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 Moreover, you should never call the autotools manually since you must run
111 them in a specific order with specific arguments. Most of the times, the
112 makefiles will automatically call the tools for you. When it's not possible
113 (such as the first time you checkout the CVS), use the ./bootstrap command
114 to call them explicitely.
116 \subsection faq_setting_MSG Setting up your own MSG code
118 Do not build your simulator by modifying the SimGrid examples. Go
119 outside the SimGrid source tree and create your own working directory
120 (say <tt>/home/joe/SimGrid/MyFirstScheduler/</tt>).
122 Suppose your simulation has the following structure (remember it is
123 just an example to illustrate a possible way to compile everything;
124 feel free to organize it as you want).
126 \li <tt>sched.h</tt>: a description of the core of the
127 scheduler (i.e. which functions are can be used by the
128 agents). For example we could find the following functions
129 (master, forwarder, slave).
131 \li <tt>sched.c</tt>: a C file including <tt>sched.h</tt> and
132 implementing the core of the scheduler. Most of these
133 functions use the MSG functions defined in section \ref
136 \li <tt>masterslave.c</tt>: a C file with the main function, i.e.
137 the MSG initialization (MSG_global_init()), the platform
138 creation (e.g. with MSG_create_environment()), the
139 deployment phase (e.g. with MSG_function_register() and
140 MSG_launch_application()) and the call to
143 To compile such a program, we suggest to use the following
144 Makefile. It is a generic Makefile that we have used many times with
145 our students when we teach the C language.
149 masterslave: masterslave.o sched.o
151 INSTALL_PATH = $$HOME
153 PEDANTIC_PARANOID_FREAK = -O0 -Wshadow -Wcast-align \
154 -Waggregate-return -Wmissing-prototypes -Wmissing-declarations \
155 -Wstrict-prototypes -Wmissing-prototypes -Wmissing-declarations \
156 -Wmissing-noreturn -Wredundant-decls -Wnested-externs \
157 -Wpointer-arith -Wwrite-strings -finline-functions
158 REASONABLY_CAREFUL_DUDE = -Wall
159 NO_PRAYER_FOR_THE_WICKED = -w -O2
160 WARNINGS = $(REASONABLY_CAREFUL_DUDE)
161 CFLAGS = -g $(WARNINGS)
163 INCLUDES = -I$(INSTALL_PATH)/include
164 DEFS = -L$(INSTALL_PATH)/lib/
165 LDADD = -lm -lsimgrid
169 $(CC) $(INCLUDES) $(DEFS) $(CFLAGS) $^ $(LIBS) $(LDADD) -o $@
172 $(CC) $(INCLUDES) $(DEFS) $(CFLAGS) -c -o $@ $<
175 rm -f $(BIN_FILES) *.o *~
181 The first two lines indicates what should be build when typing make
182 (<tt>masterslave</tt>) and of which files it is to be made of
183 (<tt>masterslave.o</tt> and <tt>sched.o</tt>). This makefile assumes
184 that you have set up correctly your <tt>LD_LIBRARY_PATH</tt> variable
185 (look, there is a <tt>LDADD = -lm -lsimgrid</tt>). If you prefer using
186 the static version, remove the <tt>-lsimgrid</tt> and add a
187 <tt>$(INSTALL_PATH)/lib/libsimgrid.a</tt> on the next line, right
188 after the <tt>LIBS = </tt>.
190 More generally, if you have never written a Makefile by yourself, type
191 in a terminal : <tt>info make</tt> and read the introduction. The
192 previous example should be enough for a first try but you may want to
193 perform some more complex compilations...
195 \subsection faq_setting_GRAS Setting up your own GRAS code
197 If you use the GRAS interface instead of the MSG one, then previous section
198 is not the better source of information. Instead, you should check the GRAS
199 tutorial in general, and the \ref GRAS_tut_tour_setup in particular.
201 \subsection faq_crosscompile Cross-compiling a Windows DLL of SimGrid from linux
203 At the moment, we do not distribute Windows pre-compiled version of SimGrid
204 because the support for this platform is still experimental. We know that
205 some parts of the GRAS environment do not work, and we think that the others
206 environments (MSG and SD) have good chances to work, but we didn't test
207 ourselves. This section explains how we generate the SimGrid DLL so that you
208 can build it for yourself. First of all, you need to have a version more
209 recent than 3.1 (ie, a CVS version as time of writting).
211 In order to cross-compile the package to windows from linux, you need to
212 install mingw32 (minimalist gnu win32). On Debian, you can do so by
213 installing the packages mingw32 (compiler), mingw32-binutils (linker and
214 so), mingw32-runtime.
216 You can use the VPATH support of configure to compile at the same time for
217 linux and windows without dupplicating the source nor cleaning the tree
218 between each. Just run bootstrap (if you use the CVS) to run the autotools.
219 Then, create a linux and a win directories. Then, type:
220 \verbatim cd linux; ../configure --srcdir=.. <usual configure flags>; make; cd ..
221 cd win; ../configure --srcdir=.. --host=i586-mingw32msvc <flags>; make; cd ..
223 The trick to VPATH builds is to call configure from another directory,
224 passing it an extra --srcdir argument to tell it where all the sources are.
225 It will understand you want to use VPATH. Then, the trick to cross-compile
226 is simply to add a --host argument specifying the target you want to build
227 for. The i586-mingw32msvc string is what you have to pass to use the mingw32
228 environment as distributed in Debian.
230 After that, you can run all make targets from both directories, and test
231 easily that what you change for one arch does not break the other one.
233 It is possible that this VPATH build thing breaks from time to time in the
234 CVS since it's quite fragile, but it's granted to work in any released
235 version. If you experience problems, drop us a mail.
237 Another possible source of issue is that at the moment, building the
238 examples request to use the gras_stub_generator tool, which is a compiled
239 program, not a script. In cross-compilation, you need to cross-execute with
240 wine for example, which is not really pleasant. We are working on this, but
241 in the meanwhile, simply don't build the examples in cross-compilation
242 (<tt>cd src</tt> before running make).
