Logo AND Algorithmique Numérique Distribuée

Public GIT Repository
simgrid.git
6 years agoFix dead assignments.
Arnaud Giersch [Fri, 9 May 2014 12:09:41 +0000 (14:09 +0200)]
Fix dead assignments.

That's when a value stored in a variable is never used.

6 years agoAdd check for error.
Arnaud Giersch [Fri, 9 May 2014 11:11:38 +0000 (13:11 +0200)]
Add check for error.

6 years agoFix warning about duplicate typedef.
Arnaud Giersch [Fri, 9 May 2014 11:10:53 +0000 (13:10 +0200)]
Fix warning about duplicate typedef.

6 years agoMerge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Stéphane Castelli [Mon, 12 May 2014 07:07:35 +0000 (09:07 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid

6 years agoFix use of wrong variable
Stéphane Castelli [Mon, 12 May 2014 07:06:55 +0000 (09:06 +0200)]
Fix use of wrong variable

6 years agoAdded a --cfg="smpi/simulation_computation" to smpirun. The default
Henri Casanova [Fri, 9 May 2014 01:17:37 +0000 (15:17 -1000)]
Added a --cfg="smpi/simulation_computation" to smpirun. The default
value is "yes", meaning that SMPI will benchmark computational parts
of the simulation application and simulate them using smpi_execute(), as
it's always been done in the past.

Disabling the simulation of computation is useful when smpirunning an "application"
that is in fact doing a "live replay" of another MPI app (e.g., ScalaTrace's replay tool, various on-line simulators that run an app at scale). In this case the computation due to the replay logic
should be be simulated by SMPI.

For the time being, for smpirunning such an "application", the approach is:
1. use --cfg="smpi/simulation_computation:no"
2. add calls to smpi_execute() in the "application"

6 years agoMerge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Henri Casanova [Thu, 8 May 2014 20:12:11 +0000 (10:12 -1000)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid

6 years agoRewrote smpi_sleep() and smpi_usleep() so that:
Henri Casanova [Thu, 8 May 2014 20:10:44 +0000 (10:10 -1000)]
Rewrote smpi_sleep() and smpi_usleep() so that:
1. the code is not duplicated but shared in a static private function
2. sleeping is done via an actual simcall_process_sleep rather than a bogus computation

Augmented SMPI tracing to allow tracing of sleep events

6 years agoforgot a comma ...
suter [Thu, 8 May 2014 20:10:02 +0000 (22:10 +0200)]
forgot a comma ...

6 years agoMerge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot//simgrid/simgrid
suter [Thu, 8 May 2014 19:36:00 +0000 (21:36 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot//simgrid/simgrid

6 years agoadd color for "test", not sure it's mandatory, but it won't hurt
suter [Thu, 8 May 2014 19:35:43 +0000 (21:35 +0200)]
add color for "test", not sure it's mandatory, but it won't hurt

6 years agoupdate the ignore patterns after the recent source reorganizations
Martin Quinson [Thu, 8 May 2014 10:41:38 +0000 (12:41 +0200)]
update the ignore patterns after the recent source reorganizations

6 years agoMerge branch 'master' of scm.gforge.inria.fr:/gitroot/simgrid/simgrid
Martin Quinson [Thu, 8 May 2014 08:44:11 +0000 (10:44 +0200)]
Merge branch 'master' of scm.gforge.inria.fr:/gitroot/simgrid/simgrid

6 years agoModified smpi_sleep() so that it returns 0, instead of the number of seconds
Henri Casanova [Thu, 8 May 2014 03:09:38 +0000 (17:09 -1000)]
Modified smpi_sleep() so that it returns 0, instead of the number of seconds
it was passed in (not that people check that return value much, but...)

