1 /*! \page faq Frequently Asked Questions
3 \htmlinclude .FAQ.doc.toc
5 \section faq_installation Installing the SimGrid library
7 Many people have been asking me questions on how to use SimGrid. Quite
8 often, the questions were not really about SimGrid but on the
9 installation process. This section is intended to help people that are
10 not familiar with compiling C files under UNIX. If you follow these
11 instructions and still have some troubles, drop an e-mail to
12 <simgrid-user@lists.gforge.inria.fr>.
14 \subsection faq_compiling Compiling SimGrid from an archive
16 First of all, you need to download the latest version of SimGrid from
17 <a href="http://gforge.inria.fr/frs/?group_id=12">here</a>.
18 Suppose you have uncompressed SimGrid in some temporary location of
19 your home directory (say <tt>/home/joe/tmp/simgrid-3.0.1 </tt>). The
20 simplest way to use SimGrid is to install it in your home
21 directory. Change your directory to
22 <tt>/home/joe/tmp/simgrid-3.0.1</tt> and type
24 \verbatim./configure --prefix=$HOME
29 If at some point, something fails, check the section "\ref
30 faq_compil_trouble". If it does not help, you can report this problem to the
31 list but, please, avoid sending a laconic mail like "There is a problem. Is it
32 okay?". Send the config.log file which is automatically generated by
33 configure. Try to capture both the standard output and the error output of the
34 <tt>make</tt> command with <tt>script</tt>. There is no way for us to help you
35 without the relevant bits of information.
37 Now, the following directory should have been created :
39 \li <tt>/home/joe/doc/simgrid/html/</tt>
40 \li <tt>/home/joe/lib/</tt>
41 \li <tt>/home/joe/include/</tt>
43 SimGrid is not a binary, it is a library. Both a static and a dynamic
44 version are available. Here is what you can find if you try a <tt>ls
47 \verbatim libsimgrid.a libsimgrid.la libsimgrid.so libsimgrid.so.0 libsimgrid.so.0.0.1
50 Thus, there is two ways to link your program with SimGrid:
51 \li Either you use the static version, e.g
52 \verbatim gcc libsimgrid.a -o MainProgram MainProgram.c
54 In this case, all the SimGrid functions are directly
55 included in <tt>MainProgram</tt> (hence a bigger binary).
56 \li Either you use the dynamic version (the preferred method)
57 \verbatim gcc -lsimgrid -o MainProgram MainProgram.c
59 In this case, the SimGrid functions are not included in
60 <tt>MainProgram</tt> and you need to set your environment
61 variable in such a way that <tt>libsimgrid.so</tt> will be
62 found at runtime. This can be done by adding the following
63 line in your .bashrc (if you use bash and if you have
64 installed the SimGrid libraries in your home directory):
65 \verbatim export LD_LIBRARY_PATH=$HOME/lib/:$LD_LIBRARY_PATH
69 \subsection faq_compiling_cvs Compiling SimGrid from the CVS
71 The project development takes place in the cvs, where all changes are
72 commited when they happen. Then every once in a while, we make sure that the
73 code quality meets our standard and release an archive from the code in the
74 CVS. We afterward go back to the development in the CVS. So, if you need a
75 recently added feature and can afford some little problem with the stability
76 of the lastest features, you may want to use the CVS version instead of a
79 For that, you first need to get the "simgrid" module from
80 <a href="http://gforge.inria.fr/scm/?group_id=12">here</a>.
82 You won't find any <tt>configure</tt> and a few other things
83 (<tt>Makefile.in</tt>'s, documentation, ...) will be missing as well. The
84 reason for that is that all these files have to be regenerated using the
85 latest versions of <tt>autoconf</tt>, <tt>libtool</tt>, <tt>automake</tt>
86 (>1.9) and <tt>doxygen</tt> (>1.4). To generate the <tt>configure</tt> and
87 the <tt>Makefile.in</tt>'s, you just have to launch the <tt>bootstrap</tt>
88 command that resides in the top of the source tree. Then just follow the
89 instructions of Section \ref faq_compiling.
91 We insist on the fact that you really need the latest versions of
92 autoconf and automake. Doing this step on exotic architectures/systems
93 (i.e. anything different from a recent linux distribution) may be
94 ... uncertain. If you want to use the CVS version on another
95 architecture/system, you should do the previous steps on a perfectly
96 standard box, then do a <tt>make dist</tt> that will build you a
97 perfectly portable SimGrid archive.
99 In summary, the following commands will checkout the CVS, regenerate the
100 configure script and friends, configure SimGrid and build an archive you can
101 use on another machine afterward.
103 \verbatim cvs -d :pserver:anonymous@scm.gforge.inria.fr:/cvsroot/simgrid login
104 cvs -d :pserver:anonymous@scm.gforge.inria.fr:/cvsroot/simgrid checkout simgrid
107 ./configure --enable-maintainer-mode
108 make dist \endverbatim
110 \subsection faq_setting_MSG Setting up your own MSG code
112 Do not build your simulator by modifying the SimGrid examples. Go
113 outside the SimGrid source tree and create your own working directory
114 (say <tt>/home/joe/SimGrid/MyFirstScheduler/</tt>).
116 Suppose your simulation has the following structure (remember it is
117 just an example to illustrate a possible way to compile everything;
118 feel free to organize it as you want).
120 \li <tt>sched.h</tt>: a description of the core of the
121 scheduler (i.e. which functions are can be used by the
122 agents). For example we could find the following functions
123 (master, forwarder, slave).
125 \li <tt>sched.c</tt>: a C file including <tt>sched.h</tt> and
126 implementing the core of the scheduler. Most of these
127 functions use the MSG functions defined in section \ref
130 \li <tt>masterslave.c</tt>: a C file with the main function, i.e.
131 the MSG initialization (MSG_global_init()), the platform
132 creation (e.g. with MSG_create_environment()), the
133 deployment phase (e.g. with MSG_function_register() and
134 MSG_launch_application()) and the call to
137 To compile such a program, we suggest to use the following
138 Makefile. It is a generic Makefile that we have used many times with
139 our students when we teach the C language.
