Logo AND Algorithmique Numérique Distribuée

Public GIT Repository
simgrid.git
23 months agoRemove redundant access specifier.
Arnaud Giersch [Wed, 4 Jul 2018 14:01:17 +0000 (16:01 +0200)]
Remove redundant access specifier.

23 months agoUse an array.
Arnaud Giersch [Wed, 4 Jul 2018 13:59:57 +0000 (15:59 +0200)]
Use an array.

23 months agoKill unused field.
Arnaud Giersch [Tue, 3 Jul 2018 14:26:32 +0000 (16:26 +0200)]
Kill unused field.

23 months agoanother attempt at fixing weird clang compilers
Martin Quinson [Tue, 10 Jul 2018 07:43:09 +0000 (09:43 +0200)]
another attempt at fixing weird clang compilers

23 months agotry to fix build with old clangs
Martin Quinson [Tue, 10 Jul 2018 05:56:36 +0000 (07:56 +0200)]
try to fix build with old clangs

It does not break on my clang version even without this, so other breakage may remain

23 months agopartially stringify s4u::Actor
Martin Quinson [Mon, 9 Jul 2018 22:46:14 +0000 (00:46 +0200)]
partially stringify s4u::Actor

23 months agobetter (?) grammar
Martin Quinson [Mon, 9 Jul 2018 22:11:45 +0000 (00:11 +0200)]
better (?) grammar

23 months agosimgrid_convert_TI_traces: better error handling
Martin Quinson [Mon, 9 Jul 2018 21:47:54 +0000 (23:47 +0200)]
simgrid_convert_TI_traces: better error handling

Thanks for the idea, Tom.

23 months agothis_actor::exit() is a better name than this_actor::kill()
Martin Quinson [Mon, 9 Jul 2018 16:03:09 +0000 (18:03 +0200)]
this_actor::exit() is a better name than this_actor::kill()

23 months agoprefer automatic memory management
Martin Quinson [Mon, 9 Jul 2018 14:26:22 +0000 (16:26 +0200)]
prefer automatic memory management

23 months agoneed to change here as well
Augustin Degomme [Mon, 9 Jul 2018 19:37:27 +0000 (21:37 +0200)]
need to change here as well

23 months agoInitialize fortran data structures in all cases.
Augustin Degomme [Mon, 9 Jul 2018 16:26:50 +0000 (18:26 +0200)]
Initialize fortran data structures in all cases.
We have a weird proxy app with the main in C, and the rest in Fortran.
TODO :
 - check if overhead is fine (+leaks).

23 months agoprevent <0 errorcodes in MPI_Error_code
Augustin Degomme [Mon, 9 Jul 2018 16:03:06 +0000 (18:03 +0200)]
prevent <0 errorcodes in MPI_Error_code

23 months agoChange size of 2 fortran MPI datatypes.
Augustin Degomme [Mon, 9 Jul 2018 16:02:46 +0000 (18:02 +0200)]
Change size of 2 fortran MPI datatypes.
Not sure if helpful, but should be more accurate.

23 months agoFix issue with Fortran indexes not computed properly for waitany/testany.
Augustin Degomme [Mon, 9 Jul 2018 16:01:54 +0000 (18:01 +0200)]
Fix issue with Fortran indexes not computed properly for waitany/testany.

23 months agoattempt to fully trace ptasks
Frederic Suter [Mon, 9 Jul 2018 15:31:36 +0000 (17:31 +0200)]
attempt to fully trace ptasks

23 months agoMerge branch 'master' of framagit.org:simgrid/simgrid
Martin Quinson [Mon, 9 Jul 2018 04:50:08 +0000 (06:50 +0200)]
Merge branch 'master' of framagit.org:simgrid/simgrid

23 months agoplug a memleak
Martin Quinson [Mon, 9 Jul 2018 04:46:24 +0000 (06:46 +0200)]
plug a memleak