6 years agoAdded a smpi_usleep function and usleep -> smpi_usleep macro
Henri Casanova [Thu, 8 May 2014 03:07:13 +0000 (17:07 -1000)]
Added a smpi_usleep function and usleep -> smpi_usleep macro

6 years agoModify one of the replay examples to use the 'test' action. the test on
suter [Wed, 7 May 2014 23:53:18 +0000 (01:53 +0200)]
Modify one of the replay examples to use the 'test' action. the test on
process of rank 1 fails (send not posted yet) while that on process of
rank 2 succeeds (the wait then becomes a noop action)

6 years agoadd support of MPI_Test in smpi_replay and TRACING-TI
suter [Wed, 7 May 2014 23:49:13 +0000 (01:49 +0200)]
add support of MPI_Test in smpi_replay and TRACING-TI

6 years agodocument smpi global variable privatization
Augustin Degomme [Wed, 7 May 2014 14:46:39 +0000 (16:46 +0200)]
document smpi global variable privatization

6 years agoFix warning introduced by previous commit
Stéphane Castelli [Wed, 7 May 2014 14:17:50 +0000 (16:17 +0200)]
Fix warning introduced by previous commit

6 years agoMerge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Stéphane Castelli [Wed, 7 May 2014 13:38:08 +0000 (15:38 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid

6 years agoCompletion of the routing for fat trees
Stéphane Castelli [Wed, 7 May 2014 13:24:34 +0000 (15:24 +0200)]
Completion of the routing for fat trees

6 years agoAdd some routing for the fat trees
Stéphane Castelli [Wed, 7 May 2014 11:54:04 +0000 (13:54 +0200)]
Add some routing for the fat trees
Unsign some ints

6 years agoremove file introduced in 4a8e5b1e8
Augustin Degomme [Wed, 7 May 2014 09:26:54 +0000 (11:26 +0200)]
remove file introduced in 4a8e5b1e8

6 years agoChangelog update
Augustin Degomme [Wed, 7 May 2014 09:25:08 +0000 (11:25 +0200)]
Changelog update

6 years agoFix leak in process_join
Paul Bédaride [Wed, 7 May 2014 09:24:57 +0000 (11:24 +0200)]
Fix leak in process_join

6 years agobetter to add this file ...
suter [Wed, 7 May 2014 02:10:06 +0000 (04:10 +0200)]
better to add this file ...

6 years agoproperly add a smpicxx "compiler"
suter [Wed, 7 May 2014 01:37:52 +0000 (03:37 +0200)]
properly add a smpicxx "compiler"

6 years agoallow MPI codes to use timezone in their gettimeofday (but we replace it
suter [Wed, 7 May 2014 01:09:51 +0000 (03:09 +0200)]
allow MPI codes to use timezone in their gettimeofday (but we replace it
to NULL anyway). The previous macro prevented the compilation of some
code.

6 years agoAdded a definition of MPI_File (MPI-IO) to allow smpicc to compile code that declare...
Henri Casanova [Tue, 6 May 2014 23:09:23 +0000 (13:09 -1000)]
Added a definition of MPI_File (MPI-IO) to allow smpicc to compile code that declare MPI_File variables.

6 years agoSome updates in ChangeLog.
Arnaud Giersch [Tue, 6 May 2014 13:50:23 +0000 (15:50 +0200)]
Some updates in ChangeLog.

6 years agohave barrier_wait return XBT/MSG_BARRIER_SERIAL_PROCESS for one process
Augustin Degomme [Tue, 6 May 2014 16:09:39 +0000 (18:09 +0200)]
have barrier_wait return XBT/MSG_BARRIER_SERIAL_PROCESS for one process
Also have it reset the counter when barrier is completed, to allow reuse

6 years agofix teshes
Augustin Degomme [Tue, 6 May 2014 15:25:13 +0000 (17:25 +0200)]
fix teshes

6 years agoactually change the value as said in previous commit
Augustin Degomme [Tue, 6 May 2014 14:58:09 +0000 (16:58 +0200)]
actually change the value as said in previous commit

6 years agochange default value of maxmin/precision to 1e-3 for smpi. Fix bug [#17132]
Augustin Degomme [Tue, 6 May 2014 14:55:59 +0000 (16:55 +0200)]
change default value of maxmin/precision to 1e-3 for smpi. Fix bug [#17132]
The default one was uselessly low, since the actual computation precision is now surf/precision.
This lead to some bug on 64 bit systems when comparing large values (>1M) with insanely high precision (1e-9).

