Augustin Degomme [Tue, 8 Jul 2014 16:09:28 +0000 (18:09 +0200)]
reduce size of recently added tests, to avoid them taking so long to complete
Augustin Degomme [Tue, 8 Jul 2014 15:26:15 +0000 (17:26 +0200)]
Previous commit exposes a bug in OpenMPI, switch algorithm for bcast to avoid it
The bug is that OpenMPI performs selection of algorithm based on the count on messages.
Or this count can be different from one process to another for bcast.
For example this test has a sender sending one element of 4096 bytes, and receivers receiving 1024 elements of 4 bytes
Yes, this is legal in MPI (but stupid ... but legal).
OpenMPI collective algorithm selector selects a different algorithm for these processes, thus deadlocking
Augustin Degomme [Tue, 8 Jul 2014 15:21:30 +0000 (17:21 +0200)]
activate two tests in collective communications
Augustin Degomme [Tue, 8 Jul 2014 13:32:42 +0000 (15:32 +0200)]
leak --
Augustin Degomme [Mon, 7 Jul 2014 15:22:04 +0000 (17:22 +0200)]
add minloc and maxloc for long_long
Augustin Degomme [Mon, 7 Jul 2014 15:14:50 +0000 (17:14 +0200)]
this should be err_rank, not err_comm
Augustin Degomme [Mon, 7 Jul 2014 15:09:11 +0000 (17:09 +0200)]
add a non-existing MPI_2LONG datatype to handle some corner cases in fortran
Paul Bédaride [Mon, 7 Jul 2014 14:51:33 +0000 (16:51 +0200)]
Xml platform cleaning teshsuite/msg
Augustin Degomme [Mon, 7 Jul 2014 14:07:14 +0000 (16:07 +0200)]
activate working tests
Lower sizes and number of messages for pingping to avoid taking too long
Augustin Degomme [Mon, 7 Jul 2014 13:33:32 +0000 (15:33 +0200)]
use the right types in fortran for some platforms (based on f2c matching, hopefully correct)
Augustin Degomme [Mon, 7 Jul 2014 13:51:34 +0000 (15:51 +0200)]
define types used by fortran even when only C code is used (needed sometimes)
Conflicts:
include/smpi/smpi.h
suter [Mon, 7 Jul 2014 13:45:17 +0000 (15:45 +0200)]
temporary hack to be able to replay traces of fortran codes through the
simulation of a C code
Marion Guthmuller [Thu, 3 Jul 2014 13:24:07 +0000 (15:24 +0200)]
don't destroy detached comm from the sender side during process cleanup
Marion Guthmuller [Thu, 3 Jul 2014 10:36:51 +0000 (12:36 +0200)]
fix debug message (wrong buffer was printed)
Marion Guthmuller [Thu, 3 Jul 2014 08:23:02 +0000 (10:23 +0200)]
model-checker : update tesh
Marion Guthmuller [Thu, 3 Jul 2014 08:22:37 +0000 (10:22 +0200)]
model-checker : cosmetic in log message
Marion Guthmuller [Thu, 3 Jul 2014 08:07:29 +0000 (10:07 +0200)]
model-checker : check dict content before removing value
Marion Guthmuller [Thu, 3 Jul 2014 08:06:01 +0000 (10:06 +0200)]
model-checker : remove useless condition
Gabriel Corona [Thu, 3 Jul 2014 10:12:34 +0000 (12:12 +0200)]
[mc] Fast path when comparing NULL against non-NULL pointers
Gabriel Corona [Thu, 3 Jul 2014 10:01:00 +0000 (12:01 +0200)]
[mc] Bug: MC was reading from the wrong region
Martin Quinson [Thu, 3 Jul 2014 08:12:33 +0000 (10:12 +0200)]
Little script to report on our MPICH3 coverage
CMakeLists.txt files are used as a source of information.
I had to slightly change one of them to make it easier to parse.