23 months agohave each model register into all_existing_models by itself
Martin Quinson [Sun, 8 Jul 2018 23:53:04 +0000 (01:53 +0200)]
have each model register into all_existing_models by itself

23 months agoprefer automatic mem handling (std::vector<>* becomes std::vector<>)
Martin Quinson [Sun, 8 Jul 2018 23:32:22 +0000 (01:32 +0200)]
prefer automatic mem handling (std::vector<>* becomes std::vector<>)

23 months agolet *Zones use their own network_model_ instead of the global surf_network_model
Martin Quinson [Sun, 8 Jul 2018 23:22:41 +0000 (01:22 +0200)]
let *Zones use their own network_model_ instead of the global surf_network_model

23 months agogive a network_model to each NetZone (unused for now)
Martin Quinson [Sun, 8 Jul 2018 23:17:58 +0000 (01:17 +0200)]
give a network_model to each NetZone (unused for now)

For now, it cannot be anything else than the network_model of the
englobing zone, or the global surf_network_model for the root zone.

For now, it's not used anywhere.

Of course, the goal is to finally implement this multi-model shit.

23 months agocosmetics
Martin Quinson [Sun, 8 Jul 2018 22:43:47 +0000 (00:43 +0200)]
cosmetics

23 months agomake the output readable for very large task amounts too
Martin Quinson [Sun, 8 Jul 2018 14:42:22 +0000 (16:42 +0200)]
make the output readable for very large task amounts too

23 months agotiny speed optim
Martin Quinson [Sun, 8 Jul 2018 14:41:57 +0000 (16:41 +0200)]
tiny speed optim

23 months agoignore unit tests
Martin Quinson [Sun, 8 Jul 2018 14:17:20 +0000 (16:17 +0200)]
ignore unit tests

23 months agoActually, SMPI is playing by the rules now. Remove friendship
Martin Quinson [Sun, 8 Jul 2018 10:12:32 +0000 (12:12 +0200)]
Actually, SMPI is playing by the rules now. Remove friendship

23 months agoanother last attempt
Augustin Degomme [Sun, 8 Jul 2018 19:04:20 +0000 (21:04 +0200)]
another last attempt

23 months agoanother last attempt
Augustin Degomme [Sun, 8 Jul 2018 18:28:56 +0000 (20:28 +0200)]
another last attempt

23 months agoone last attempt
Augustin Degomme [Sun, 8 Jul 2018 17:36:02 +0000 (19:36 +0200)]
one last attempt

23 months agooops
Augustin Degomme [Sun, 8 Jul 2018 15:21:25 +0000 (17:21 +0200)]
oops

23 months agofix coverage paths and add sloccount
Augustin Degomme [Sun, 8 Jul 2018 14:46:06 +0000 (16:46 +0200)]
fix coverage paths and add sloccount

23 months agomake the output readable for very large task amounts too
Martin Quinson [Sun, 8 Jul 2018 14:42:22 +0000 (16:42 +0200)]
make the output readable for very large task amounts too

23 months agotiny speed optim
Martin Quinson [Sun, 8 Jul 2018 14:41:57 +0000 (16:41 +0200)]
tiny speed optim

23 months agoignore unit tests
Martin Quinson [Sun, 8 Jul 2018 14:17:20 +0000 (16:17 +0200)]
ignore unit tests

23 months agoadd codacy upload to coverage
Augustin Degomme [Sun, 8 Jul 2018 13:49:44 +0000 (15:49 +0200)]
add codacy upload to coverage

23 months agoActually, SMPI is playing by the rules now. Remove friendship
Martin Quinson [Sun, 8 Jul 2018 10:12:32 +0000 (12:12 +0200)]
Actually, SMPI is playing by the rules now. Remove friendship

