Martin Quinson [Fri, 17 Jun 2016 07:52:10 +0000 (09:52 +0200)]
update doc to nowadays lexique
Martin Quinson [Fri, 17 Jun 2016 07:47:49 +0000 (09:47 +0200)]
small cleanups in s4u::hosts: improve doc, kill useless function
Martin Quinson [Thu, 16 Jun 2016 20:29:31 +0000 (22:29 +0200)]
small improvments to the s4u_actor doc
adfaure [Thu, 16 Jun 2016 15:09:36 +0000 (17:09 +0200)]
[simix] Refcounting with SIMIX_cond_{ref,unref}()
Gabriel Corona [Thu, 16 Jun 2016 13:12:27 +0000 (15:12 +0200)]
[simix] Add warning about atomic operations
Frederic Suter [Thu, 16 Jun 2016 10:21:09 +0000 (12:21 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Frederic Suter [Thu, 16 Jun 2016 09:00:52 +0000 (11:00 +0200)]
remove some smells in Java
Gabriel Corona [Thu, 16 Jun 2016 08:57:39 +0000 (10:57 +0200)]
[simix] Group simcall_process_create()
Frederic Suter [Thu, 16 Jun 2016 08:11:33 +0000 (10:11 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Frederic Suter [Thu, 16 Jun 2016 08:08:14 +0000 (10:08 +0200)]
make pastry example functional
no idea what it does though
Gabriel Corona [Thu, 16 Jun 2016 07:37:17 +0000 (09:37 +0200)]
[simix] Create actors by value (instead of new()-ing and leaking them)
Frederic Suter [Thu, 16 Jun 2016 07:36:55 +0000 (09:36 +0200)]
clean chord a bit
Gabriel Corona [Thu, 16 Jun 2016 07:25:03 +0000 (09:25 +0200)]
[simix] Refcounting with SIMIX_{process,mutex}_{ref,unref}()
Martin Quinson [Wed, 15 Jun 2016 23:51:00 +0000 (01:51 +0200)]
Improve the doc of the S4U actors
Martin Quinson [Wed, 15 Jun 2016 20:49:19 +0000 (22:49 +0200)]
start including S4U in the doc
Martin Quinson [Wed, 15 Jun 2016 19:48:33 +0000 (21:48 +0200)]
Merge branch 'master' of scm.gforge.inria.fr:/gitroot/simgrid/simgrid
Gabriel Corona [Wed, 15 Jun 2016 14:48:07 +0000 (16:48 +0200)]
[s4u] Cleanup s4u mutex example
Gabriel Corona [Wed, 15 Jun 2016 14:44:44 +0000 (16:44 +0200)]
[s4u] Use simix simix::Mutex refcount support for s4u::Mutex
Gabriel Corona [Tue, 14 Jun 2016 11:26:58 +0000 (13:26 +0200)]
[simix] Add refcount to Process
adfaure [Wed, 15 Jun 2016 13:45:56 +0000 (15:45 +0200)]
[s4u] add mutex
Gabriel Corona [Wed, 15 Jun 2016 13:00:04 +0000 (15:00 +0200)]
[simix] Fix refcount in Mutex
Christian Heinrich [Wed, 15 Jun 2016 08:26:44 +0000 (10:26 +0200)]
Revert "[SMPI] Fix wrong return value in MPI_Testany."
This reverts commit
efcf0fb465cf0465c462fc28a830ccd4d6b6b13c.
The reason I reverted again is that the standard does actually state
"The array may contain null or inactive handles. If the array contains
no active handles then the call returns immediately with flag = true,
index = MPI_UNDEFINED , and an empty status."