Augustin Degomme [Wed, 2 Jul 2014 08:04:07 +0000 (10:04 +0200)]
fix platform paths for windows packaging
Gabriel Corona [Tue, 1 Jul 2014 13:37:58 +0000 (15:37 +0200)]
[mc] Fix unit MC tests after platform file move/rename
Paul Bédaride [Tue, 1 Jul 2014 13:21:41 +0000 (15:21 +0200)]
Fix test-smpi-mpich3 tests
Paul Bédaride [Tue, 1 Jul 2014 12:52:39 +0000 (14:52 +0200)]
Fix CMakeLists.txt
Gabriel Corona [Tue, 1 Jul 2014 09:36:46 +0000 (11:36 +0200)]
[mc] Remove useless question in assert
Paul Bédaride [Tue, 1 Jul 2014 09:28:50 +0000 (11:28 +0200)]
Path relative to xml directory works
Paul Bédaride [Mon, 30 Jun 2014 14:14:18 +0000 (16:14 +0200)]
Xml platform cleaning examples/msg/ns3
Paul Bédaride [Mon, 30 Jun 2014 12:06:38 +0000 (14:06 +0200)]
Fix ns3
Paul Bédaride [Mon, 30 Jun 2014 11:48:39 +0000 (13:48 +0200)]
Xml platform cleaning example/scala
Paul Bédaride [Mon, 30 Jun 2014 08:33:22 +0000 (10:33 +0200)]
Revert commit
ed3e911d for better performances
Paul Bédaride [Wed, 25 Jun 2014 09:58:51 +0000 (11:58 +0200)]
Fix xmls for lua examples
Paul Bédaride [Tue, 24 Jun 2014 13:06:00 +0000 (15:06 +0200)]
Xml platform cleaning
Paul Bédaride [Tue, 17 Jun 2014 12:56:26 +0000 (14:56 +0200)]
Add difftool option to tesh to help updating tesh files
Augustin Degomme [Fri, 27 Jun 2014 15:42:34 +0000 (17:42 +0200)]
Wcloberred is unknown for clang...
Leave it, but avoid failing the build by setting -Wno-unknown-warning-option
Augustin Degomme [Fri, 27 Jun 2014 14:57:16 +0000 (16:57 +0200)]
Add other warning flags for cpp
Augustin Degomme [Fri, 27 Jun 2014 14:20:34 +0000 (16:20 +0200)]
as linking may be done through c++, add flags to remove some old/new warnings to cxx flags.
Augustin Degomme [Fri, 27 Jun 2014 12:50:14 +0000 (14:50 +0200)]
remove warnings in mpich c tests
Marion Guthmuller [Fri, 27 Jun 2014 12:09:21 +0000 (14:09 +0200)]
model-checker : fix visited states reduction with comm determinism verification
Marion Guthmuller [Fri, 27 Jun 2014 11:56:49 +0000 (13:56 +0200)]
model-checker : if no inttypes found in c++
Gabriel Corona [Fri, 27 Jun 2014 11:51:24 +0000 (13:51 +0200)]
Add missing dist files
Gabriel Corona [Fri, 27 Jun 2014 10:41:21 +0000 (12:41 +0200)]
Merge branch 'mc'
Conflicts:
src/mc/mc_global.c
Martin Quinson [Fri, 27 Jun 2014 09:30:41 +0000 (11:30 +0200)]
add some tests about the smpi robustness to PEBKAC
Martin Quinson [Fri, 27 Jun 2014 09:18:57 +0000 (11:18 +0200)]
make it easier to write tesh files by removing the full path to scripts from logs
Martin Quinson [Fri, 27 Jun 2014 08:03:23 +0000 (10:03 +0200)]
ignore some generated files
Martin Quinson [Fri, 27 Jun 2014 08:02:16 +0000 (10:02 +0200)]
Properly fix the case where the user provides an empty hostfile
Gabriel Corona [Fri, 27 Jun 2014 09:05:34 +0000 (11:05 +0200)]
[mc] Do not optimize src/xbt/dynar.c
For some reason, it breaks state comparison after the merge.
Stéphane Castelli [Fri, 27 Jun 2014 08:51:33 +0000 (10:51 +0200)]
Add an option to get the compiler version, change the manpages accordingly
Stéphane Castelli [Mon, 23 Jun 2014 07:58:22 +0000 (09:58 +0200)]
Some minor corrections on the documentation
Gabriel Corona [Fri, 27 Jun 2014 08:20:15 +0000 (10:20 +0200)]
Merge branch 'mc-fastsnapshot' into mc
Conflicts:
src/mc/mc_visited.c
Martin Quinson [Fri, 27 Jun 2014 06:53:09 +0000 (08:53 +0200)]
dividing by 0 is not always the best possible choice, even if the user shoulnt give empty hostfiles
Martin Quinson [Fri, 27 Jun 2014 06:48:26 +0000 (08:48 +0200)]
Merge branch 'master' of scm.gforge.inria.fr:/gitroot/simgrid/simgrid
Augustin Degomme [Thu, 26 Jun 2014 16:31:37 +0000 (18:31 +0200)]
Change the wonderful magic value of the max number of simcall parameters.
