions so they
do the right thing.  They now properly return
the value after the update.  This also fixes all warnings
reported by the Sun Studio compiler.  George provided the
new assembly routines.  I added some configure code to make
sure the compilers could handle it.

This fixes trac:2560.

This commit was SVN r23721.

The following Trac tickets were found above:
  Ticket 2560 --> https://svn.open-mpi.org/trac/ompi/ticket/2560
2010-09-08 10:47:15 +00:00
Rainer Keller
5eb571c458 - As suggested in CMR #2558, attribute-macros should be
be tested on function pointers and assigned accordingly,
   instead of using the pre-processor in the header files.

   A functional change is (re-) specifying __opal_attribute_noreturn__
   on orte_errmgr_base_abort(): All modules in the errmgr framework
   either use this function, or define their own abort function,
   which sets __opal_attribute_noreturn__.
   This attributes was taken out with the errmgr overhaul in r22872.

This commit was SVN r23689.

The following SVN revision numbers were found above:
  r22872 --> open-mpi/ompi@e4f2d03d28
2010-08-31 10:28:51 +00:00
Ethan Mallove
f42c2a737f Fixes trac:2532 - "MPI_Put can result in SIGBUS on SPARC"
Reviewed by Rolf V and Brian B

This commit was SVN r23649.

The following Trac tickets were found above:
  Ticket 2532 --> https://svn.open-mpi.org/trac/ompi/ticket/2532
2010-08-24 18:10:43 +00:00
Shiqing Fan
b8db8d0ef8 Need to change another variable name.
This commit was SVN r23556.
2010-08-05 12:38:28 +00:00