244 Program (cross-)compiled with mingw32 do request an extra DLL at run-time to be
245 usable. For example, if you want to test your build with wine, you should do
246 the following to put this library where wine looks for DLLs.
247 \verbatim cp /usr/share/doc/mingw32-runtime/mingwm10.dll.gz ~/.wine/c/windows/system/
248 gunzip ~/.wine/c/windows/system/mingwm10.dll.gz
251 The DLL is builded in src/.libs, and installed in the <prefix>/bin directory
252 when you run make install.
254 If you want to use it in a native project on windows, you need to use
255 simgrid.dll and mingwm10.dll. For each DLL, you need to build .def file
256 under linux (listing the defined symbols), and convert it into a .lib file
257 under windows (specifying this in a way that windows compilers like). To
258 generate the def files, run (under linux):
259 \verbatim echo "LIBRARY libsimgrid-0.dll" > simgrid.def
260 echo EXPORTS >> simgrid.def
261 nm libsimgrid-0.dll | grep ' T _' | sed 's/.* T _//' >> simgrid.def
262 nm libsimgrid-0.dll | grep ' D _' | sed 's/.* D _//' | sed 's/$/ DATA/' >> simgrid.def
264 echo "LIBRARY mingwm10.dll" > mingwm10.def
265 echo EXPORTS >> mingwm10.def
266 nm mingwm10.dll | grep ' T _' | sed 's/.* T _//' >> mingwm10.def
267 nm mingwm10.dll | grep ' D _' | sed 's/.* D _//' | sed 's/$/ DATA/' >> mingwm10.def
270 To create the import .lib files, use the <tt>lib</tt> windows tool (from
271 MSVC) the following way to produce simgrid.lib and mingwm10.lib
272 \verbatim lib /def:simgrid.def
273 lib /def:mingwm10.def
276 If you happen to use Borland C Builder, the right command line is the
277 following (note that you don't need any file.def to get this working).
278 \verbatim implib simgrid.lib libsimgrid-0.dll
279 implib mingwm10.lib mingwm10.dll
282 Then, set the following parameters in Visual C++ 2005:
283 Linker -> Input -> Additional dependencies = simgrid.lib mingwm10.lib
285 Just in case you wonder how to generate a DLL from libtool in another
286 project, we added -no-undefined to any lib*_la_LDFLAGS variables so that
287 libtool accepts to generate a dynamic library under windows. Then, to make
288 it true, we pass any dependencies (such as -lws2 under windows or -lpthread
289 on need) on the linking line. Passing such deps is a good idea anyway so
290 that they get noted in the library itself, avoiding the users to know about
291 our dependencies and put them manually on their compilation line. Then we
292 added the AC_LIBTOOL_WIN32_DLL macro just before AC_PROG_LIBTOOL in the
293 configure.ac. It means that we exported any symbols which need to be.
294 Nowadays, functions get automatically exported, so we don't need to load our
295 header files with tons of __declspec(dllexport) cruft. We only need to do so
296 for data, but there is no public data in SimGrid so we are good.
298 \section faq_simgrid I'm new to SimGrid. I have some questions. Where should I start?
300 You are at the right place... Having a look to these
301 <a href="http://graal.ens-lyon.fr/~alegrand/articles/slides_g5k_simul.pdf">slides</a>
303 <a href="http://graal.ens-lyon.fr/~alegrand/articles/Simgrid-Introduction.pdf">"obsolete" slides</a>)
304 may give you some insights on what SimGrid can help you to do and what
305 are its limitations. Then you definitely should read the \ref
306 MSG_examples. There is also a mailing list: <simgrid-user@lists.gforge.inria.fr>.
308 \subsection faq_generic Building a generic simulator
310 Please read carefully the \ref MSG_examples. You'll find in \ref
311 MSG_ex_master_slave a very simple consisting of a master (that owns a bunch of
312 tasks and distributes them) , some slaves (that process tasks whenever
313 they receive one) and some forwarder agents (that simply pass the
314 tasks they receive to some slaves).
316 \subsection faq_visualization Visualizing the schedule
318 It is sometime convenient to "see" how the agents are behaving. If you
319 like colors, you can use <tt>tools/MSG_visualization/colorize.pl </tt>
320 as a filter to your MSG outputs. It works directly with INFO. Beware,
321 INFO() prints on stderr. Do not forget to redirect if you want to
322 filter (e.g. with bash):
324 ./msg_test small_platform.xml small_deployment.xml 2>&1 | ../../tools/MSG_visualization/colorize.pl
327 We also have a more graphical output. Have a look at MSG_paje_output(). It
328 generates an input to <a href="http://www-id.imag.fr/Logiciels/paje/">Paje</a>.
331 <a href="Paje_MSG_screenshot.jpg"><img src="Paje_MSG_screenshot_thn.jpg"></a>
335 \subsection faq_postmortem_analysis Online/postmortem analysis
337 Vizualization with Paje can be seen as a kind of postmortem
338 analysis. However, as soon as you start playing with big simulations,
339 you'll realize that processing such output is kind of tricky. There is
340 so much generic informations that it is hard to find the information
343 As a matter of fact, loging really depends on simulations (e.g. what
344 kind of events is important...). That is why we do not propose a big
345 dump of your whole simulation (it would slow everything down) but give
346 you neat tools to structure you logs. Have a look at \ref XBT_log. In
347 fact, rather than a post-mortem analysis, you may want to do it on the
348 fly. The process you are running can do whatever you want. Have you
349 thought about adding a global structure where you directly compute the
350 informations that are really important rather than writing everything
351 down and then processing huge files ?
353 \subsection faq_C Argh! Do I really have to code in C ?