6 years agorequalify tesh to account for new finalize
Augustin Degomme [Tue, 6 May 2014 13:46:21 +0000 (15:46 +0200)]
requalify tesh to account for new finalize

6 years agoUse the xbt barrier to finalize smpi processes cleanly.
Augustin Degomme [Tue, 6 May 2014 13:41:53 +0000 (15:41 +0200)]
Use the xbt barrier to finalize smpi processes cleanly.
This removes the MPI_Barrier that was used before, which might have caused problems in some cases

6 years agoadd msg and xbt versions of barrier.
Augustin Degomme [Tue, 6 May 2014 13:31:00 +0000 (15:31 +0200)]
add msg and xbt versions of barrier.
MSG/xbt_barrier_init(unsigned int count) creates the barrier object for all processes
MSG/xbt_barrier_wait(msg/xbt_bar_t bar) waits on it
MSG/xbt_barrier_destroy(msg/xbt_bar_t bar) deletes the object (only one process should call it)

6 years agotypo --
Augustin Degomme [Mon, 5 May 2014 13:49:43 +0000 (15:49 +0200)]
typo --

6 years agotypo --
Augustin Degomme [Mon, 5 May 2014 12:43:15 +0000 (14:43 +0200)]
typo --

6 years agoadd MSG_get_as_by_name call, answering [#17311]
Augustin Degomme [Mon, 5 May 2014 12:11:37 +0000 (14:11 +0200)]
add MSG_get_as_by_name call, answering [#17311]

6 years agoFix tesh-simdag-flatifier
Paul Bédaride [Tue, 6 May 2014 13:23:50 +0000 (15:23 +0200)]
Fix tesh-simdag-flatifier

6 years agoUpdate changelog
Paul Bédaride [Tue, 6 May 2014 12:13:26 +0000 (14:13 +0200)]
Update changelog

6 years agoReorganize AddTests.cmake
Paul Bédaride [Tue, 6 May 2014 09:07:44 +0000 (11:07 +0200)]
Reorganize AddTests.cmake

6 years agoCorrections to fat trees, set the size of connection vectors right
Stéphane Castelli [Tue, 6 May 2014 09:05:52 +0000 (11:05 +0200)]
Corrections to fat trees, set the size of connection vectors right

6 years agoMerge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Stéphane Castelli [Mon, 5 May 2014 15:10:55 +0000 (17:10 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid

6 years agoRewrite of the fat trees implementation
Stéphane Castelli [Mon, 5 May 2014 15:10:20 +0000 (17:10 +0200)]
Rewrite of the fat trees implementation

6 years agoThere was too much levels of backslash escapes. Remove one of them.
Arnaud Giersch [Mon, 5 May 2014 12:17:31 +0000 (14:17 +0200)]
There was too much levels of backslash escapes.  Remove one of them.

It seems that it changed with commit 782233, when the add_test statements
were changed from add_test(testname exe arg...) to
add_test(NAME testname COMMAND exe arg...).

This broke test memcheck-smpi-replay-9.

6 years agoDon't use xbt_assert for error checking, it may be disabled.
Arnaud Giersch [Mon, 5 May 2014 12:15:12 +0000 (14:15 +0200)]
Don't use xbt_assert for error checking, it may be disabled.

6 years agoSpecify dependencies between memcheck tests.
Arnaud Giersch [Mon, 5 May 2014 11:43:46 +0000 (13:43 +0200)]
Specify dependencies between memcheck tests.