This is a corner case and was correctly handled here.
degomme [Tue, 14 Jun 2016 23:32:33 +0000 (01:32 +0200)]
Add a test for jenkins to try a few configuration options
for the nightly build. This should trigger 16 builds for now, which may be a lot.
degomme [Tue, 14 Jun 2016 22:58:13 +0000 (00:58 +0200)]
fix build with MC but without SMPI
Martin Quinson [Tue, 14 Jun 2016 19:40:57 +0000 (21:40 +0200)]
tiny doc formating improvement
Martin Quinson [Tue, 14 Jun 2016 19:33:40 +0000 (21:33 +0200)]
Merge branch 'master' of scm.gforge.inria.fr:/gitroot/simgrid/simgrid
Christian Heinrich [Tue, 14 Jun 2016 16:43:28 +0000 (18:43 +0200)]
[SMPI] Update notice in smpirun.in for cluster tags.
Christian Heinrich [Tue, 14 Jun 2016 14:06:37 +0000 (16:06 +0200)]
[SMPI] Fix wrong return value in MPI_Testany.
See the MPI Standard on this function: flag must be true
if and only if there was at least one operation completed.
Christian Heinrich [Tue, 14 Jun 2016 13:51:43 +0000 (15:51 +0200)]
[SMPI] Remove outdated comment.
Christian Heinrich [Tue, 14 Jun 2016 13:46:21 +0000 (15:46 +0200)]
[SMPI] Cosmetics in smpi_base.cpp
Christian Heinrich [Tue, 14 Jun 2016 13:34:32 +0000 (15:34 +0200)]
[SIMIX] Add a comment
Christian Heinrich [Tue, 14 Jun 2016 13:34:17 +0000 (15:34 +0200)]
[SMPI] C++ify smpi_mpi_testany(). There's more to do though.
Christian Heinrich [Tue, 14 Jun 2016 11:35:43 +0000 (13:35 +0200)]
[SMPI] Alignment cosmetics.
Christian Heinrich [Tue, 14 Jun 2016 10:17:22 +0000 (12:17 +0200)]
[SMX] Alignment cosmetics.
Gabriel Corona [Tue, 14 Jun 2016 11:01:30 +0000 (13:01 +0200)]
[simix] Add refcount to Mutex
Gabriel Corona [Tue, 14 Jun 2016 08:47:38 +0000 (10:47 +0200)]
[simix] Make Mutex a C++ class (kind-of)
Frederic Suter [Tue, 14 Jun 2016 08:11:58 +0000 (10:11 +0200)]
bugs removal of the day
degomme [Tue, 14 Jun 2016 00:23:24 +0000 (02:23 +0200)]
[SMPI] Die more peacefully when a process calls exit or stop, either before or after MPI_Finalize()
degomme [Mon, 13 Jun 2016 22:18:59 +0000 (00:18 +0200)]
warnings --
Christian Heinrich [Mon, 13 Jun 2016 18:49:30 +0000 (20:49 +0200)]
[SMPI] Rewrote parsing of smpi-factors in C++
We do want to move the parsing to some other location, as this
might be required for other parameters as well
Christian Heinrich [Mon, 13 Jun 2016 13:19:36 +0000 (15:19 +0200)]
[SMPI] Alignments in smpi_base.cpp
Christian Heinrich [Mon, 13 Jun 2016 12:40:09 +0000 (14:40 +0200)]
[CHANGELOG] Added recent SMPI fixes
Martin Quinson [Mon, 13 Jun 2016 16:06:47 +0000 (18:06 +0200)]
small pass over the doc
Martin Quinson [Mon, 13 Jun 2016 14:56:55 +0000 (16:56 +0200)]
Merge branch 'master' of scm.gforge.inria.fr:/gitroot/simgrid/simgrid
Gabriel Corona [Fri, 10 Jun 2016 11:54:52 +0000 (13:54 +0200)]
[xbt] Move backtrace support for exceptions in a mixin class
* mixin class to add backtraces to any exception;
* code to display exceptions with name, what(), backtraces and nested
exceptions;
* demangle C++ names (class type names, function names).