355 Up until now, there is no binding for other languages. If you use C++,
356 you should be able to use the SimGrid library as a standard C library
357 and everything should work fine (simply <i>link</i> against this
358 library; recompiling SimGrid with a C++ compiler won't work and it
359 wouldn't help if you could).
361 In fact, the bindings needed to allow one to use SimGrid from Perl,
362 Python, Java, etc. are double-layered. The first layer would allow
363 you to call for example the MSG_task_get_name(task) function while
364 what you really want is a proper object wrapping allowing you to call
365 task->name(). That's the purpose of the second layer. The first one
366 is granted with C++ but can be done with tools like
367 <a href="www.swig.org/">swig</a> for other languages like Perl, Ruby,
368 Python, CAML. None of us really need the second one (which is a bit
369 more demanding and cannot be automatically generated) yet and there is
370 no real point in doing the first one without the second. :)
372 As usual, you're welcome to participate.
374 \section faq_MIA How to ....? Is there a function in the API to simply ....?
376 Here is the deal. The whole SimGrid project (MSG, SURF, GRAS, ...) is
377 meant to be kept as simple and generic as possible. We cannot add
378 functions for everybody's need when these functions can easily be
379 built from the ones already in the API. Most of the time, it is
380 possible and when it was not possible we always have upgraded the API
381 accordingly. When somebody asks us a question like "How to do that ?
382 Is there a function in the API to simply do this ?", we're always glad
383 to answer and help. However if we don't need this code for our own
384 need, there is no chance we're going to write it... it's your job! :)
385 The counterpart to our answers is that once you come up with a neat
386 implementation of this feature (task duplication, RPC, thread
387 synchronization, ...), you should send it to us and we will be glad to
388 add it to the distribution. Thus, other people will take advantage of
389 it (and we don't have to answer this question again and again ;).
391 You'll find in this section a few "Missing In Action" features. Many
392 people have asked about it and we have given hints on how to simply do
393 it with MSG. Feel free to contribute...
395 \subsection faq_MIA_examples I want some more complex examples!
397 Many people have come to ask me a more complex example and each time,
398 they have realized afterward that the basics were in the previous three
401 Of course they have often been needing more complex functions like
402 MSG_process_suspend(), MSG_process_resume() and
403 MSG_process_isSuspended() (to perform synchronization), or
404 MSG_task_Iprobe() and MSG_process_sleep() (to avoid blocking
405 receptions), or even MSG_process_create() (to design asynchronous
406 communications or computations). But the examples are sufficient to
409 We know. We should add some more examples, but not really some more
410 complex ones... We should add some examples that illustrate some other
411 functionalities (like how to simply encode asynchronous
412 communications, RPC, process migrations, thread synchronization, ...)
413 and we will do it when we will have a little bit more time. We have
414 tried to document the examples so that they are understandable. Tell
415 us if something is not clear and once again feel free to participate!
418 \subsection faq_MIA_taskdup Missing in action: Task duplication/replication
420 There is no task duplication in MSG. When you create a task, you can
421 process it or send it somewhere else. As soon as a process has sent
422 this task, he doesn't have this task anymore. It's gone. The receiver
423 process has got the task. However, you could decide upon receiving to
424 create a "copy" of a task but you have to handle by yourself the
425 semantic associated to this "duplication".
427 As we already told, we prefer keeping the API as simple as
428 possible. This kind of feature is rather easy to implement by users
429 and the semantic you associate really depends on people. Having a
430 *generic* task duplication mechanism is not that trivial (in
431 particular because of the data field). That is why I would recommand
432 that you write it by yourself even if I can give you advice on how to
435 You have the following functions to get informations about a task:
436 MSG_task_get_name(), MSG_task_get_compute_duration(),
437 MSG_task_get_remaining_computation(), MSG_task_get_data_size(),
438 and MSG_task_get_data().
440 You could use a dictionnary (#xbt_dict_t) of dynars (#xbt_dynar_t). If
441 you still don't see how to do it, please come back to us...
443 \subsection faq_MIA_asynchronous I want to do asynchronous communications in MSG
445 Up until now, there is no asynchronous communications in MSG. However,
446 you can create as many process as you want so you should be able to do
447 whatever you want... I've written a queue module to help implementing
448 some asynchronous communications at low cost (creating thousands of
449 process only to handle communications may be problematic in term of
450 performance at some point). I'll add it in the distribution asap.
452 \subsection faq_MIA_thread_synchronization I need to synchronize my MSG processes
454 You obviously cannot use pthread_mutexes of pthread_conds. The best
455 thing would be to propose similar structures. Unfortunately, we
456 haven't found time to do it yet. However you can try to play with
457 MSG_process_suspend() and MSG_process_resume(). You can even do some
458 synchronization with fake communications (using MSG_task_get(),
459 MSG_task_put() and MSG_task_Iprobe()).
461 \subsection faq_MIA_host_load Where is the get_host_load function hidden in MSG?
463 There is no such thing because its semantic wouldn't be really
464 clear. Of course, it is something about the amount of host throughput,
465 but there is as many definition of "host load" as people asking for
466 this function. First, you have to remember that resource availability
467 may vary over time, which make any load notion harder to define.
469 It may be instantaneous value or an average one. Moreover it may be only the
470 power of the computer, or may take the background load into account, or may
471 even take the currently running tasks into account. In some SURF models,
472 communications have an influence on computational power. Should it be taken
475 So, we decided not to include such a function into MSG and let people do it
476 thereselves so that they get the value matching exactly what they mean. One
477 possibility is to run active measurement as in next code snippet. It is very
478 close from what you would have to do out of the simulator, and thus gives
479 you information that you could also get in real settings to not hinder the
480 realism of your simulation.
483 double get_host_load() {
484 m_task_t task = MSG_task_create("test", 0.001, 0, NULL);
485 double date = MSG_get_clock();
487 MSG_task_execute(task);
488 date = MSG_get_clock() - date;
489 MSG_task_destroy(task);
494 Of course, it may not match your personal definition of "host load". In this
495 case, please detail what you mean on the mailing list, and we will extend
496 this FAQ section to fit your taste if possible.