Allows to run memcheck tests in parallel with e.g.
    ctest -D ExperimentalMemCheck -j 8

6 years agoRemove useless script (replaced by cmake macro ADD_TESH_FACTORIES).
Arnaud Giersch [Mon, 5 May 2014 09:40:20 +0000 (11:40 +0200)]
Remove useless script (replaced by cmake macro ADD_TESH_FACTORIES).

6 years agoMove tests so that the real binary is tested by valgrind, and not the tesh tool.
Arnaud Giersch [Mon, 5 May 2014 09:35:33 +0000 (11:35 +0200)]
Move tests so that the real binary is tested by valgrind, and not the tesh tool.

6 years agoRemove last references to the now obsolete testsuite/ directory.
Arnaud Giersch [Mon, 5 May 2014 09:09:38 +0000 (11:09 +0200)]
Remove last references to the now obsolete testsuite/ directory.

6 years agoRestructure teshsuite smpi
Paul Bédaride [Mon, 5 May 2014 11:38:04 +0000 (13:38 +0200)]
Restructure teshsuite smpi

6 years agoRestructure teshsuite simdag
Paul Bédaride [Wed, 30 Apr 2014 12:45:43 +0000 (14:45 +0200)]
Restructure teshsuite simdag

6 years agoRestructure teshsuite simix
Paul Bédaride [Wed, 30 Apr 2014 09:57:55 +0000 (11:57 +0200)]
Restructure teshsuite simix

6 years agoRestructure teshsuite xbt
Paul Bédaride [Wed, 30 Apr 2014 09:09:23 +0000 (11:09 +0200)]
Restructure teshsuite xbt

6 years agomodel-checker : restore previous version of smpi_finalize if MC is enabled
Marion Guthmuller [Tue, 29 Apr 2014 19:29:27 +0000 (21:29 +0200)]
model-checker : restore previous version of smpi_finalize if MC is enabled

6 years agomodel-checker : deal with the soundness of DPOR only if DPOR is enabled
Marion Guthmuller [Tue, 29 Apr 2014 15:21:59 +0000 (17:21 +0200)]
model-checker : deal with the soundness of DPOR only if DPOR is enabled

6 years agomodel-checker : fix the detection of determinism
Marion Guthmuller [Tue, 29 Apr 2014 15:03:09 +0000 (17:03 +0200)]
model-checker : fix the detection of determinism

6 years agoRemove absolute path from xbt-log tesh
Paul Bédaride [Tue, 29 Apr 2014 15:41:30 +0000 (17:41 +0200)]
Remove absolute path from xbt-log tesh

6 years agoAdd teshfiles to CMakelists.txt of teshsuite/mc
Paul Bédaride [Tue, 29 Apr 2014 15:29:58 +0000 (17:29 +0200)]
Add teshfiles to CMakelists.txt of teshsuite/mc

6 years agopermission change
Augustin Degomme [Tue, 29 Apr 2014 14:44:10 +0000 (16:44 +0200)]
permission change

6 years agoFix memcheck with new Add_test macros
Paul Bédaride [Tue, 29 Apr 2014 14:52:31 +0000 (16:52 +0200)]
Fix memcheck with new Add_test macros

6 years agoadd script for jenkins DynamicAnalysis build (memcheck+coverage)
Augustin Degomme [Tue, 29 Apr 2014 14:12:36 +0000 (16:12 +0200)]
add script for jenkins DynamicAnalysis build (memcheck+coverage)

6 years agoadd enable_memcheck_xml flag, to output valgrind in xml format (necessary for jenkins).
Augustin Degomme [Tue, 29 Apr 2014 13:39:43 +0000 (15:39 +0200)]
add enable_memcheck_xml flag, to output valgrind in xml format (necessary for jenkins).
This triggers enable_memcheck if not provided.