143 masterslave: masterslave.o sched.o
145 INSTALL_PATH = $$HOME
147 PEDANTIC_PARANOID_FREAK = -O0 -Wshadow -Wcast-align \
148 -Waggregate-return -Wmissing-prototypes -Wmissing-declarations \
149 -Wstrict-prototypes -Wmissing-prototypes -Wmissing-declarations \
150 -Wmissing-noreturn -Wredundant-decls -Wnested-externs \
151 -Wpointer-arith -Wwrite-strings -finline-functions
152 REASONABLY_CAREFUL_DUDE = -Wall
153 NO_PRAYER_FOR_THE_WICKED = -w -O2
154 WARNINGS = $(REASONABLY_CAREFUL_DUDE)
155 CFLAGS = -g $(WARNINGS)
157 INCLUDES = -I$(INSTALL_PATH)/include
158 DEFS = -L$(INSTALL_PATH)/lib/
159 LDADD = -lm -lsimgrid
163 $(CC) $(INCLUDES) $(DEFS) $(CFLAGS) $^ $(LIBS) $(LDADD) -o $@
166 $(CC) $(INCLUDES) $(DEFS) $(CFLAGS) -c -o $@ $<
169 rm -f $(BIN_FILES) *.o *~
175 The first two lines indicates what should be build when typing make
176 (<tt>masterslave</tt>) and of which files it is to be made of
177 (<tt>masterslave.o</tt> and <tt>sched.o</tt>). This makefile assumes
178 that you have set up correctly your <tt>LD_LIBRARY_PATH</tt> variable
179 (look, there is a <tt>LDADD = -lm -lsimgrid</tt>). If you prefer using
180 the static version, remove the <tt>-lsimgrid</tt> and add a
181 <tt>$(INSTALL_PATH)/lib/libsimgrid.a</tt> on the next line, right
182 after the <tt>LIBS = </tt>.
184 More generally, if you have never written a Makefile by yourself, type
185 in a terminal : <tt>info make</tt> and read the introduction. The
186 previous example should be enough for a first try but you may want to
187 perform some more complex compilations...
189 \subsection faq_setting_GRAS Setting up your own GRAS code
191 If you use the GRAS interface instead of the MSG one, then previous section
192 is not the better source of information. Instead, you should check the GRAS
193 tutorial in general, and the \ref GRAS_tut_tour_setup in particular.
195 \subsection faq_crosscompile Cross-compiling a Windows DLL of SimGrid from linux
197 At the moment, we do not distribute Windows pre-compiled version of SimGrid
198 because the support for this platform is still experimental. We know that
199 some parts of the GRAS environment do not work, and we think that the others
200 environments (MSG and SD) have good chances to work, but we didn't test
201 ourselves. This section explains how we generate the SimGrid DLL so that you
202 can build it for yourself. First of all, you need to have a version more
203 recent than 3.1 (ie, a CVS version as time of writting).
205 In order to cross-compile the package to windows from linux, you need to
206 install mingw32 (minimalist gnu win32). On Debian, you can do so by
207 installing the packages mingw32 (compiler), mingw32-binutils (linker and
208 so), mingw32-runtime.
210 You can use the VPATH support of configure to compile at the same time for
211 linux and windows without dupplicating the source nor cleaning the tree
212 between each. Just run bootstrap (if you use the CVS) to run the autotools.
213 Then, create a linux and a win directories. Then, type:
214 \verbatim cd linux; ../configure --srcdir=.. <usual configure flags>; make; cd ..
215 cd win; ../configure --srcdir=.. --host=i586-mingw32msvc <flags>; make; cd ..
217 The trick to VPATH builds is to call configure from another directory,
218 passing it an extra --srcdir argument to tell it where all the sources are.
219 It will understand you want to use VPATH. Then, the trick to cross-compile
220 is simply to add a --host argument specifying the target you want to build
221 for. The i586-mingw32msvc string is what you have to pass to use the mingw32
222 environment as distributed in Debian.
224 After that, you can run all make targets from both directories, and test
225 easily that what you change for one arch does not break the other one.
227 It is possible that this VPATH build thing breaks from time to time in the
228 CVS since it's quite fragile, but it's granted to work in any released
229 version. If you experience problems, drop us a mail.
231 Another possible source of issue is that at the moment, building the
232 examples request to use the gras_stub_generator tool, which is a compiled
233 program, not a script. In cross-compilation, you need to cross-execute with
234 wine for example, which is not really pleasant. We are working on this, but
235 in the meanwhile, simply don't build the examples in cross-compilation
236 (<tt>cd src</tt> before running make).
238 Program (cross-)compiled with mingw32 do request an extra DLL at run-time to be
239 usable. For example, if you want to test your build with wine, you should do
240 the following to put this library where wine looks for DLLs.
241 \verbatim cp /usr/share/doc/mingw32-runtime/mingwm10.dll.gz ~/.wine/c/windows/system/
242 gunzip ~/.wine/c/windows/system/mingwm10.dll.gz
245 The DLL is builded in src/.libs, and installed in the <prefix>/bin directory
246 when you run make install.
248 If you want to use it in a native project on windows, you need to use
249 simgrid.dll and mingwm10.dll. For each DLL, you need to build .def file
250 under linux (listing the defined symbols), and convert it into a .lib file
251 under windows (specifying this in a way that windows compilers like). To
252 generate the def files, run (under linux):
253 \verbatim echo "LIBRARY libsimgrid-0.dll" > simgrid.def
254 echo EXPORTS >> simgrid.def
255 nm libsimgrid-0.dll | grep ' T _' | sed 's/.* T _//' >> simgrid.def
256 nm libsimgrid-0.dll | grep ' D _' | sed 's/.* D _//' | sed 's/$/ DATA/' >> simgrid.def
258 echo "LIBRARY mingwm10.dll" > mingwm10.def
259 echo EXPORTS >> mingwm10.def
260 nm mingwm10.dll | grep ' T _' | sed 's/.* T _//' >> mingwm10.def
261 nm mingwm10.dll | grep ' D _' | sed 's/.* D _//' | sed 's/$/ DATA/' >> mingwm10.def
264 To create the import .lib files, use the <tt>lib</tt> windows tool (from
265 MSVC) the following way to produce simgrid.lib and mingwm10.lib
266 \verbatim lib /def:simgrid.def
267 lib /def:mingwm10.def
270 Then, set the following parameters in Visual C++ 2005:
271 Linker -> Input -> Additional dependencies = simgrid.lib mingwm10.lib
273 Just in case you wonder how to generate a DLL from libtool in another
274 project, we added -no-undefined to any lib*_la_LDFLAGS variables so that
275 libtool accepts to generate a dynamic library under windows. Then, to make
276 it true, we pass any dependencies (such as -lws2 under windows or -lpthread
277 on need) on the linking line. Passing such deps is a good idea anyway so
278 that they get noted in the library itself, avoiding the users to know about
279 our dependencies and put them manually on their compilation line. Then we
280 added the AC_LIBTOOL_WIN32_DLL macro just before AC_PROG_LIBTOOL in the
281 configure.ac. It means that we exported any symbols which need to be.