23 months agoMerge branch 'master' of scm.gforge.inria.fr:/gitroot/simgrid/simgrid
Martin Quinson [Sun, 8 Jul 2018 09:57:25 +0000 (11:57 +0200)]
Merge branch 'master' of scm.gforge.inria.fr:/gitroot/simgrid/simgrid

23 months agoadd include for papi
Augustin Degomme [Sun, 8 Jul 2018 09:49:31 +0000 (11:49 +0200)]
add include for papi

23 months agocodacy: remove trailing spaces
Martin Quinson [Sun, 8 Jul 2018 09:46:17 +0000 (11:46 +0200)]
codacy: remove trailing spaces

23 months agosurf/Cbindings: kill another unused function
Martin Quinson [Sun, 8 Jul 2018 09:39:28 +0000 (11:39 +0200)]
surf/Cbindings: kill another unused function

23 months agosurf/Cbinding: kill an unused and misnamed function
Martin Quinson [Sun, 8 Jul 2018 09:34:00 +0000 (11:34 +0200)]
surf/Cbinding: kill an unused and misnamed function

23 months agoI forgot once again to mark my constructors as explicit. thx sonar
Martin Quinson [Sun, 8 Jul 2018 09:27:18 +0000 (11:27 +0200)]
I forgot once again to mark my constructors as explicit. thx sonar

23 months agoC++ify some forgotten part of surf
Martin Quinson [Sun, 8 Jul 2018 09:24:24 +0000 (11:24 +0200)]
C++ify some forgotten part of surf

23 months agocontinue emptying surf.hpp from symbols that should not have been made public
Martin Quinson [Sun, 8 Jul 2018 08:37:22 +0000 (10:37 +0200)]
continue emptying surf.hpp from symbols that should not have been made public

23 months agotravis: allow_unauthenticated, as boost 1.60 are not authenticated anymore
Martin Quinson [Sun, 8 Jul 2018 08:21:01 +0000 (10:21 +0200)]
travis: allow_unauthenticated, as boost 1.60 are not authenticated anymore

23 months agoAllow to register plugins from their file, w/o touching the core implem
Martin Quinson [Sun, 8 Jul 2018 00:27:39 +0000 (02:27 +0200)]
Allow to register plugins from their file, w/o touching the core implem

23 months agothe model descriptions also don't need to be public
Martin Quinson [Sat, 7 Jul 2018 23:44:15 +0000 (01:44 +0200)]
the model descriptions also don't need to be public

23 months agocleanup __surf_is_absolute_file_path + empty a bit surf.hpp
Martin Quinson [Sat, 7 Jul 2018 23:29:44 +0000 (01:29 +0200)]
cleanup __surf_is_absolute_file_path + empty a bit surf.hpp

Sorry for smashing two commits together this way, I forgot to commit
between 2 cleanups (because of clang-format).

The symbols removed from the public surf.hpp should never have been
made public in the first place. I think (and hope) that nobody use them.

23 months agothe author of these lines loves complexity, but I don't
Martin Quinson [Sat, 7 Jul 2018 23:06:00 +0000 (01:06 +0200)]
the author of these lines loves complexity, but I don't

23 months agoCpuTiTmgr: further cleanups (make the enum class private)
Martin Quinson [Sat, 7 Jul 2018 22:58:53 +0000 (00:58 +0200)]
CpuTiTmgr: further cleanups (make the enum class private)

23 months agoCpuTiTmgr: change a C enum into an enum class
Martin Quinson [Sat, 7 Jul 2018 22:46:51 +0000 (00:46 +0200)]
CpuTiTmgr: change a C enum into an enum class

23 months agocosmetics around CpuTiModel
Martin Quinson [Sat, 7 Jul 2018 22:40:57 +0000 (00:40 +0200)]
cosmetics around CpuTiModel

23 months agoimprove a comment
Martin Quinson [Sat, 7 Jul 2018 22:37:23 +0000 (00:37 +0200)]
improve a comment