498 \subsection faq_MIA_communication_time How can I get the *real* communication time ?
500 Communications are synchronous and thus if you simply get the time
501 before and after a communication, you'll only get the transmission
502 time and the time spent to really communicate (it will also take into
503 account the time spent waiting for the other party to be
504 ready). However, getting the *real* communication time is not really
505 hard either. The following solution is a good starting point.
510 m_task_t task = MSG_task_create("Task", task_comp_size, task_comm_size,
511 calloc(1,sizeof(double)));
512 *((double*) task->data) = MSG_get_clock();
513 MSG_task_put(task, slaves[i % slaves_count], PORT_22);
514 INFO0("Send completed");
519 m_task_t task = NULL;
522 time1 = MSG_get_clock();
523 a = MSG_task_get(&(task), PORT_22);
524 time2 = MSG_get_clock();
525 if(time1<*((double *)task->data))
526 time1 = *((double *) task->data);
527 INFO1("Communication time : \"%f\" ", time2-time1);
529 MSG_task_destroy(task);
534 \subsection faq_MIA_batch_scheduler Is there a native support for batch schedulers in SimGrid ?
536 No, there is no native support for batch schedulers and none is
537 planned because this is a very specific need (and doing it in a
538 generic way is thus very hard). However some people have implemented
539 their own batch schedulers. Vincent Garonne wrote one during his PhD
540 and put his code in the contrib directory of our CVS so that other can
541 keep working on it. You may find inspinring ideas in it.
543 \subsection faq_MIA_checkpointing I need a checkpointing thing
545 Actually, it depends on whether you want to checkpoint the simulation, or to
546 simulate checkpoints.
548 The first one could help if your simulation is a long standing process you
549 want to keep running even on hardware issues. It could also help to
550 <i>rewind</i> the simulation by jumping sometimes on an old checkpoint to
551 cancel recent calculations.\n
552 Unfortunately, such thing will probably never exist in SG. One would have to
553 duplicate all data structures because doing a rewind at the simulator level
554 is very very hard (not talking about the malloc free operations that might
555 have been done in between). Instead, you may be interested in the Libckpt
556 library (http://www.cs.utk.edu/~plank/plank/www/libckpt.html). This is the
557 checkpointing solution used in the condor project, for example. It makes it
558 easy to create checkpoints (at the OS level, creating something like core
559 files), and rerunning them on need.
561 If you want to simulate checkpoints instead, it means that you want the
562 state of an executing task (in particular, the progress made towards
563 completion) to be saved somewhere. So if a host (and the task executing on
564 it) fails (cf. #MSG_HOST_FAILURE), then the task can be restarted
565 from the last checkpoint.\n
567 Actually, such a thing does not exists in SimGrid either, but it's just
568 because we don't think it is fundamental and it may be done in the user code
569 at relatively low cost. You could for example use a watcher that
570 periodically get the remaining amount of things to do (using
571 MSG_task_get_remaining_computation()), or fragment the task in smaller
574 \section faq_SG Where has SG disappeared?!?
576 OK, it's time to explain what's happening to the SimGrid project. Let's
577 start with a little bit of history.
579 * Historically, SimGrid was a low-level toolkit for scheduling with
580 classical models such as DAGs. That was SimGrid v.1.* aka SG, written
581 by Henri Casanova. I (Arnaud) had been using it in its earliest
582 versions during an internship at UCSD.
584 Then we have realized that encoding distributed algorithm in SG was a
587 * So we have built MSG on top of SG and have released SimGrid v.2.*. MSG
588 offered a very basic API to encode a distributed application easily.
589 However encoding MSG on top of SG was not really convenient and did not
590 use the DAG part since the control of the task synchronization was done
591 on top of MSG and no more in SG. We have been playing a little bit with
592 MSG. We have realized that:
594 \li 1) the platform modeling was quite flexible and could be "almost"
595 automated (e.g. using random generator and post-annotations);
597 \li 2) SG was the bottleneck because of the way we were using
598 it. We needed to simulate concurrent transfers, complex load
599 sharing mechanisms. Many optimizations (e.g. trace integration)
600 were totally inefficient when combined with MSG and made extending SG
601 to implement new sharing policies, parallel tasks models, or failures
602 (many people were asking for these kind of features) a real pain;
604 \li 3) the application modeling was not really easy. Even though the
605 application modeling depends on people's applications, we thought
606 we could improve things here. One of our target here was realistic
607 distributed applications ranging from computer sensor networks like
608 the NWS to peer-to-peer applications;
610 * So we have been planning mainly two things for SimGrid 3:
612 \li 1) I have proposed to get rid of SG and to re-implement a new kernel
613 that would be faster and more flexible. That is what I did in the
614 end of 2004: SURF. SURF is based on a fast max-min linear solver
615 using O(1) data-structures. I have quickly replaced SG by SURF in
616 MSG and the result has been that on the MSG example, the new
617 version was more than 10 times faster while we had gain a lot of
618 flexibility. I think I could still easily make MSG faster but I
619 have to work on MSG now (e.g. using some of the O(1)
620 data-structures I've been using to build SURF) since it has become
621 the bottleneck. Some MSG functions have been removed from the API
622 but they were mainly intended to build the platform by hand (they
623 had appeared in the earliest versions of MSG) and were therefore
624 not useful anymore since we are providing a complete mechanism to
625 automatically build the platform and deploy the agents on it.;
627 \li 2) GRAS is a new project Martin and I have come up with. The idea is
628 to have a programming environment that let you program real
629 distributed applications while letting you the ability to run it in
630 the simulator without having to change the slightest line of your
631 code. From the simulation point of view, GRAS performs the
632 application modeling automatically... Up until now, GRAS works on
633 top MSG for historical reasons but I'm going to make it work
634 directly on top of SURF so that it can use all the flex and the
635 speed provided by SURF.