282 Nowadays, functions get automatically exported, so we don't need to load our
283 header files with tons of __declspec(dllexport) cruft. We only need to do so
284 for data, but there is no public data in SimGrid so we are good.
286 \section faq_simgrid I'm new to SimGrid. I have some questions. Where should I start?
288 You are at the right place... Having a look to these
289 <a href="http://graal.ens-lyon.fr/~alegrand/articles/slides_g5k_simul.pdf">slides</a>
291 <a href="http://graal.ens-lyon.fr/~alegrand/articles/Simgrid-Introduction.pdf">"obsolete" slides</a>)
292 may give you some insights on what SimGrid can help you to do and what
293 are its limitations. Then you definitely should read the \ref
294 MSG_examples. There is also a mailing list: <simgrid-user@lists.gforge.inria.fr>.
296 \subsection faq_generic Building a generic simulator
298 Please read carefully the \ref MSG_examples. You'll find in \ref
299 MSG_ex_master_slave a very simple consisting of a master (that owns a bunch of
300 tasks and distributes them) , some slaves (that process tasks whenever
301 they receive one) and some forwarder agents (that simply pass the
302 tasks they receive to some slaves).
304 \subsection faq_visualization Visualizing the schedule
306 It is sometime convenient to "see" how the agents are behaving. If you
307 like colors, you can use <tt>tools/MSG_visualization/colorize.pl </tt>
308 as a filter to your MSG outputs. It works directly with INFO. Beware,
309 INFO() prints on stderr. Do not forget to redirect if you want to
310 filter (e.g. with bash):
312 ./msg_test small_platform.xml small_deployment.xml 2>&1 | ../../tools/MSG_visualization/colorize.pl
315 We also have a more graphical output. Have a look at MSG_paje_output(). It
316 generates an input to <a href="http://www-id.imag.fr/Logiciels/paje/">Paje</a>.
319 <a href="Paje_MSG_screenshot.jpg"><img src="Paje_MSG_screenshot_thn.jpg"></a>
323 \subsection faq_postmortem_analysis Online/postmortem analysis
325 Vizualization with Paje can be seen as a kind of postmortem
326 analysis. However, as soon as you start playing with big simulations,
327 you'll realize that processing such output is kind of tricky. There is
328 so much generic informations that it is hard to find the information
331 As a matter of fact, loging really depends on simulations (e.g. what
332 kind of events is important...). That is why we do not propose a big
333 dump of your whole simulation (it would slow everything down) but give
334 you neat tools to structure you logs. Have a look at \ref XBT_log. In
335 fact, rather than a post-mortem analysis, you may want to do it on the
336 fly. The process you are running can do whatever you want. Have you
337 thought about adding a global structure where you directly compute the
338 informations that are really important rather than writing everything
339 down and then processing huge files ?
341 \subsection faq_C Argh! Do I really have to code in C ?
343 Up until now, there is no binding for other languages. If you use C++,
344 you should be able to use the SimGrid library as a standard C library
345 and everything should work fine (simply <i>link</i> against this
346 library; recompiling SimGrid with a C++ compiler won't work and it
347 wouldn't help if you could).
349 In fact, the bindings needed to allow one to use SimGrid from Perl,
350 Python, Java, etc. are double-layered. The first layer would allow
351 you to call for example the MSG_task_get_name(task) function while
352 what you really want is a proper object wrapping allowing you to call
353 task->name(). That's the purpose of the second layer. The first one
354 is granted with C++ but can be done with tools like
355 <a href="www.swig.org/">swig</a> for other languages like Perl, Ruby,
356 Python, CAML. None of us really need the second one (which is a bit
357 more demanding and cannot be automatically generated) yet and there is
358 no real point in doing the first one without the second. :)
360 As usual, you're welcome to participate.
362 \section faq_MIA How to ....? Is there a function in the API to simply ....?
364 Here is the deal. The whole SimGrid project (MSG, SURF, GRAS, ...) is
365 meant to be kept as simple and generic as possible. We cannot add
366 functions for everybody's need when these functions can easily be
367 built from the ones already in the API. Most of the time, it is
368 possible and when it was not possible we always have upgraded the API
369 accordingly. When somebody asks us a question like "How to do that ?
370 Is there a function in the API to simply do this ?", we're always glad
371 to answer and help. However if we don't need this code for our own
372 need, there is no chance we're going to write it... it's your job! :)
373 The counterpart to our answers is that once you come up with a neat
374 implementation of this feature (task duplication, RPC, thread
375 synchronization, ...), you should send it to us and we will be glad to
376 add it to the distribution. Thus, other people will take advantage of
377 it (and we don't have to answer this question again and again ;).
379 You'll find in this section a few "Missing In Action" features. Many
380 people have asked about it and we have given hints on how to simply do
381 it with MSG. Feel free to contribute...
383 \subsection faq_MIA_examples I want some more complex examples!
385 Many people have come to ask me a more complex example and each time,
386 they have realized afterward that the basics were in the previous three
389 Of course they have often been needing more complex functions like
390 MSG_process_suspend(), MSG_process_resume() and
391 MSG_process_isSuspended() (to perform synchronization), or
392 MSG_task_Iprobe() and MSG_process_sleep() (to avoid blocking
393 receptions), or even MSG_process_create() (to design asynchronous
394 communications or computations). But the examples are sufficient to
397 We know. We should add some more examples, but not really some more
398 complex ones... We should add some examples that illustrate some other
399 functionalities (like how to simply encode asynchronous
400 communications, RPC, process migrations, thread synchronization, ...)
401 and we will do it when we will have a little bit more time. We have
402 tried to document the examples so that they are understandable. Tell
403 us if something is not clear and once again feel free to participate!
406 \subsection faq_MIA_taskdup Missing in action: Task duplication/replication
408 There is no task duplication in MSG. When you create a task, you can
409 process it or send it somewhere else. As soon as a process has sent
410 this task, he doesn't have this task anymore. It's gone. The receiver
411 process has got the task. However, you could decide upon receiving to
412 create a "copy" of a task but you have to handle by yourself the
413 semantic associated to this "duplication".