Fixes build on ci with better compilers than mine.
Marion Guthmuller [Thu, 26 Jun 2014 15:50:45 +0000 (17:50 +0200)]
model-checker : minor fix in comments
Marion Guthmuller [Thu, 26 Jun 2014 15:49:52 +0000 (17:49 +0200)]
model-checker : update tesh
Marion Guthmuller [Thu, 26 Jun 2014 15:49:25 +0000 (17:49 +0200)]
model-checker : visited states reduction available with comm determinism verification
Augustin Degomme [Thu, 26 Jun 2014 15:09:30 +0000 (17:09 +0200)]
JSURF_JAVA_GENERATED_SRC is already included in JMSG_JAVA_SRC
Some architectures complain and fail build if duplicated files are found
Augustin Degomme [Thu, 26 Jun 2014 13:48:26 +0000 (15:48 +0200)]
Start the remote send as soon as possible in MPI_Get
It used to be done only when synchronisation calls were issued, as it needed to be done sender's side
Augustin Degomme [Thu, 26 Jun 2014 13:47:32 +0000 (15:47 +0200)]
Use source of a MPI_Request as sender for SMPI messages.
make the SMPI messages use the sender process registered in the request perform the send, even if is not the same as self
Augustin Degomme [Thu, 26 Jun 2014 13:43:13 +0000 (15:43 +0200)]
Change send/isend simcalls to add a new parameter : src.
It is set everywhere it is used to SIMIX_process_self, so it should not change anything except for SMPI
This allows a process to ask for an immediate send from another process without any action from the sender's side
This is needed to simulate MPI_Get, where the receivers asynchronously triggers a send from another process (handled by the MPI library in another thread or directly through RMA)
Gabriel Corona [Thu, 26 Jun 2014 12:35:40 +0000 (14:35 +0200)]
[mc] Inline mc_snapshot_read_region()
Gabriel Corona [Thu, 26 Jun 2014 12:10:38 +0000 (14:10 +0200)]
[mc] Inline mc_translate_address() and mc_translate_address_region()
Gabriel Corona [Thu, 26 Jun 2014 10:12:16 +0000 (12:12 +0200)]
[mc] Forgot to decrease the page refcounts when freeing a snapshot
Augustin Degomme [Wed, 25 Jun 2014 07:56:49 +0000 (09:56 +0200)]
don't build mtest static lib if mpich3 tests are not enabled
Augustin Degomme [Tue, 24 Jun 2014 14:52:16 +0000 (16:52 +0200)]
change path where to look for an include file
Augustin Degomme [Tue, 24 Jun 2014 14:42:56 +0000 (16:42 +0200)]
activate tests for Accumulation with complex datatypes
Augustin Degomme [Tue, 24 Jun 2014 14:41:49 +0000 (16:41 +0200)]
Change the way MPI_Accumulate works with serialization, to handle complex datatypes.
Gabriel Corona [Tue, 24 Jun 2014 12:51:10 +0000 (14:51 +0200)]
[mc] Add support of privatized global variables in per-page snapshot
Not very useful because state comparison does not work with SMPI
privatisation yet.
Gabriel Corona [Tue, 24 Jun 2014 11:29:28 +0000 (13:29 +0200)]
[mc] Avoid loosing meaningful bits ot pair.p2 when hashing pointer pair
Gabriel Corona [Tue, 24 Jun 2014 08:34:51 +0000 (10:34 +0200)]
[mc] Use an unordered_set for compared_pointers instead of a dynarr
It has a better algorithmic complexity.
At the same time, avoid using TLS which is not implemented with
destructor-enabled data structures before C++11.
Gabriel Corona [Tue, 24 Jun 2014 07:23:20 +0000 (09:23 +0200)]
[mc] Remove declaration of mc_read_pagemap() in mc_page_store.cpp
It is now in mc_page_snapshot.cpp.
Gabriel Corona [Tue, 24 Jun 2014 07:13:41 +0000 (09:13 +0200)]
[mc] Optimise more files
degomme [Mon, 23 Jun 2014 17:19:24 +0000 (19:19 +0200)]
Switch to the right version of cmakefiles for mpich3 tests.