637 Those two things are working, but we want to make everything as clean as
638 possible before releasing SimGrid v.3.
640 So what about those nice DAGs we used to have in SimGrid v.1.? They're
641 not anymore in SimGrid v.3. At least not in their original form... Let
642 me recall you the way SimGrid 3 is organized:
656 XBT is our tool box and now, you should have an idea of what the other
657 ones are. As you can see, the primitive SG is not here
658 anymore. However we have written a brand new and cleaner API for this
659 purpose: \ref SD_API. It is built directly on top of SURF and provides
660 an API rather close to the old SG:
663 ______________________
665 |____________________|
666 | | MSG | GRAS | SD |
667 | -------------------|
669 | -------------------|
671 ----------------------
674 The nice thing is that, as it is writen on top of SURF, it seamlessly
675 support DAG of parallel tasks as well as complex communications
676 patterns. Some old codes using SG are currently under rewrite using
677 \ref SD_API to check that all needful functions are provided.
679 \subsection faq_SG_DAG How to implement a distributed dynamic scheduler of DAGs.
681 Distributed is somehow "contagious". If you start making distributed
682 decisions, there is no way to handle DAGs directly anymore (unless I
683 am missing something). You have to encode your DAGs in term of
684 communicating process to make the whole scheduling process
685 distributed. Here is an example of how you could do that. Assume T1
686 has to be done before T2.
689 int your_agent(int argc, char *argv[] {
691 T1 = MSG_task_create(...);
692 T2 = MSG_task_create(...);
696 if(cond) MSG_task_execute(T1);
698 if((MSG_task_get_remaining_computation(T1)=0.0) && (you_re_in_a_good_mood))
701 /* do something else */
707 If you decide that the distributed part is not that much important and that
708 DAG is really the level of abstraction you want to work with, then you should
709 give a try to \ref SD_API.
711 \section faq_dynamic Dynamic resources and platform building
713 \subsection faq_platform Building a realistic platform
715 We can speak more than an hour on this subject and we still do not have
716 the right answer, just some ideas. You can read the following
717 <a href="http://graal.ens-lyon.fr/~alegrand/articles/Simgrid-Introduction.pdf">slides</a>.
718 It may give you some hints. You can also have a look at the
719 <tt>tools/platform_generation/</tt> directory. There is a perl-script
720 we use to annotate a Tiers generated platform.
722 \subsection faq_SURF_dynamic How can I have variable resource availability?
724 A nice feature of SimGrid is that it enables you to seamlessly have
725 resources whose availability change over time. When you build a
726 platform, you generally declare CPUs like that:
729 <cpu name="Cpu A" power="100.00"/>
732 If you want the availability of "CPU A" to change over time, the only
733 thing you have to do is change this definition like that:
736 <cpu name="Cpu A" power="100.00" availability_file="trace_A.txt" state_file="trace_A_failure.txt"/>
739 For CPUs, availability files are expressed in fraction of available
740 power. Let's have a look at what "trace_A.txt" may look like:
749 At time 0, our CPU will deliver 100 Mflop/s. At time 11.0, it will
750 deliver only 50 Mflop/s until time 20.0 where it will will start
751 delivering 90 Mflop/s. Last at time 21.0 (20.0 plus the periodicity
752 1.0), we'll be back to the beginning and it will deliver 100Mflop/s.
754 Now let's look at the state file:
761 A negative value means "off" while a positive one means "on". At time
762 1.0, the CPU is on. At time 1.0, it is turned off and at time 2.0, it
763 is turned on again until time 12 (2.0 plus the periodicity 10.0). It
764 will be turned on again at time 13.0 until time 23.0, and so on.
766 Now, let's look how the same kind of thing can be done for network
767 links. A usual declaration looks like:
770 <network_link name="LinkA" bandwidth="10.0" latency="0.2"/>
773 You have at your disposal the following options: bandwidth_file,
774 latency_file and state_file. The only difference with CPUs is that
775 bandwidth_file and latency_file do not express fraction of available
776 power but are expressed directly in Mb/s and seconds.
778 \subsection faq_flexml_bypassing How can I have some C functions do what the platform file does?
780 So you want to bypass the XML files parser, uh? Maybe doin some parameter
781 sweep experiments on your simulations or so? This is possible, but it's not
782 really easy. Here is how it goes.
784 For this, you have to first remember that the XML parsing in SimGrid is done
785 using a tool called FleXML. Given a DTD, this gives a flex-based parser. If
786 you want to bypass the parser, you need to provide some code mimicking what
787 it does and replacing it in its interactions with the SURF code. So, let's
788 have a look at these interactions.
790 FleXML parser are close to classical SAX parsers. It means that a
791 well-formed SimGrid platform XML file might result in the following
794 - start "platform_description"
795 - start "cpu" with attributes name="host1" power="1.0"
797 - start "cpu" with attributes name="host2" power="2.0"
799 - start "network_link" with ...
801 - start "route" with ...
803 - start "route" with ...