415 As we already told, we prefer keeping the API as simple as
416 possible. This kind of feature is rather easy to implement by users
417 and the semantic you associate really depends on people. Having a
418 *generic* task duplication mechanism is not that trivial (in
419 particular because of the data field). That is why I would recommand
420 that you write it by yourself even if I can give you advice on how to
423 You have the following functions to get informations about a task:
424 MSG_task_get_name(), MSG_task_get_compute_duration(),
425 MSG_task_get_remaining_computation(), MSG_task_get_data_size(),
426 and MSG_task_get_data().
428 You could use a dictionnary (#xbt_dict_t) of dynars (#xbt_dynar_t). If
429 you still don't see how to do it, please come back to us...
431 \subsection faq_MIA_asynchronous I want to do asynchronous communications in MSG
433 Up until now, there is no asynchronous communications in MSG. However,
434 you can create as many process as you want so you should be able to do
435 whatever you want... I've written a queue module to help implementing
436 some asynchronous communications at low cost (creating thousands of
437 process only to handle communications may be problematic in term of
438 performance at some point). I'll add it in the distribution asap.
440 \subsection faq_MIA_thread_synchronization I need to synchronize my MSG processes
442 You obviously cannot use pthread_mutexes of pthread_conds. The best
443 thing would be to propose similar structures. Unfortunately, we
444 haven't found time to do it yet. However you can try to play with
445 MSG_process_suspend() and MSG_process_resume(). You can even do some
446 synchronization with fake communications (using MSG_task_get(),
447 MSG_task_put() and MSG_task_Iprobe()).
449 \subsection faq_MIA_host_load Where is the get_host_load function hidden in MSG?
451 There is no such thing because its semantic wouldn't be really
452 clear. Of course, it is something about the amount of host throughput,
453 but there is as many definition of "host load" as people asking for
454 this function. First, you have to remember that resource availability
455 may vary over time, which make any load notion harder to define.
457 It may be instantaneous value or an average one. Moreover it may be only the
458 power of the computer, or may take the background load into account, or may
459 even take the currently running tasks into account. In some SURF models,
460 communications have an influence on computational power. Should it be taken
463 So, we decided not to include such a function into MSG and let people do it
464 thereselves so that they get the value matching exactly what they mean. One
465 possibility is to run active measurement as in next code snippet. It is very
466 close from what you would have to do out of the simulator, and thus gives
467 you information that you could also get in real settings to not hinder the
468 realism of your simulation.
471 double get_host_load() {
472 m_task_t task = MSG_task_create("test", 0.001, 0, NULL);
473 double date = MSG_get_clock();
475 MSG_task_execute(task);
476 date = MSG_get_clock() - date;
477 MSG_task_destroy(task);
482 Of course, it may not match your personal definition of "host load". In this
483 case, please detail what you mean on the mailing list, and we will extend
484 this FAQ section to fit your taste if possible.
486 \subsection faq_MIA_communication_time How can I get the *real* communication time ?
488 Communications are synchronous and thus if you simply get the time
489 before and after a communication, you'll only get the transmission
490 time and the time spent to really communicate (it will also take into
491 account the time spent waiting for the other party to be
492 ready). However, getting the *real* communication time is not really
493 hard either. The following solution is a good starting point.
498 m_task_t task = MSG_task_create("Task", task_comp_size, task_comm_size,
499 calloc(1,sizeof(double)));
500 *((double*) task->data) = MSG_get_clock();
501 MSG_task_put(task, slaves[i % slaves_count], PORT_22);
502 INFO0("Send completed");
507 m_task_t task = NULL;
510 time1 = MSG_get_clock();
511 a = MSG_task_get(&(task), PORT_22);
512 time2 = MSG_get_clock();
513 if(time1<*((double *)task->data))
514 time1 = *((double *) task->data);
515 INFO1("Communication time : \"%f\" ", time2-time1);
517 MSG_task_destroy(task);
522 \subsection faq_MIA_batch_scheduler Is there a native support for batch schedulers in SimGrid ?
524 No, there is no native support for batch schedulers and none is
525 planned because this is a very specific need (and doing it in a
526 generic way is thus very hard). However some people have implemented
527 their own batch schedulers. Vincent Garonne wrote one during his PhD
528 and put his code in the contrib directory of our CVS so that other can
529 keep working on it. You may find inspinring ideas in it.
531 \subsection faq_MIA_checkpointing I need a checkpointing thing
533 Actually, it depends on whether you want to checkpoint the simulation, or to
534 simulate checkpoints.
536 The first one could help if your simulation is a long standing process you
537 want to keep running even on hardware issues. It could also help to
538 <i>rewind</i> the simulation by jumping sometimes on an old checkpoint to
539 cancel recent calculations.\n
540 Unfortunately, such thing will probably never exist in SG. One would have to
541 duplicate all data structures because doing a rewind at the simulator level
542 is very very hard (not talking about the malloc free operations that might
543 have been done in between). Instead, you may be interested in the Libckpt
544 library (http://www.cs.utk.edu/~plank/plank/www/libckpt.html). This is the
545 checkpointing solution used in the condor project, for example. It makes it
546 easy to create checkpoints (at the OS level, creating something like core
547 files), and rerunning them on need.
549 If you want to simulate checkpoints instead, it means that you want the
550 state of an executing task (in particular, the progress made towards
551 completion) to be saved somewhere. So if a host (and the task executing on
552 it) fails (cf. #MSG_HOST_FAILURE), then the task can be restarted
553 from the last checkpoint.\n
555 Actually, such a thing does not exists in SimGrid either, but it's just
556 because we don't think it is fundamental and it may be done in the user code
557 at relatively low cost. You could for example use a watcher that
558 periodically get the remaining amount of things to do (using
559 MSG_task_get_remaining_computation()), or fragment the task in smaller
562 \section faq_SG Where has SG disappeared?!?
564 OK, it's time to explain what's happening to the SimGrid project. Let's
565 start with a little bit of history.
567 * Historically, SimGrid was a low-level toolkit for scheduling with
568 classical models such as DAGs. That was SimGrid v.1.* aka SG, written
569 by Henri Casanova. I (Arnaud) had been using it in its earliest
570 versions during an internship at UCSD.
572 Then we have realized that encoding distributed algorithm in SG was a
575 * So we have built MSG on top of SG and have released SimGrid v.2.*. MSG
576 offered a very basic API to encode a distributed application easily.