23 months agotiny cleanups in HostCLM03Model
Martin Quinson [Sat, 7 Jul 2018 22:23:59 +0000 (00:23 +0200)]
tiny cleanups in HostCLM03Model

23 months agoDijkstra is so ugly, I could spend weeks cleaning it up
Martin Quinson [Sat, 7 Jul 2018 22:13:01 +0000 (00:13 +0200)]
Dijkstra is so ugly, I could spend weeks cleaning it up

23 months agodijkstra: convert a public struct into a private class
Martin Quinson [Sat, 7 Jul 2018 21:14:10 +0000 (23:14 +0200)]
dijkstra: convert a public struct into a private class

23 months agosimplifications in DijkstraZone
Martin Quinson [Sat, 7 Jul 2018 20:03:49 +0000 (22:03 +0200)]
simplifications in DijkstraZone

23 months agosonar: allow # and ## in macros to please augustin
Martin Quinson [Sat, 7 Jul 2018 19:44:17 +0000 (21:44 +0200)]
sonar: allow # and ## in macros to please augustin

23 months agosnake_case kernel::resource::NetworkModel
Martin Quinson [Sat, 7 Jul 2018 19:32:50 +0000 (21:32 +0200)]
snake_case kernel::resource::NetworkModel

23 months agocodacy: reuse the ignore patterns from sonar
Martin Quinson [Sat, 7 Jul 2018 19:31:26 +0000 (21:31 +0200)]
codacy: reuse the ignore patterns from sonar

23 months agoMerge pull request #283 from mpoquet/smpi_convert_fix
Martin Quinson [Sat, 7 Jul 2018 19:01:00 +0000 (21:01 +0200)]
Merge pull request #283 from mpoquet/smpi_convert_fix

Smpi convert script fixes

23 months agoremove references to SimGrid-Multi jenkins job, now just SimGrid
Augustin Degomme [Sat, 7 Jul 2018 17:34:06 +0000 (19:34 +0200)]
remove references to SimGrid-Multi jenkins job, now just SimGrid

23 months agoadd codacy config file to ignore teshsuite folder for coverage computation
Augustin Degomme [Sat, 7 Jul 2018 17:20:47 +0000 (19:20 +0200)]
add codacy config file to ignore teshsuite folder for coverage computation

23 months agodeactivate bad call to MPI_Attr_put
Augustin Degomme [Fri, 6 Jul 2018 13:26:24 +0000 (15:26 +0200)]
deactivate bad call to MPI_Attr_put

23 months agoFix warning by putting the array only where it's needed
Augustin Degomme [Fri, 6 Jul 2018 13:25:38 +0000 (15:25 +0200)]
Fix warning by putting the array only where it's needed

23 months agocleanup
Augustin Degomme [Fri, 6 Jul 2018 13:06:10 +0000 (15:06 +0200)]
cleanup

23 months agoGenerate error strings at the same time enum is generated.
Augustin Degomme [Fri, 6 Jul 2018 12:56:06 +0000 (14:56 +0200)]
Generate error strings at the same time enum is generated.
- TODO : put back MPI_Errhandler_f2c and MPI_File_f2c unimplemented signatures...
Somehow converting from "#define MPI_SUCCESS 0" to their return type was OK but from enum is not for the compiler.

23 months agoHave MPI_* send a debug messages whenever output is not MPI_SUCCESS
Augustin Degomme [Fri, 6 Jul 2018 12:19:09 +0000 (14:19 +0200)]
Have MPI_* send a debug messages whenever output is not MPI_SUCCESS
Should help debugging quite a lot
TODO
- pass the message in WARNING
- requalify teshes for this (some intentionally give bad params to test)
- fix bugs in SMPI (already found one useless MPI_Attr_put call for example).