Martin Quinson [Mon, 13 Jun 2016 10:01:12 +0000 (12:01 +0200)]
ignorable cosmetics
Martin Quinson [Mon, 13 Jun 2016 09:54:31 +0000 (11:54 +0200)]
dePERLify a bit the routing
- Use the object, not its name (src & dst in s_sg_platf_route_cbarg_t)
- this allows to move the test of their existance earlier in the process
Martin Quinson [Sun, 12 Jun 2016 20:08:19 +0000 (22:08 +0200)]
remove redundent forward declarations
Martin Quinson [Sun, 12 Jun 2016 19:54:09 +0000 (21:54 +0200)]
spellcheck mc. Don't ask why
degomme [Sat, 11 Jun 2016 10:20:46 +0000 (12:20 +0200)]
add missing include
Martin Quinson [Fri, 10 Jun 2016 23:06:41 +0000 (01:06 +0200)]
No need for the OSSP copyright header on the backtrace thing, only for the exceptions
Martin Quinson [Fri, 10 Jun 2016 22:04:05 +0000 (00:04 +0200)]
rewrite the doc of option exceptions/cutpath
Gabriel Corona [Fri, 10 Jun 2016 12:21:17 +0000 (14:21 +0200)]
Fix backtrace_dummy
Gabriel Corona [Fri, 10 Jun 2016 11:07:15 +0000 (13:07 +0200)]
Factor get_binary_path() out of resolveBacktrace()
Gabriel Corona [Fri, 10 Jun 2016 10:14:46 +0000 (12:14 +0200)]
Move the backtrace code out of the exception code
Christian Heinrich [Fri, 10 Jun 2016 08:51:21 +0000 (10:51 +0200)]
[DOC] Update exception/cutpath option
Christian Heinrich [Fri, 10 Jun 2016 06:43:08 +0000 (08:43 +0200)]
[DOC] Update documentation on smpi/privatize-global-variables
Frederic Suter [Thu, 9 Jun 2016 14:51:26 +0000 (16:51 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Frederic Suter [Thu, 9 Jun 2016 14:49:38 +0000 (16:49 +0200)]
reduce smell of python scripts too
Frederic Suter [Thu, 9 Jun 2016 14:49:09 +0000 (16:49 +0200)]
cosmetics in java
Frederic Suter [Thu, 9 Jun 2016 14:47:34 +0000 (16:47 +0200)]
C++ify this example + fix bug in tesh
Gabriel Corona [Mon, 6 Jun 2016 14:22:59 +0000 (16:22 +0200)]
Getting rid of C exceptions
The goal if to used native C++ exception instead our own SLJL-based C
exceptions:
* this removes the problem of exception conversion between C and C++
context and the lack of interoperability of SimGrid THROW/TRY/CATCH
with its C++ counterpart;
* we can still throw exceptions from C;
* but we can't catch exception from C anymore, you should use C++
instead if order to do that;
Things to do afterwards:
* remove exceptions from the C public APIs;
* restore the multiple usage debugging;
* fix the backtrace code and separate from the exception code in order
to be able to add backtraces to any exception and get backtraces
without generating exceptions;
* add support for attaching backtraces to any exception;
* attach context (baktraces and cause) everywhere;
Gabriel Corona [Thu, 9 Jun 2016 12:08:29 +0000 (14:08 +0200)]
maxmin_bench, use stderr only (not stdout)
Gabriel Corona [Thu, 9 Jun 2016 09:57:03 +0000 (11:57 +0200)]
[jenkins] Add lib/ in the PATH for Windows
Christian Heinrich [Thu, 9 Jun 2016 13:06:58 +0000 (15:06 +0200)]
[SMPI] Fix local path that slipped in :)
Frederic Suter [Thu, 9 Jun 2016 09:14:51 +0000 (11:14 +0200)]
remove a few sonar stuff
Christian Heinrich [Wed, 8 Jun 2016 21:26:36 +0000 (23:26 +0200)]
[INCLUDES] NULL -> nullptr substitution.