577 However encoding MSG on top of SG was not really convenient and did not
578 use the DAG part since the control of the task synchronization was done
579 on top of MSG and no more in SG. We have been playing a little bit with
580 MSG. We have realized that:
582 \li 1) the platform modeling was quite flexible and could be "almost"
583 automated (e.g. using random generator and post-annotations);
585 \li 2) SG was the bottleneck because of the way we were using
586 it. We needed to simulate concurrent transfers, complex load
587 sharing mechanisms. Many optimizations (e.g. trace integration)
588 were totally inefficient when combined with MSG and made extending SG
589 to implement new sharing policies, parallel tasks models, or failures
590 (many people were asking for these kind of features) a real pain;
592 \li 3) the application modeling was not really easy. Even though the
593 application modeling depends on people's applications, we thought
594 we could improve things here. One of our target here was realistic
595 distributed applications ranging from computer sensor networks like
596 the NWS to peer-to-peer applications;
598 * So we have been planning mainly two things for SimGrid 3:
600 \li 1) I have proposed to get rid of SG and to re-implement a new kernel
601 that would be faster and more flexible. That is what I did in the
602 end of 2004: SURF. SURF is based on a fast max-min linear solver
603 using O(1) data-structures. I have quickly replaced SG by SURF in
604 MSG and the result has been that on the MSG example, the new
605 version was more than 10 times faster while we had gain a lot of
606 flexibility. I think I could still easily make MSG faster but I
607 have to work on MSG now (e.g. using some of the O(1)
608 data-structures I've been using to build SURF) since it has become
609 the bottleneck. Some MSG functions have been removed from the API
610 but they were mainly intended to build the platform by hand (they
611 had appeared in the earliest versions of MSG) and were therefore
612 not useful anymore since we are providing a complete mechanism to
613 automatically build the platform and deploy the agents on it.;
615 \li 2) GRAS is a new project Martin and I have come up with. The idea is
616 to have a programming environment that let you program real
617 distributed applications while letting you the ability to run it in
618 the simulator without having to change the slightest line of your
619 code. From the simulation point of view, GRAS performs the
620 application modeling automatically... Up until now, GRAS works on
621 top MSG for historical reasons but I'm going to make it work
622 directly on top of SURF so that it can use all the flex and the
623 speed provided by SURF.
625 Those two things are working, but we want to make everything as clean as
626 possible before releasing SimGrid v.3.
628 So what about those nice DAGs we used to have in SimGrid v.1.? They're
629 not anymore in SimGrid v.3. At least not in their original form... Let
630 me recall you the way SimGrid 3 is organized:
644 XBT is our tool box and now, you should have an idea of what the other
645 ones are. As you can see, the primitive SG is not here
646 anymore. However we have written a brand new and cleaner API for this
647 purpose: \ref SD_API. It is built directly on top of SURF and provides
648 an API rather close to the old SG:
651 ______________________
653 |____________________|
654 | | MSG | GRAS | SD |
655 | -------------------|
657 | -------------------|
659 ----------------------
662 The nice thing is that, as it is writen on top of SURF, it seamlessly
663 support DAG of parallel tasks as well as complex communications
664 patterns. Some old codes using SG are currently under rewrite using
665 \ref SD_API to check that all needful functions are provided.
667 \subsection faq_SG_DAG How to implement a distributed dynamic scheduler of DAGs.
669 Distributed is somehow "contagious". If you start making distributed
670 decisions, there is no way to handle DAGs directly anymore (unless I
671 am missing something). You have to encode your DAGs in term of
672 communicating process to make the whole scheduling process
673 distributed. Here is an example of how you could do that. Assume T1
674 has to be done before T2.
677 int your_agent(int argc, char *argv[] {
679 T1 = MSG_task_create(...);
680 T2 = MSG_task_create(...);
684 if(cond) MSG_task_execute(T1);
686 if((MSG_task_get_remaining_computation(T1)=0.0) && (you_re_in_a_good_mood))
689 /* do something else */
695 If you decide that the distributed part is not that much important and that
696 DAG is really the level of abstraction you want to work with, then you should
697 give a try to \ref SD_API.
699 \section faq_dynamic Dynamic resources and platform building
701 \subsection faq_platform Building a realistic platform
703 We can speak more than an hour on this subject and we still do not have
704 the right answer, just some ideas. You can read the following
705 <a href="http://graal.ens-lyon.fr/~alegrand/articles/Simgrid-Introduction.pdf">slides</a>.
706 It may give you some hints. You can also have a look at the
707 <tt>tools/platform_generation/</tt> directory. There is a perl-script
708 we use to annotate a Tiers generated platform.
710 \subsection faq_SURF_dynamic How can I have variable resource availability?
712 A nice feature of SimGrid is that it enables you to seamlessly have
713 resources whose availability change over time. When you build a
714 platform, you generally declare CPUs like that:
717 <cpu name="Cpu A" power="100.00"/>
720 If you want the availability of "CPU A" to change over time, the only
721 thing you have to do is change this definition like that:
724 <cpu name="Cpu A" power="100.00" availability_file="trace_A.txt" state_file="trace_A_failure.txt"/>
727 For CPUs, availability files are expressed in fraction of available
728 power. Let's have a look at what "trace_A.txt" may look like:
737 At time 0, our CPU will deliver 100 Mflop/s. At time 11.0, it will
738 deliver only 50 Mflop/s until time 20.0 where it will will start
739 delivering 90 Mflop/s. Last at time 21.0 (20.0 plus the periodicity
740 1.0), we'll be back to the beginning and it will deliver 100Mflop/s.
742 Now let's look at the state file:
749 A negative value means "off" while a positive one means "on". At time
750 1.0, the CPU is on. At time 1.0, it is turned off and at time 2.0, it
751 is turned on again until time 12 (2.0 plus the periodicity 10.0). It
752 will be turned on again at time 13.0 until time 23.0, and so on.
754 Now, let's look how the same kind of thing can be done for network
755 links. A usual declaration looks like:
758 <network_link name="LinkA" bandwidth="10.0" latency="0.2"/>
761 You have at your disposal the following options: bandwidth_file,
762 latency_file and state_file. The only difference with CPUs is that
763 bandwidth_file and latency_file do not express fraction of available
764 power but are expressed directly in Mb/s and seconds.
766 \subsection faq_flexml_bypassing How can I have some C functions do what the platform file does?
768 So you want to bypass the XML files parser, uh? Maybe doin some parameter
769 sweep experiments on your simulations or so? This is possible, but it's not
770 really easy. Here is how it goes.