6 years agoFix DefinePackage mc CMakeLists.txt
Paul Bédaride [Tue, 29 Apr 2014 11:18:53 +0000 (13:18 +0200)]
Fix DefinePackage mc CMakeLists.txt

6 years agoFix teshsuite/surf/sur_usage/CMakeList.txt
Paul Bédaride [Tue, 29 Apr 2014 11:05:43 +0000 (13:05 +0200)]
Fix teshsuite/surf/sur_usage/CMakeList.txt

6 years agoMove mc testsuite to teshsuite
Paul Bédaride [Tue, 29 Apr 2014 09:04:25 +0000 (11:04 +0200)]
Move mc testsuite to teshsuite

6 years agoMove xbt testsuite to teshsuite
Paul Bédaride [Tue, 29 Apr 2014 08:15:47 +0000 (10:15 +0200)]
Move xbt testsuite to teshsuite

6 years agoRemove simeng_usage.tesh (not used)
Paul Bédaride [Mon, 28 Apr 2014 15:35:42 +0000 (17:35 +0200)]
Remove simeng_usage.tesh (not used)

6 years agoMove surf testsuite to teshsuite
Paul Bédaride [Mon, 28 Apr 2014 15:31:50 +0000 (17:31 +0200)]
Move surf testsuite to teshsuite

6 years agoRemove redundant test.
Arnaud Giersch [Mon, 28 Apr 2014 22:10:00 +0000 (00:10 +0200)]
Remove redundant test.

6 years agoSort file lists.
Arnaud Giersch [Fri, 25 Apr 2014 13:25:01 +0000 (15:25 +0200)]
Sort file lists.

6 years agoRemove stale temp file.
Arnaud Giersch [Fri, 25 Apr 2014 12:54:18 +0000 (14:54 +0200)]
Remove stale temp file.

6 years agoFix some javadoc warnings (remains 2).
Arnaud Giersch [Fri, 25 Apr 2014 12:47:45 +0000 (14:47 +0200)]
Fix some javadoc warnings (remains 2).

6 years agoNew ADD_TESH and ADD_TESH_FACTORIES cmake macros
Paul Bédaride [Mon, 28 Apr 2014 13:05:00 +0000 (15:05 +0200)]
New ADD_TESH and ADD_TESH_FACTORIES cmake macros

6 years agofix path to deployment fil in tesh example
Augustin Degomme [Mon, 28 Apr 2014 10:11:54 +0000 (12:11 +0200)]
fix path to deployment fil in tesh example

6 years agoleak --
Augustin Degomme [Mon, 28 Apr 2014 08:35:51 +0000 (10:35 +0200)]
leak --

6 years agoAdd missing CMakeLists.txt
Paul Bédaride [Mon, 28 Apr 2014 08:00:02 +0000 (10:00 +0200)]
Add missing CMakeLists.txt

6 years agoReorganize teshsuite/msg
Paul Bédaride [Fri, 25 Apr 2014 16:06:29 +0000 (18:06 +0200)]
Reorganize teshsuite/msg

6 years agoAdd an example of use of the previous interface.
Augustin Degomme [Fri, 25 Apr 2014 14:57:12 +0000 (16:57 +0200)]
Add an example of use of the previous interface.
This is based on the masterslave_mailbox example for MSG, but adds two instances of MPI application:
one doing a simple send/recv, the other performing an alltoall.

6 years agoForce valgrind to output in XML.
Augustin Degomme [Fri, 25 Apr 2014 14:04:55 +0000 (16:04 +0200)]
Force valgrind to output in XML.
This will break cdash use, but enable jenkins parsing and displaying of the reports. Hopefully.

6 years agoBegin to add an interface to instanciate SMPI programatically.
Augustin Degomme [Fri, 25 Apr 2014 14:33:35 +0000 (16:33 +0200)]
Begin to add an interface to instanciate SMPI programatically.

This adds SMPI_init, SMPI_finalize, and SMPI_app_instance_register

This allows user to deploy MSG+MPI applications in the same simulation,
as well as deploying several instances of SMPI with the same or different application,
without using smpirun. This means that we can now handle several MPI_COMM_WORLD.