Christian Heinrich [Wed, 8 Jun 2016 21:22:15 +0000 (23:22 +0200)]
[TESHSUITE] NULL -> nullptr substitution
I used the following command: (the '**' means recursion in ZSH)
sed -i -e 's/\([^_]\s*\)NULL/\1nullptr/g' src/**/*.cpp
We check for the underscore to avoid replacing MPI_*_NULL
Christian Heinrich [Wed, 8 Jun 2016 21:20:59 +0000 (23:20 +0200)]
[XBT] NULL -> nullptr substitution
I used the following command: (the '**' means recursion in ZSH)
sed -i -e 's/\([^_]\s*\)NULL/\1nullptr/g' src/**/*.cpp
We check for the underscore to avoid replacing MPI_*_NULL
Christian Heinrich [Wed, 8 Jun 2016 21:20:59 +0000 (23:20 +0200)]
[SURF] NULL -> nullptr substitution
I used the following command: (the '**' means recursion in ZSH)
sed -i -e 's/\([^_]\s*\)NULL/\1nullptr/g' src/**/*.cpp
We check for the underscore to avoid replacing MPI_*_NULL
Christian Heinrich [Wed, 8 Jun 2016 21:20:59 +0000 (23:20 +0200)]
[SMPI] NULL -> nullptr substitution
I used the following command: (the '**' means recursion in ZSH)
sed -i -e 's/\([^_]\s*\)NULL/\1nullptr/g' src/**/*.cpp
We check for the underscore to avoid replacing MPI_*_NULL
Christian Heinrich [Wed, 8 Jun 2016 21:20:59 +0000 (23:20 +0200)]
[SIMIX] NULL -> nullptr substitution
I used the following command: (the '**' means recursion in ZSH)
sed -i -e 's/\([^_]\s*\)NULL/\1nullptr/g' src/**/*.cpp
We check for the underscore to avoid replacing MPI_*_NULL
Christian Heinrich [Wed, 8 Jun 2016 21:20:59 +0000 (23:20 +0200)]
[SIMGRID] NULL -> nullptr substitution
I used the following command: (the '**' means recursion in ZSH)
sed -i -e 's/\([^_]\s*\)NULL/\1nullptr/g' src/**/*.cpp
We check for the underscore to avoid replacing MPI_*_NULL
Christian Heinrich [Wed, 8 Jun 2016 21:20:59 +0000 (23:20 +0200)]
[SIMDAG] NULL -> nullptr substitution
I used the following command: (the '**' means recursion in ZSH)
sed -i -e 's/\([^_]\s*\)NULL/\1nullptr/g' src/**/*.cpp
We check for the underscore to avoid replacing MPI_*_NULL
Christian Heinrich [Wed, 8 Jun 2016 21:20:59 +0000 (23:20 +0200)]
[S4U] NULL -> nullptr substitution
I used the following command: (the '**' means recursion in ZSH)
sed -i -e 's/\([^_]\s*\)NULL/\1nullptr/g' src/**/*.cpp
We check for the underscore to avoid replacing MPI_*_NULL
Christian Heinrich [Wed, 8 Jun 2016 21:20:59 +0000 (23:20 +0200)]
[MSG] NULL -> nullptr substitution
I used the following command: (the '**' means recursion in ZSH)
sed -i -e 's/\([^_]\s*\)NULL/\1nullptr/g' src/**/*.cpp
We check for the underscore to avoid replacing MPI_*_NULL
Christian Heinrich [Wed, 8 Jun 2016 21:20:59 +0000 (23:20 +0200)]
[INSTR] NULL -> nullptr substitution
I used the following command: (the '**' means recursion in ZSH)
sed -i -e 's/\([^_]\s*\)NULL/\1nullptr/g' src/**/*.cpp
We check for the underscore to avoid replacing MPI_*_NULL
Christian Heinrich [Wed, 8 Jun 2016 21:20:59 +0000 (23:20 +0200)]
[BINDINGS] NULL -> nullptr substitution
I used the following command: (the '**' means recursion in ZSH)
sed -i -e 's/\([^_]\s*\)NULL/\1nullptr/g' src/**/*.cpp
We check for the underscore to avoid replacing MPI_*_NULL
degomme [Wed, 8 Jun 2016 15:30:11 +0000 (17:30 +0200)]
needed for win
degomme [Wed, 8 Jun 2016 15:01:02 +0000 (17:01 +0200)]
Finally test for timing injection in MPI_Send, MPI_Isend, MPI_Recv
MPI_Irecv timings are injected in the consecutive MPI_Wait, actually
degomme [Wed, 8 Jun 2016 14:34:49 +0000 (16:34 +0200)]
Enhance pt2pt test for dsend to test MPI_Isend/MPI_Irecv as well.