772 For this, you have to first remember that the XML parsing in SimGrid is done
773 using a tool called FleXML. Given a DTD, this gives a flex-based parser. If
774 you want to bypass the parser, you need to provide some code mimicking what
775 it does and replacing it in its interactions with the SURF code. So, let's
776 have a look at these interactions.
778 FleXML parser are close to classical SAX parsers. It means that a
779 well-formed SimGrid platform XML file might result in the following
782 - start "platform_description"
783 - start "cpu" with attributes name="host1" power="1.0"
785 - start "cpu" with attributes name="host2" power="2.0"
787 - start "network_link" with ...
789 - start "route" with ...
791 - start "route" with ...
793 - end "platform_description"
795 The communication from the parser to the SURF code uses two means:
796 Attributes get copied into some global variables, and a surf-provided
797 function gets called by the parser for each event. For example, the event
798 - start "cpu" with attributes name="host1" power="1.0"
800 let the parser do the equivalent of:
802 strcpy("host1",A_cpu_name);
807 In SURF, we attach callbacks to the different events by initializing the
808 pointer functions to some the right surf functions. Example in
809 workstation_KCCFLN05.c (surf_parse_open() ends up calling surf_parse()):
811 // Building the routes
812 surf_parse_reset_parser();
813 STag_route_fun=parse_route_set_endpoints;
814 ETag_route_element_fun=parse_route_elem;
815 ETag_route_fun=parse_route_set_route;
816 surf_parse_open(file);
817 xbt_assert1((!surf_parse()),"Parse error in %s",file);
821 So, to bypass the FleXML parser, you need to write your own version of the
822 surf_parse function, which should do the following:
823 - Call the corresponding STag_<tag>_fun function to simulate tag start
824 - Fill the A_<tag>_<attribute> variables with the wanted values
825 - Call the corresponding ETag_<tag>_fun function to simulate tag end
826 - (do the same for the next set of values, and loop)
828 Then, tell SimGrid that you want to use your own "parser" instead of the stock one:
830 surf_parse = surf_parse_bypass;
831 MSG_create_environment(NULL);
834 An example of this trick is distributed in the file examples/msg/msg_test_surfxml_bypassed.c
836 \section faq_limits Pushing the limits
838 \subsection faq_context_1000 I want thousands of simulated processes
840 SimGrid can use either pthreads library or the UNIX98 contextes. On most
841 systems, the number of pthreads is limited and then your simulation may be
842 limited for a stupid reason. This is especially true with the current linux
843 pthreads, and I cannot get more than 2000 simulated processes with pthreads
844 on my box. The UNIX98 contexts allow me to raise the limit to 25,000
845 simulated processes on my laptop.
847 The <tt>--with-context</tt> option of the <tt>./configure</tt> script allows
848 you to choose between UNIX98 contextes (<tt>--with-context=ucontext</tt>)
849 and the pthread version ( (<tt>--with-context=pthread</tt>). The default
850 value is ucontext when the script detect a working UNIX98 context
851 implementation. On Windows boxes, the provided value is discarded and an
852 adapted version is picked up.
854 We experienced some issues with contextes on some rare systems (solaris 8
855 and lower or old alpha linuxes comes to mind). The main problem is that the
856 configure script detect the contextes as being functional when it's not
857 true. If you happen to use such a system, switch manually to the pthread
858 version, and provide us with a good patch for the configure script so that
859 it is done automatically ;)
861 \subsection faq_context_10000 I want hundred thousands of simulated processes
863 As explained above, SimGrid can use UNIX98 contextes to represent and handle
864 the simulated processes. Thanks to this, the main limitation to the number
865 of simulated processes becomes the available memory.
867 Here are some tricks I had to use in order to run a token ring between
868 25,000 processes on my laptop (1Gb memory, 1.5Gb swap).
870 - First of all, make sure your code runs for a few hundreds processes
871 before trying to push the limit. Make sure it's valgrind-clean, ie that
872 valgrind does not report neither memory error nor memory leaks. Indeed,
873 numerous simulated processes result in *fat* simulation hindering debugging.
875 - It was really boring to write 25,000 entries in the deployment file, so I wrote
876 a little script <tt>examples/gras/tokenS/make_deployment.pl</tt>, which you may
877 want to adapt to your case. You could also think about hijacking
878 the SURFXML parser (have look at \ref faq_flexml_bypassing).
880 - The deployment file became quite big, so I had to do what is in the FAQ
881 entry \ref faq_flexml_limit
883 - Each UNIX98 context has its own stack entry. As debugging this is quite
884 hairly, the default value is a bit overestimated so that user don't get
885 into trouble about this. You want to tune this size to increse the number
886 of processes. This is the <tt>STACK_SIZE</tt> define in
887 <tt>src/xbt/context_private.h</tt>, which is 128kb by default.
888 Reduce this as much as you can, but be warned that if this value is too
889 low, you'll get a segfault. The token ring example, which is quite simple,
890 runs with 40kb stacks.
892 \section faq_troubleshooting Troubleshooting
894 \subsection faq_compil_trouble ./configure fails!
896 We now only one reason for the configure to fail:
898 - <b>You are using a borken build environment</b>\n
899 If symptom is that configure complains about gcc not being able to build
900 executables, you are probably missing the libc6-dev package. Damn Ubuntu.
902 If you experience other kind of issue, please get in touch with us. We are
903 always interested in improving our portability to new systems.
905 \subsection faq_distcheck_fails Dude! "make check" fails on my machine!
907 Don't assume we never run this target, because we do. Really. Promise!
909 There is several reasons which may cause the make check to fail on your
912 - <b>You are using a borken libc (probably concerning the contextes)</b>.\n
913 The symptom is that the "make check" fails within the examples/msg directory.\n
914 By default, SimGrid uses something called ucontexts. This is part of the
915 libc, but it's quite undertested. For example, some (old) versions of the
916 glibc on alpha do not implement these functions, but provide the stubs
917 (which return ENOSYS: not implemented). It fools our detection mecanism
918 and leads to segfaults.\n
919 On some x86_64, the pointer to function is stored into a integer, but int
920 are 32bits only on this arch while pointers are 64bits. Our detection
921 mecanism also fails to detect the problem, which leads to segfaults.\n
922 In both cases, there is not much we can do to fix the bug. We are working
923 on a workaround for x86_64 machines, but in the meanwhile, you can
924 compile with --with-context=pthread to avoid ucontext completely. You'll
925 be a bit more limitated in the number of simulated processes you can start
926 concurently, but 5000 processes is still enough for most purposes, isn't
928 This limitation is the reason why we insist on using this piece of ...
929 software even if it's so troublesome.\n
930 <b>=> use --with-pthread on AMD64 architecture that do not have an
931 ultra-recent libc.</b>
933 - <b>There is a bug in SimGrid we aren't aware of</b>.\n
934 If none of the above apply, please drop us a mail on the mailing list so
935 that we can check it out.