23 months ago- Support MPI_Error_string
Augustin Degomme [Fri, 6 Jul 2018 12:16:49 +0000 (14:16 +0200)]
- Support MPI_Error_string
TODO :
- better and extensible way to store errors/errorstrings, now it's stupid

23 months agoRevert "address FIXME and kill useless code"
Frederic Suter [Fri, 6 Jul 2018 10:14:18 +0000 (12:14 +0200)]
Revert "address FIXME and kill useless code"

This reverts commit 4bfa8a32e0083bcd28a78689625c4806c2fbbf46.

23 months agoaddress FIXME and kill useless code
Frederic Suter [Fri, 6 Jul 2018 09:55:07 +0000 (11:55 +0200)]
address FIXME and kill useless code

23 months ago[tools] replace asserts by error and exit in TI convert script
MERCIER Michael [Fri, 6 Jul 2018 09:20:18 +0000 (11:20 +0200)]
[tools] replace asserts by error and exit in TI convert script

23 months ago[SMPI] Remove specific encoding of MPI_DATATYPE_NULL
Christian Heinrich [Fri, 6 Jul 2018 06:38:56 +0000 (08:38 +0200)]
[SMPI] Remove specific encoding of MPI_DATATYPE_NULL

This code is now being treated by MPI_DATATYPE_NULL
being an object.

23 months ago[SMPI] Make MPI_DATATYPE_NULL a non-null object
Christian Heinrich [Fri, 6 Jul 2018 06:37:37 +0000 (08:37 +0200)]
[SMPI] Make MPI_DATATYPE_NULL a non-null object

We're calling methods on the Datatype objects, such as
datatype->size(), especially in the replay component.
With MPI_DATATYPE_NULL being the nullptr, we'd have to
check every time whether or not a given object is
actually MPI_DATATYPE_NULL before using it.

Making MPI_DATATYPE_NULL a real object allows us to call methods
on this object that will be meaningful (for instance, size() returns 0)
and to avoid this constant checking.

Thanks to Michael Mercier for reporting this bug.

23 months ago[tools] Better corner case handeling + lower everything in TI trace convert
MERCIER Michael [Thu, 5 Jul 2018 16:40:47 +0000 (18:40 +0200)]
[tools] Better corner case handeling + lower everything in TI trace convert

23 months agoMerge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Frederic Suter [Thu, 5 Jul 2018 13:24:39 +0000 (15:24 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid

23 months ago[SMPI] Assert: host_speed must be > 0, not >= 0.
Christian Heinrich [Thu, 5 Jul 2018 12:59:50 +0000 (14:59 +0200)]
[SMPI] Assert: host_speed must be > 0, not >= 0.

23 months ago[CONFIG] Add smpi/plugin/lb/migration-frequency option
Christian Heinrich [Thu, 5 Jul 2018 12:42:27 +0000 (14:42 +0200)]
[CONFIG] Add smpi/plugin/lb/migration-frequency option

This needs to be moved to the plugin file, but unfortunately, when the
smpirun command is parsed, the plugin is not yet initialized and hence
an error is thrown (since this option is not yet known).

23 months ago[SMPI] Replace surf_get_clock with SIMIX_get_clock
Christian Heinrich [Thu, 5 Jul 2018 12:38:25 +0000 (14:38 +0200)]
[SMPI] Replace surf_get_clock with SIMIX_get_clock

23 months ago[S4U] Initialize s4u::Barrier in initializer list
Christian Heinrich [Thu, 5 Jul 2018 12:19:44 +0000 (14:19 +0200)]
[S4U] Initialize s4u::Barrier in initializer list

23 months agoMerge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Frederic Suter [Thu, 5 Jul 2018 13:08:51 +0000 (15:08 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid

23 months agoadd C interface to s4u::Barrier
Frederic Suter [Thu, 5 Jul 2018 13:08:21 +0000 (15:08 +0200)]
add C interface to s4u::Barrier
move MSG version to legacy

23 months agoremove a dependency on MSG from SMPI
Frederic Suter [Thu, 5 Jul 2018 11:35:50 +0000 (13:35 +0200)]
remove a dependency on MSG from SMPI

23 months ago[CI] Activate PAPI in Coverage checks.
Christian Heinrich [Thu, 5 Jul 2018 09:53:28 +0000 (11:53 +0200)]
[CI] Activate PAPI in Coverage checks.