Augustin Degomme [Mon, 23 Jun 2014 16:11:21 +0000 (18:11 +0200)]
remove warning
Augustin Degomme [Mon, 23 Jun 2014 15:56:11 +0000 (17:56 +0200)]
remove now useless functions
Augustin Degomme [Mon, 23 Jun 2014 15:50:26 +0000 (17:50 +0200)]
remove now useless comments
Augustin Degomme [Mon, 23 Jun 2014 15:39:40 +0000 (17:39 +0200)]
add MPICH3 rma tests (15 out of 88 should be passing now)
Augustin Degomme [Mon, 23 Jun 2014 15:30:01 +0000 (17:30 +0200)]
First steps for implemtation of MPI Onesided Operations :
Support for MPI_Win structures, MPI_Put, MPI_Get, MPI_Accumulate, MPI_Win_Fence
Gabriel Corona [Mon, 23 Jun 2014 10:02:49 +0000 (12:02 +0200)]
[mc] Always inline mc_page_store_get_page (C version of page_store->get_page())
Marion Guthmuller [Wed, 4 Jun 2014 15:41:09 +0000 (17:41 +0200)]
model-checker : clarify comments
Gabriel Corona [Mon, 23 Jun 2014 08:30:57 +0000 (10:30 +0200)]
[mc] Move method definition out of class declaratio and force inline
degomme [Sat, 21 Jun 2014 11:41:51 +0000 (13:41 +0200)]
Fix java tests : this dynar may be empty, don't try to iterate on it
Pierre Veyre [Fri, 20 Jun 2014 15:01:27 +0000 (17:01 +0200)]
Remove call to concurrent-rw test
Gabriel Corona [Fri, 20 Jun 2014 14:09:12 +0000 (16:09 +0200)]
[mc] Reset the softdirty bits *before* restoring the ignored bytes (not after)
We want to know what has changed since the last per-page snapshot: we
have to reset the soft dirty bits just after the per-page restoration.
Pierre Veyre [Fri, 20 Jun 2014 13:37:29 +0000 (15:37 +0200)]
Remove storage.xml example file
Gabriel Corona [Fri, 20 Jun 2014 13:31:04 +0000 (15:31 +0200)]
[mc] Try to use optimisation with some part of the MC
Pierre Veyre [Fri, 20 Jun 2014 13:30:27 +0000 (15:30 +0200)]
Initiate java bindings for MSG storage API
Pierre Veyre [Mon, 2 Jun 2014 07:34:10 +0000 (09:34 +0200)]
Add new tesh suite for msg-storage
Gabriel Corona [Fri, 20 Jun 2014 12:02:05 +0000 (14:02 +0200)]
[mc] Always inline 'inline' functions
The MC is usually compiled without optimisation as they tend to
confuse the state comparison. But we *really* want to inline functions
declared 'inline'.
suter [Fri, 20 Jun 2014 09:05:07 +0000 (11:05 +0200)]
remove documentation of non-existing date parameter
suter [Fri, 20 Jun 2014 09:04:20 +0000 (11:04 +0200)]
prevent warining for undocumented (because internal) enum members when
making doc
suter [Fri, 20 Jun 2014 08:24:59 +0000 (10:24 +0200)]
doc is whining +indent
Augustin Degomme [Fri, 20 Jun 2014 09:04:42 +0000 (11:04 +0200)]
don't do the call if an error is caught before
Augustin Degomme [Fri, 20 Jun 2014 08:58:43 +0000 (10:58 +0200)]
Remove benchmarking calls from cart_create
Stéphane Castelli [Fri, 20 Jun 2014 08:32:44 +0000 (10:32 +0200)]
Small fixes to the documentation
Stéphane Castelli [Thu, 19 Jun 2014 07:49:14 +0000 (09:49 +0200)]
Remove some unbreakable spaces breaking the pdf generation
Gabriel Corona [Fri, 20 Jun 2014 08:33:27 +0000 (10:33 +0200)]
[mc] Fix compilation warnings/errors when optimizing
Gabriel Corona [Thu, 19 Jun 2014 14:07:55 +0000 (16:07 +0200)]
[mc] Documentation of mc_page_snapshot
Gabriel Corona [Thu, 19 Jun 2014 13:49:21 +0000 (15:49 +0200)]
[mc] Config to disable soft-dirty page tracking --cfg=model-check/soft-dirty:yes
It is here mostly to test the effect on soft-dirty page tracking on
the performance of the simulator. It might be removed in the future
once we will be certain that soft-dirty page tracking does not have a
significant negative impact on some workloads.
Gabriel Corona [Thu, 19 Jun 2014 13:33:12 +0000 (15:33 +0200)]
[mc] Avoid big alloca in mc_snapshot_region_memcp and mc_snapshot_memcp
Be more stack friendly.