937 \subsection faq_longjmp longjmp madness in valgrind
939 This is when valgrind starts complaining about longjmp things, just like:
941 \verbatim ==21434== Conditional jump or move depends on uninitialised value(s)
942 ==21434== at 0x420DBE5: longjmp (longjmp.c:33)
944 ==21434== Use of uninitialised value of size 4
945 ==21434== at 0x420DC3A: __longjmp (__longjmp.S:48)
948 or even when it reports scary things like:
950 \verbatim ==24023== Warning: client switching stacks? SP change: 0xBE3FF618 --> 0xBE7FF710
951 x86->IR: unhandled instruction bytes: 0xF4 0xC7 0x83 0xD0
952 ==24023== to suppress, use: --max-stackframe=4194552 or greater
953 ==24023== Your program just tried to execute an instruction that Valgrind
954 ==24023== did not recognise. There are two possible reasons for this.
955 ==24023== 1. Your program has a bug and erroneously jumped to a non-code
956 ==24023== location. If you are running Memcheck and you just saw a
957 ==24023== warning about a bad jump, it's probably your program's fault.
958 ==24023== 2. The instruction is legitimate but Valgrind doesn't handle it,
959 ==24023== i.e. it's Valgrind's fault. If you think this is the case or
960 ==24023== you are not sure, please let us know.
961 ==24023== Either way, Valgrind will now raise a SIGILL signal which will
962 ==24023== probably kill your program.
964 ==24023== Process terminating with default action of signal 4 (SIGILL)
965 ==24023== Illegal opcode at address 0x420D234
966 ==24023== at 0x420D234: abort (abort.c:124)
969 This is the sign that you didn't used the exception mecanism well. Most
970 probably, you have a <tt>return;</tt> somewhere within a <tt>TRY{}</tt>
971 block. This is <b>evil</b>, and you must not do this. Did you read the section
974 \subsection faq_valgrind Valgrind spits tons of errors!
976 It may happen that valgrind, the memory debugger beloved by any decent C
977 programmer, spits tons of warnings like the following :
978 \verbatim ==8414== Conditional jump or move depends on uninitialised value(s)
979 ==8414== at 0x400882D: (within /lib/ld-2.3.6.so)
980 ==8414== by 0x414EDE9: (within /lib/tls/i686/cmov/libc-2.3.6.so)
981 ==8414== by 0x400B105: (within /lib/ld-2.3.6.so)
982 ==8414== by 0x414F937: _dl_open (in /lib/tls/i686/cmov/libc-2.3.6.so)
983 ==8414== by 0x4150F4C: (within /lib/tls/i686/cmov/libc-2.3.6.so)
984 ==8414== by 0x400B105: (within /lib/ld-2.3.6.so)
985 ==8414== by 0x415102D: __libc_dlopen_mode (in /lib/tls/i686/cmov/libc-2.3.6.so)
986 ==8414== by 0x412D6B9: backtrace (in /lib/tls/i686/cmov/libc-2.3.6.so)
987 ==8414== by 0x8076446: xbt_dictelm_get_ext (dict_elm.c:714)
988 ==8414== by 0x80764C1: xbt_dictelm_get (dict_elm.c:732)
989 ==8414== by 0x8079010: xbt_cfg_register (config.c:208)
990 ==8414== by 0x806821B: MSG_config (msg_config.c:42)
993 This problem is somewhere in the libc when using the backtraces and there is
994 very few things we can do ourselves to fix it. Instead, here is how to tell
995 valgrind to ignore the error. Add the following to your ~/.valgrind.supp (or
996 create this file on need). Make sure to change the obj line according to
997 your personnal mileage (change 2.3.6 to the actual version you are using,
998 which you can retrieve with a simple "ls /lib/ld*.so").
1001 name: Backtrace madness
1003 obj:/lib/ld-2.3.6.so
1008 fun:__libc_dlopen_mode
1011 Then, you have to specify valgrind to use this suppression file by passing
1012 the <tt>--suppressions=$HOME/.valgrind.supp</tt> option on the command line.
1013 You can also add the following to your ~/.bashrc so that it gets passed
1014 automatically. Actually, it passes a bit more options to valgrind, and this
1015 happen to be my personnal settings. Check the valgrind documentation for
1018 \verbatim export VALGRIND_OPTS="--leak-check=yes --leak-resolution=high --num-callers=40 --tool=memcheck --suppressions=$HOME/.valgrind.supp" \endverbatim
1020 \subsection faq_flexml_limit I get the message "surf_parse_lex: Assertion `next<limit' failed."
1022 This is because your platform file is too big for the parser.
1024 Actually, the message comes directly from FleXML, the technology on top of
1025 which the parser is built. FleXML has the bad idea of fetching the whole
1026 document in memory before parsing it. And moreover, the memory buffer size
1027 must be determinded at compilation time.
1029 We use a value which seems big enough for our need without bloating the
1030 simulators footprints. But of course your mileage may vary. In this case,
1031 just edit src/surf/surfxml.l modify the definition of
1032 FLEXML_BUFFERSTACKSIZE. E.g.
1035 #define FLEXML_BUFFERSTACKSIZE 1000000000
1038 Then recompile and everything should be fine, provided that your version of
1039 Flex is recent enough (>= 2.5.31). If not the compilation process should
1042 A while ago, we worked on FleXML to reduce a bit its memory consumtion, but
1043 these issues remain. There is two things we should do:
1045 - use a dynamic buffer instead of a static one so that the only limit
1046 becomes your memory, not a stupid constant fixed at compilation time
1047 (maybe not so difficult).
1048 - change the parser so that it does not need to get the whole file in
1049 memory before parsing
1050 (seems quite difficult, but I'm a complete newbe wrt flex stuff).
1052 These are changes to FleXML itself, not SimGrid. But since we kinda hijacked
1053 the development of FleXML, I can grant you that any patches would be really
1054 welcome and quickly integrated.