There are no real tests (yet), but at least that
code compiles again and will be tested for compilation
errors.

23 months ago[DVFS] Make the Dvfs plugin available
Christian Heinrich [Thu, 5 Jul 2018 06:55:53 +0000 (08:55 +0200)]
[DVFS] Make the Dvfs plugin available

This commit allows users to actually enable the plugin
via smpirun.

23 months ago[SMPI/PAPI] Make PAPI counters compile again ;)
Christian Heinrich [Wed, 4 Jul 2018 14:07:22 +0000 (16:07 +0200)]
[SMPI/PAPI] Make PAPI counters compile again ;)

This commit contains all the changes that were necessary
to reflect the changes in SimGrid/S4U since 2016.

23 months ago[CMAKE] Fix grammar for SMPI built option
Christian Heinrich [Tue, 3 Jul 2018 17:35:40 +0000 (19:35 +0200)]
[CMAKE] Fix grammar for SMPI built option

23 months ago[ENERGY] Move comment to the correct location
Christian Heinrich [Mon, 2 Jul 2018 13:31:15 +0000 (15:31 +0200)]
[ENERGY] Move comment to the correct location

23 months agodisable the autorestart test for now, sorry
Martin Quinson [Wed, 4 Jul 2018 23:55:20 +0000 (01:55 +0200)]
disable the autorestart test for now, sorry

2 years agotry to fix the build on centos
Martin Quinson [Wed, 4 Jul 2018 20:36:56 +0000 (22:36 +0200)]
try to fix the build on centos

2 years agostupid typo
Frederic Suter [Wed, 4 Jul 2018 14:25:41 +0000 (16:25 +0200)]
stupid typo

2 years agoAttempt to move msg_bar_t to S4U
Frederic Suter [Wed, 4 Jul 2018 14:01:51 +0000 (16:01 +0200)]
Attempt to move msg_bar_t to S4U

2 years agoBoost.test must be >= v1.59 to be useful to us
Martin Quinson [Wed, 4 Jul 2018 13:14:51 +0000 (15:14 +0200)]
Boost.test must be >= v1.59 to be useful to us

2 years agoRevert "cmake: cleanups around Boost detection"
Martin Quinson [Wed, 4 Jul 2018 13:12:56 +0000 (15:12 +0200)]
Revert "cmake: cleanups around Boost detection"

It requires cmake v3.10.2, released in january 2018...

This reverts commit 853baf79434abbab6ad23c839eb86f68ea546896.

2 years agocmake: cleanups around Boost detection
Martin Quinson [Wed, 4 Jul 2018 12:56:01 +0000 (14:56 +0200)]
cmake: cleanups around Boost detection

2 years agobetter integration of the new MC unit tests
Martin Quinson [Wed, 4 Jul 2018 08:07:30 +0000 (10:07 +0200)]
better integration of the new MC unit tests

Same functionalities than before, but leveraging boost_unit_test
instead of our own unit testing.

2 years agofix make distcheck
Martin Quinson [Wed, 4 Jul 2018 08:00:34 +0000 (10:00 +0200)]
fix make distcheck

2 years agoMerge branch 'master' of https://github.com/Onesphore/simgrid into Onesphore-master
Martin Quinson [Wed, 4 Jul 2018 07:55:33 +0000 (09:55 +0200)]
Merge branch 'master' of https://github.com/Onesphore/simgrid into Onesphore-master

2 years agoremove the old-style MC unit tests
Martin Quinson [Wed, 4 Jul 2018 07:46:17 +0000 (09:46 +0200)]
remove the old-style MC unit tests

They will come back renewed when I merge PR #273.