degomme [Wed, 8 Jun 2016 14:28:34 +0000 (16:28 +0200)]
Fix bug, MPI_Recv timings for small messages were actually ignored
Christian Heinrich [Wed, 8 Jun 2016 13:47:02 +0000 (15:47 +0200)]
[SMPI] Add a simple comment for the else clause
Christian Heinrich [Wed, 8 Jun 2016 13:46:44 +0000 (15:46 +0200)]
[SMPI] Fix missing timings for MPI_Send.
Timings were never injected for MPI_Send since request->detached
was only set to something != 0 after the timings were already
injected. I therefore moved the block that computes this value
up. See also
825ad35c811f
Christian Heinrich [Wed, 8 Jun 2016 13:43:06 +0000 (15:43 +0200)]
[SMPI] Cosmetics & replace call to smpi_os with sleeptime
Christian Heinrich [Tue, 7 Jun 2016 09:32:42 +0000 (11:32 +0200)]
[SMPI] Make an if-condition more readable
Christian Heinrich [Tue, 7 Jun 2016 08:44:27 +0000 (10:44 +0200)]
[DOC] Corrected SMPI option name
Christian Heinrich [Tue, 7 Jun 2016 08:31:28 +0000 (10:31 +0200)]
[SMPI] Even more cosmetics...
Christian Heinrich [Tue, 7 Jun 2016 08:15:56 +0000 (10:15 +0200)]
[SMPI] More alignment cosmetics
Christian Heinrich [Tue, 7 Jun 2016 08:08:24 +0000 (10:08 +0200)]
[SMPI] Indentation in smpi_base.cpp
Christian Heinrich [Tue, 7 Jun 2016 08:08:10 +0000 (10:08 +0200)]
[SMPI] Cosmetics / NULL <-> nullptr in smpi_base.cpp
Christian Heinrich [Tue, 7 Jun 2016 05:27:05 +0000 (07:27 +0200)]
[SMPI] Grammar in smpi_base.cpp
Christian Heinrich [Tue, 7 Jun 2016 05:24:14 +0000 (07:24 +0200)]
[SMPI] Change some NULL to nullptr
Christian Heinrich [Tue, 7 Jun 2016 05:22:03 +0000 (07:22 +0200)]
[SMPI] Alignment cosmetics in smpi_base.cpp
Christian Heinrich [Sat, 4 Jun 2016 14:42:25 +0000 (16:42 +0200)]
[SMPI] Convert smpi_c99.cpp to actual C++ code
Christian Heinrich [Fri, 3 Jun 2016 10:32:44 +0000 (12:32 +0200)]
[SMPI] PMPI_Ssend: Cosmetics
Christian Heinrich [Fri, 3 Jun 2016 10:28:26 +0000 (12:28 +0200)]
[SMPI] PMPI_Send: Cosmetics
Christian Heinrich [Fri, 3 Jun 2016 10:28:00 +0000 (12:28 +0200)]
[SMPI] NULL -> nullptr in smpi_base.cpp
Christian Heinrich [Thu, 2 Jun 2016 14:49:06 +0000 (16:49 +0200)]
[DOC] Document the 'smpi/init' config option
Martin Quinson [Tue, 7 Jun 2016 22:23:44 +0000 (00:23 +0200)]
woops, I broke everything