This is only a very early step, as for now the interface still depends on MSG calls
 to deploy the application, and run the processes.

We should work on an interface to do this for SMPI, or add a generic interface
that would allow several SimGrid interfaces to run alltogether.

6 years agoexplode smpi_main into modules
Augustin Degomme [Fri, 25 Apr 2014 12:44:00 +0000 (14:44 +0200)]
explode smpi_main into modules

6 years agoforgotten in 0395dfc0
Augustin Degomme [Wed, 23 Apr 2014 16:53:36 +0000 (18:53 +0200)]
forgotten in 0395dfc0

6 years agoFix windows java build
Augustin Degomme [Wed, 23 Apr 2014 16:53:08 +0000 (18:53 +0200)]
Fix windows java build

6 years agoadd private (optional) data copy function callback to _send _isend _recv _irecv simcalls
Augustin Degomme [Wed, 23 Apr 2014 16:52:38 +0000 (18:52 +0200)]
add private (optional) data copy function callback to _send _isend _recv _irecv simcalls
It used to be globally set, but if we want to mix layers of SimGrid (SMPI + MSG calls)
we would want to have different copy functions used during the same simulation

6 years agonew flex run
Augustin Degomme [Wed, 23 Apr 2014 13:37:11 +0000 (15:37 +0200)]
new flex run

6 years agoMake MSG action module initialization/finalization an explicit call
Augustin Degomme [Thu, 17 Apr 2014 13:54:10 +0000 (15:54 +0200)]
Make MSG action module initialization/finalization an explicit call
This avoid setting up these modules every time, which could cause some issues when mixing

6 years agoMerge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Stéphane Castelli [Fri, 25 Apr 2014 12:50:42 +0000 (14:50 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid

6 years agoFix failed conflict (oups)
Paul Bédaride [Fri, 25 Apr 2014 12:56:57 +0000 (14:56 +0200)]
Fix failed conflict (oups)

6 years agoAdd MSG_process_join [#13601]
Paul Bédaride [Fri, 25 Apr 2014 12:52:03 +0000 (14:52 +0200)]
Add MSG_process_join [#13601]

6 years agoAdd in DefinePackage.cmake the man pages
Stéphane Castelli [Fri, 25 Apr 2014 12:50:23 +0000 (14:50 +0200)]
Add in DefinePackage.cmake the man pages

6 years agoMerge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Stéphane Castelli [Fri, 25 Apr 2014 12:15:22 +0000 (14:15 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid

6 years agoAdd a man page for smpicc and smpirun
Stéphane Castelli [Fri, 25 Apr 2014 11:26:32 +0000 (13:26 +0200)]
Add a man page for smpicc and smpirun

6 years agoFix warning about undefined identifier [-Wundef].
Arnaud Giersch [Thu, 24 Apr 2014 14:38:31 +0000 (16:38 +0200)]
Fix warning about undefined identifier [-Wundef].

Thanks to Samuel Thibault for the report.

6 years ago[mc] Enable Fortran 90 code with model cheking
Gabriel Corona [Thu, 24 Apr 2014 13:54:04 +0000 (15:54 +0200)]
[mc] Enable Fortran 90 code with model cheking

6 years ago[mmalloc] Extends the junk area and use finer grained allocations
Gabriel Corona [Thu, 24 Apr 2014 13:51:54 +0000 (15:51 +0200)]
[mmalloc] Extends the junk area and use finer grained allocations

With this we are able to run Fortran 90 code at the expense of 2 more
pages per state.

6 years agoNo need to copy these files.
Arnaud Giersch [Thu, 24 Apr 2014 13:19:28 +0000 (15:19 +0200)]
No need to copy these files.

Add --setenv srcdir=... to tests when needed.

6 years agoFix surf inclusion path for test smpi-tracing-ptp.
Arnaud Giersch [Thu, 24 Apr 2014 13:08:14 +0000 (15:08 +0200)]
Fix surf inclusion path for test smpi-tracing-ptp.