805 - end "platform_description"
807 The communication from the parser to the SURF code uses two means:
808 Attributes get copied into some global variables, and a surf-provided
809 function gets called by the parser for each event. For example, the event
810 - start "cpu" with attributes name="host1" power="1.0"
812 let the parser do the equivalent of:
814 strcpy("host1",A_cpu_name);
819 In SURF, we attach callbacks to the different events by initializing the
820 pointer functions to some the right surf functions. Example in
821 workstation_KCCFLN05.c (surf_parse_open() ends up calling surf_parse()):
823 // Building the routes
824 surf_parse_reset_parser();
825 STag_route_fun=parse_route_set_endpoints;
826 ETag_route_element_fun=parse_route_elem;
827 ETag_route_fun=parse_route_set_route;
828 surf_parse_open(file);
829 xbt_assert1((!surf_parse()),"Parse error in %s",file);
833 So, to bypass the FleXML parser, you need to write your own version of the
834 surf_parse function, which should do the following:
835 - Call the corresponding STag_<tag>_fun function to simulate tag start
836 - Fill the A_<tag>_<attribute> variables with the wanted values
837 - Call the corresponding ETag_<tag>_fun function to simulate tag end
838 - (do the same for the next set of values, and loop)
840 Then, tell SimGrid that you want to use your own "parser" instead of the stock one:
842 surf_parse = surf_parse_bypass;
843 MSG_create_environment(NULL);
846 An example of this trick is distributed in the file examples/msg/msg_test_surfxml_bypassed.c
848 \section faq_limits Pushing the limits
850 \subsection faq_context_1000 I want thousands of simulated processes
852 SimGrid can use either pthreads library or the UNIX98 contextes. On most
853 systems, the number of pthreads is limited and then your simulation may be
854 limited for a stupid reason. This is especially true with the current linux
855 pthreads, and I cannot get more than 2000 simulated processes with pthreads
856 on my box. The UNIX98 contexts allow me to raise the limit to 25,000
857 simulated processes on my laptop.
859 The <tt>--with-context</tt> option of the <tt>./configure</tt> script allows
860 you to choose between UNIX98 contextes (<tt>--with-context=ucontext</tt>)
861 and the pthread version ( (<tt>--with-context=pthread</tt>). The default
862 value is ucontext when the script detect a working UNIX98 context
863 implementation. On Windows boxes, the provided value is discarded and an
864 adapted version is picked up.
866 We experienced some issues with contextes on some rare systems (solaris 8
867 and lower or old alpha linuxes comes to mind). The main problem is that the
868 configure script detect the contextes as being functional when it's not
869 true. If you happen to use such a system, switch manually to the pthread
870 version, and provide us with a good patch for the configure script so that
871 it is done automatically ;)
873 \subsection faq_context_10000 I want hundred thousands of simulated processes
875 As explained above, SimGrid can use UNIX98 contextes to represent and handle
876 the simulated processes. Thanks to this, the main limitation to the number
877 of simulated processes becomes the available memory.
879 Here are some tricks I had to use in order to run a token ring between
880 25,000 processes on my laptop (1Gb memory, 1.5Gb swap).
882 - First of all, make sure your code runs for a few hundreds processes
883 before trying to push the limit. Make sure it's valgrind-clean, ie that
884 valgrind does not report neither memory error nor memory leaks. Indeed,
885 numerous simulated processes result in *fat* simulation hindering debugging.
887 - It was really boring to write 25,000 entries in the deployment file, so I wrote
888 a little script <tt>examples/gras/tokenS/make_deployment.pl</tt>, which you may
889 want to adapt to your case. You could also think about hijacking
890 the SURFXML parser (have look at \ref faq_flexml_bypassing).
892 - The deployment file became quite big, so I had to do what is in the FAQ
893 entry \ref faq_flexml_limit
895 - Each UNIX98 context has its own stack entry. As debugging this is quite
896 hairly, the default value is a bit overestimated so that user don't get
897 into trouble about this. You want to tune this size to increse the number
898 of processes. This is the <tt>STACK_SIZE</tt> define in
899 <tt>src/xbt/context_private.h</tt>, which is 128kb by default.
900 Reduce this as much as you can, but be warned that if this value is too
901 low, you'll get a segfault. The token ring example, which is quite simple,
902 runs with 40kb stacks.
904 \section faq_troubleshooting Troubleshooting
906 \subsection faq_compil_trouble ./configure fails!
908 We now only one reason for the configure to fail:
910 - <b>You are using a borken build environment</b>\n
911 If symptom is that configure complains about gcc not being able to build
912 executables, you are probably missing the libc6-dev package. Damn Ubuntu.
914 If you experience other kind of issue, please get in touch with us. We are
915 always interested in improving our portability to new systems.
917 \subsection faq_distcheck_fails Dude! "make check" fails on my machine!
919 Don't assume we never run this target, because we do. Really. Promise!
921 There is several reasons which may cause the make check to fail on your
924 - <b>You are using a borken libc (probably concerning the contextes)</b>.\n
925 The symptom is that the "make check" fails within the examples/msg directory.\n
926 By default, SimGrid uses something called ucontexts. This is part of the
927 libc, but it's quite undertested. For example, some (old) versions of the
928 glibc on alpha do not implement these functions, but provide the stubs
929 (which return ENOSYS: not implemented). It fools our detection mecanism
930 and leads to segfaults.\n
931 On some x86_64, the pointer to function is stored into a integer, but int
932 are 32bits only on this arch while pointers are 64bits. Our detection
933 mecanism also fails to detect the problem, which leads to segfaults.\n
934 In both cases, there is not much we can do to fix the bug. We are working
935 on a workaround for x86_64 machines, but in the meanwhile, you can
936 compile with --with-context=pthread to avoid ucontext completely. You'll
937 be a bit more limitated in the number of simulated processes you can start
938 concurently, but 5000 processes is still enough for most purposes, isn't
940 This limitation is the reason why we insist on using this piece of ...
941 software even if it's so troublesome.\n
942 <b>=> use --with-pthread on AMD64 architecture that do not have an
943 ultra-recent libc.</b>
945 - <b>There is a bug in SimGrid we aren't aware of</b>.\n
946 If none of the above apply, please drop us a mail on the mailing list so
947 that we can check it out.
949 \subsection faq_longjmp longjmp madness in valgrind
951 This is when valgrind starts complaining about longjmp things, just like:
953 \verbatim ==21434== Conditional jump or move depends on uninitialised value(s)
954 ==21434== at 0x420DBE5: longjmp (longjmp.c:33)
956 ==21434== Use of uninitialised value of size 4
957 ==21434== at 0x420DC3A: __longjmp (__longjmp.S:48)
960 or even when it reports scary things like:
962 \verbatim ==24023== Warning: client switching stacks? SP change: 0xBE3FF618 --> 0xBE7FF710
963 x86->IR: unhandled instruction bytes: 0xF4 0xC7 0x83 0xD0
964 ==24023== to suppress, use: --max-stackframe=4194552 or greater
965 ==24023== Your program just tried to execute an instruction that Valgrind
966 ==24023== did not recognise. There are two possible reasons for this.