1056 \subsection faq_gras_transport GRAS spits networking error messages
1058 Gras, on real platforms, naturally use regular sockets to communicate. They
1059 are deeply hiden in the gras abstraction, but when things go wrong, you may
1060 get some weird error messages. Here are some example, with the probable
1063 - <b>Transport endpoint is not connected</b>: several processes try to open
1064 a server socket on the same port number of the same machine. This is
1065 naturally bad and each process should pick its own port number for this.\n
1066 Maybe, you just have some processes remaining from a previous experiment
1068 Killing them may help, but again if you kill -KILL them, you'll have to
1069 wait for a while: they didn't close there sockets properly and the system
1070 needs a while to notice that this port is free again.
1072 - <b>Socket closed by remote side</b>: if the remote process is not
1073 supposed to close the socket at this point, it may be dead.
1075 - <b>Connection reset by peer</b>: I found this on internet about this
1076 error. I think it's what's happening here, too:\n
1077 <i>This basically means that a network error occurred while the client was
1078 receiving data from the server. But what is really happening is that the
1079 server actually accepts the connection, processes the request, and sends
1080 a reply to the client. However, when the server closes the socket, the
1081 client believes that the connection has been terminated abnormally
1082 because the socket implementation sends a TCP reset segment telling the
1083 client to throw away the data and report an error.\n
1084 Sometimes, this problem is caused by not properly closing the
1085 input/output streams and the socket connection. Make sure you close the
1086 input/output streams and socket connection properly. If everything is
1087 closed properly, however, and the problem persists, you can work around
1088 it by adding a one-second sleep before closing the streams and the
1089 socket. This technique, however, is not reliable and may not work on all
1091 Since GRAS sockets are closed properly (repeat after me: there is no bug
1092 in GRAS), it is either that you are closing your sockets on server side
1093 before the client get a chance to read them (use gras_os_sleep() to delay
1094 the server), or the server died awfully before the client got the data.
1096 \subsection faq_deadlock There is a deadlock !!!
1098 Unfortunately, we cannot debug every code written in SimGrid. We
1099 furthermore believe that the framework provides ways enough
1100 information to debug such informations yourself. If the textual output
1101 is not enough, Make sure to check the \ref faq_visualization FAQ entry to see
1102 how to get a graphical one.
1104 Now, if you come up with a really simple example that deadlocks and
1105 you're absolutely convinced that it should not, you can ask on the
1106 list. Just be aware that you'll be severely punished if the mistake is
1107 on your side... We have plenty of FAQ entries to redact and new
1108 features to implement for the impenitents! ;)
1110 \subsection faq_big_fat_warning A BIG FAT WARNING is reported telling me that my platform and deployment files are too old.
1112 We have decided to change the units in SimGrid. Now we use Bytes, Flops and
1113 seconds instead of MBytes, MFlops and seconds... Units should be updated
1114 accordingly and the version of platform_description should be set to a
1115 valuer greater than 1:
1117 <platform_description version="1">
1119 You should try to use the surfxml_update.pl script that can be found
1120 <a href="http://gforge.inria.fr/plugins/scmcvs/cvsweb.php/contrib/platform_generation/?cvsroot=cvsroot%2Fsimgrid">here</a>.
1122 \subsection faq_surf_network_latency I get weird timings when I play with the latencies.
1124 OK, first of all, remember that units should be Bytes, Flops and
1125 Seconds. If you don't use such units, some SimGrid constants (e.g. the
1126 SG_TCP_CTE_GAMMA constant used in most network models) won't have the
1127 right unit and you'll end up with weird results.
1129 Here is what happens with a single transfer of size L on a link
1130 (bw,lat) when nothing else happens.
1133 0-----lat--------------------------------------------------t
1134 |-----|**** real_bw =min(bw,SG_TCP_CTE_GAMMA/(2*lat)) *****|
1137 In more complex situations, this min is the solution of a complex
1138 max-min linear system. Have a look
1139 <a href="http://lists.gforge.inria.fr/pipermail/simgrid-devel/2006-April/thread.html">here</a>
1140 and read the two threads "Bug in SURF?" and "Surf bug not
1141 fixed?". You'll have a few other examples of such computations. You
1142 can also read "A Network Model for Simulation of Grid Application" by
1143 Henri Casanova and Loris Marchal to have all the details. The fact
1144 that the real_bw is smaller than bw is easy to understand. The fact
1145 that real_bw is smaller than SG_TCP_CTE_GAMMA/(2*lat) is due to the
1146 window-based congestion mechanism of TCP. With TCP, you can't exploit
1147 your huge network capacity if you don't have a good round-trip-time
1148 because of the acks...
1150 Anyway, what you get is t=lat + L/min(bw,SG_TCP_CTE_GAMMA/(2*lat)).
1152 * if I you set (bw,lat)=(100 000 000, 0.00001), you get t = 1.00001 (you fully
1154 * if I you set (bw,lat)=(100 000 000, 0.0001), you get t = 1.0001 (you're on the
1156 * if I you set (bw,lat)=(100 000 000, 0.001), you get t = 10.001 (ouch!)
1158 This bound on the effective bandwidth of a flow is not the only thing
1159 that may make your result be unexpected. For example, two flows
1160 competing on a saturated link receive an amount of bandwidth inversely
1161 proportional to their round trip time.
1163 \subsection faq_bugrepport So I've found a bug in SimGrid. How to report it?
1165 We do our best to make sure to hammer away any bugs of SimGrid, but this is
1166 still an academic project so please be patient if/when you find bugs in it.
1167 If you do, the best solution is to drop an email either on the simgrid-user
1168 or the simgrid-devel mailing list and explain us about the issue. You can
1169 also decide to open a formal bug report using the
1170 <a href="https://gforge.inria.fr/tracker/?atid=165&group_id=12&func=browse">relevant
1171 interface</a>. You need to login on the server to get the ability to submit
1174 We will do our best to solve any problem repported, but you need to help us
1175 finding the issue. Just telling "it segfault" isn't enough. Telling "It
1176 segfaults when running the attached simulator" doesn't really help either.
1177 You may find the following article interesting to see how to repport
1178 informative bug repports:
1179 http://www.chiark.greenend.org.uk/~sgtatham/bugs.html (it is not SimGrid
1180 specific at all, but it's full of good advices).
1182 \author Arnaud Legrand (arnaud.legrand::imag.fr)
1183 \author Martin Quinson (martin.quinson::loria.fr)