967 ==24023== 1. Your program has a bug and erroneously jumped to a non-code
968 ==24023== location. If you are running Memcheck and you just saw a
969 ==24023== warning about a bad jump, it's probably your program's fault.
970 ==24023== 2. The instruction is legitimate but Valgrind doesn't handle it,
971 ==24023== i.e. it's Valgrind's fault. If you think this is the case or
972 ==24023== you are not sure, please let us know.
973 ==24023== Either way, Valgrind will now raise a SIGILL signal which will
974 ==24023== probably kill your program.
976 ==24023== Process terminating with default action of signal 4 (SIGILL)
977 ==24023== Illegal opcode at address 0x420D234
978 ==24023== at 0x420D234: abort (abort.c:124)
981 This is the sign that you didn't used the exception mecanism well. Most
982 probably, you have a <tt>return;</tt> somewhere within a <tt>TRY{}</tt>
983 block. This is <b>evil</b>, and you must not do this. Did you read the section
986 \subsection faq_valgrind Valgrind spits tons of errors!
988 It may happen that valgrind, the memory debugger beloved by any decent C
989 programmer, spits tons of warnings like the following :
990 \verbatim ==8414== Conditional jump or move depends on uninitialised value(s)
991 ==8414== at 0x400882D: (within /lib/ld-2.3.6.so)
992 ==8414== by 0x414EDE9: (within /lib/tls/i686/cmov/libc-2.3.6.so)
993 ==8414== by 0x400B105: (within /lib/ld-2.3.6.so)
994 ==8414== by 0x414F937: _dl_open (in /lib/tls/i686/cmov/libc-2.3.6.so)
995 ==8414== by 0x4150F4C: (within /lib/tls/i686/cmov/libc-2.3.6.so)
996 ==8414== by 0x400B105: (within /lib/ld-2.3.6.so)
997 ==8414== by 0x415102D: __libc_dlopen_mode (in /lib/tls/i686/cmov/libc-2.3.6.so)
998 ==8414== by 0x412D6B9: backtrace (in /lib/tls/i686/cmov/libc-2.3.6.so)
999 ==8414== by 0x8076446: xbt_dictelm_get_ext (dict_elm.c:714)
1000 ==8414== by 0x80764C1: xbt_dictelm_get (dict_elm.c:732)
1001 ==8414== by 0x8079010: xbt_cfg_register (config.c:208)
1002 ==8414== by 0x806821B: MSG_config (msg_config.c:42)
1005 This problem is somewhere in the libc when using the backtraces and there is
1006 very few things we can do ourselves to fix it. Instead, here is how to tell
1007 valgrind to ignore the error. Add the following to your ~/.valgrind.supp (or
1008 create this file on need). Make sure to change the obj line according to
1009 your personnal mileage (change 2.3.6 to the actual version you are using,
1010 which you can retrieve with a simple "ls /lib/ld*.so").
1013 name: Backtrace madness
1015 obj:/lib/ld-2.3.6.so
1020 fun:__libc_dlopen_mode
1023 Then, you have to specify valgrind to use this suppression file by passing
1024 the <tt>--suppressions=$HOME/.valgrind.supp</tt> option on the command line.
1025 You can also add the following to your ~/.bashrc so that it gets passed
1026 automatically. Actually, it passes a bit more options to valgrind, and this
1027 happen to be my personnal settings. Check the valgrind documentation for
1030 \verbatim export VALGRIND_OPTS="--leak-check=yes --leak-resolution=high --num-callers=40 --tool=memcheck --suppressions=$HOME/.valgrind.supp" \endverbatim
1032 \subsection faq_flexml_limit I get the message "surf_parse_lex: Assertion `next<limit' failed."
1034 This is because your platform file is too big for the parser.
1036 Actually, the message comes directly from FleXML, the technology on top of
1037 which the parser is built. FleXML has the bad idea of fetching the whole
1038 document in memory before parsing it. And moreover, the memory buffer size
1039 must be determinded at compilation time.
1041 We use a value which seems big enough for our need without bloating the
1042 simulators footprints. But of course your mileage may vary. In this case,
1043 just edit src/surf/surfxml.l modify the definition of
1044 FLEXML_BUFFERSTACKSIZE. E.g.
1047 #define FLEXML_BUFFERSTACKSIZE 1000000000
1050 Then recompile and everything should be fine, provided that your version of
1051 Flex is recent enough (>= 2.5.31). If not the compilation process should
1054 A while ago, we worked on FleXML to reduce a bit its memory consumtion, but
1055 these issues remain. There is two things we should do:
1057 - use a dynamic buffer instead of a static one so that the only limit
1058 becomes your memory, not a stupid constant fixed at compilation time
1059 (maybe not so difficult).
1060 - change the parser so that it does not need to get the whole file in
1061 memory before parsing
1062 (seems quite difficult, but I'm a complete newbe wrt flex stuff).
1064 These are changes to FleXML itself, not SimGrid. But since we kinda hijacked
1065 the development of FleXML, I can grant you that any patches would be really
1066 welcome and quickly integrated.
1068 \subsection faq_gras_transport GRAS spits networking error messages
1070 Gras, on real platforms, naturally use regular sockets to communicate. They
1071 are deeply hiden in the gras abstraction, but when things go wrong, you may
1072 get some weird error messages. Here are some example, with the probable
1075 - <b>Transport endpoint is not connected</b>: several processes try to open
1076 a server socket on the same port number of the same machine. This is
1077 naturally bad and each process should pick its own port number for this.\n
1078 Maybe, you just have some processes remaining from a previous experiment
1080 Killing them may help, but again if you kill -KILL them, you'll have to
1081 wait for a while: they didn't close there sockets properly and the system
1082 needs a while to notice that this port is free again.
1084 - <b>Socket closed by remote side</b>: if the remote process is not
1085 supposed to close the socket at this point, it may be dead.
1087 - <b>Connection reset by peer</b>: I found this on internet about this
1088 error. I think it's what's happening here, too:\n
1089 <i>This basically means that a network error occurred while the client was
1090 receiving data from the server. But what is really happening is that the
1091 server actually accepts the connection, processes the request, and sends
1092 a reply to the client. However, when the server closes the socket, the
1093 client believes that the connection has been terminated abnormally
1094 because the socket implementation sends a TCP reset segment telling the
1095 client to throw away the data and report an error.\n
1096 Sometimes, this problem is caused by not properly closing the
1097 input/output streams and the socket connection. Make sure you close the
1098 input/output streams and socket connection properly. If everything is
1099 closed properly, however, and the problem persists, you can work around
1100 it by adding a one-second sleep before closing the streams and the
1101 socket. This technique, however, is not reliable and may not work on all
1103 Since GRAS sockets are closed properly (repeat after me: there is no bug
1104 in GRAS), it is either that you are closing your sockets on server side
1105 before the client get a chance to read them (use gras_os_sleep() to delay
1106 the server), or the server died awfully before the client got the data.
1108 \subsection faq_deadlock There is a deadlock !!!
1110 Unfortunately, we cannot debug every code written in SimGrid. We
1111 furthermore believe that the framework provides ways enough
1112 information to debug such informations yourself. If the textual output
1113 is not enough, Make sure to check the \ref faq_visualization FAQ entry to see
1114 how to get a graphical one.
1116 Now, if you come up with a really simple example that deadlocks and
1117 you're absolutely convinced that it should not, you can ask on the
1118 list. Just be aware that you'll be severely punished if the mistake is
1119 on your side... We have plenty of FAQ entries to redact and new
1120 features to implement for the impenitents! ;)
1122 \subsection faq_big_fat_warning A BIG FAT WARNING is reported telling me that my platform and deployment files are too old.
1124 We have decided to change the units in SimGrid. Now we use Bytes, Flops and
1125 seconds instead of MBytes, MFlops and seconds... Units should be updated
1126 accordingly and the version of platform_description should be set to a
1127 valuer greater than 1:
1129 <platform_description version="1">
1131 You should try to use the surfxml_update.pl script that can be found
1132 <a href="http://gforge.inria.fr/plugins/scmcvs/cvsweb.php/contrib/platform_generation/?cvsroot=cvsroot%2Fsimgrid">here</a>.
1134 \subsection faq_surf_network_latency I get weird timings when I play with the latencies.
1136 OK, first of all, remember that units should be Bytes, Flops and
1137 Seconds. If you don't use such units, some SimGrid constants (e.g. the
1138 SG_TCP_CTE_GAMMA constant used in most network models) won't have the
1139 right unit and you'll end up with weird results.
1141 Here is what happens with a single transfer of size L on a link
1142 (bw,lat) when nothing else happens.
1145 0-----lat--------------------------------------------------t
1146 |-----|**** real_bw =min(bw,SG_TCP_CTE_GAMMA/(2*lat)) *****|
1149 In more complex situations, this min is the solution of a complex
1150 max-min linear system. Have a look
1151 <a href="http://lists.gforge.inria.fr/pipermail/simgrid-devel/2006-April/thread.html">here</a>
1152 and read the two threads "Bug in SURF?" and "Surf bug not
1153 fixed?". You'll have a few other examples of such computations. You
1154 can also read "A Network Model for Simulation of Grid Application" by
1155 Henri Casanova and Loris Marchal to have all the details. The fact
1156 that the real_bw is smaller than bw is easy to understand. The fact
1157 that real_bw is smaller than SG_TCP_CTE_GAMMA/(2*lat) is due to the
1158 window-based congestion mechanism of TCP. With TCP, you can't exploit
1159 your huge network capacity if you don't have a good round-trip-time
1160 because of the acks...
1162 Anyway, what you get is t=lat + L/min(bw,SG_TCP_CTE_GAMMA/(2*lat)).
1164 * if I you set (bw,lat)=(100 000 000, 0.00001), you get t = 1.00001 (you fully
1166 * if I you set (bw,lat)=(100 000 000, 0.0001), you get t = 1.0001 (you're on the
1168 * if I you set (bw,lat)=(100 000 000, 0.001), you get t = 10.001 (ouch!)
1170 This bound on the effective bandwidth of a flow is not the only thing
1171 that may make your result be unexpected. For example, two flows
1172 competing on a saturated link receive an amount of bandwidth inversely
1173 proportional to their round trip time.
1175 \subsection faq_bugrepport So I've found a bug in SimGrid. How to report it?
1177 We do our best to make sure to hammer away any bugs of SimGrid, but this is
1178 still an academic project so please be patient if/when you find bugs in it.
1179 If you do, the best solution is to drop an email either on the simgrid-user
1180 or the simgrid-devel mailing list and explain us about the issue. You can
1181 also decide to open a formal bug report using the
1182 <a href="https://gforge.inria.fr/tracker/?atid=165&group_id=12&func=browse">relevant
1183 interface</a>. You need to login on the server to get the ability to submit
1186 We will do our best to solve any problem repported, but you need to help us
1187 finding the issue. Just telling "it segfault" isn't enough. Telling "It
1188 segfaults when running the attached simulator" doesn't really help either.
1189 You may find the following article interesting to see how to repport
1190 informative bug repports:
1191 http://www.chiark.greenend.org.uk/~sgtatham/bugs.html (it is not SimGrid
1192 specific at all, but it's full of good advices).
1194 \author Arnaud Legrand (arnaud.legrand::imag.fr)
1195 \author Martin Quinson (martin.quinson::loria.fr)