1 /*! \page faq Frequently Asked Questions
3 \htmlinclude .FAQ.doc.toc
5 \section faq_simgrid I'm new to SimGrid. I have some questions. Where should I start?
7 You are at the right place... Having a look to these
8 <a href="http://www.loria.fr/~quinson/articles/simgrid-tutorial.pdf">the tutorial slides</a>
9 (or to these <a href="http://graal.ens-lyon.fr/~alegrand/articles/slides_g5k_simul.pdf">old slides</a>,
11 <a href="http://graal.ens-lyon.fr/~alegrand/articles/Simgrid-Introduction.pdf">"obsolete" slides</a>)
12 may give you some insights on what SimGrid can help you to do and what
13 are its limitations. Then you definitely should read the \ref
14 MSG_examples. The \ref GRAS_tut can also help you.
16 If you are stuck at any point and if this FAQ cannot help you, please drop us a
17 mail to the user mailing list: <simgrid-user@lists.gforge.inria.fr>.
19 \subsection faq_interfaces What is the difference between MSG, SimDag, and GRAS? Do they serve the same purpose?
21 It depend on how you define "purpose", I guess ;)
23 They all allow you to build a prototype of application which you can run
24 within the simulator afterward. They all share the same simulation kernel,
25 which is the core of the SimGrid project. They differ by the way you express
28 With SimDag, you express your code as a collection of interdependent
29 parallel tasks. So, in this model, applications can be seen as a DAG of
30 tasks. This is the interface of choice for people wanting to port old
31 code designed for SimGrid v1 or v2 to the framework current version.
33 With both GRAS and MSG, your application is seen as a set of communicating
34 processes, exchanging data by the way of messages and performing computation
37 The difference between both is that MSG is somehow easier to use, but GRAS
38 is not limited to the simulator. Once you're done writing your GRAS code,
39 you can run your code both in the simulator or on a real platform. For this,
40 there is two implementations of the GRAS interface, one for simulation, one
41 for real execution. So, you just have to relink your code to chose one of
44 \subsection faq_generic First steps with SimGrid
46 If you decide to go for the MSG interface, please read carefully the
47 \ref MSG_examples. You'll find in \ref MSG_ex_master_slave a very
48 simple consisting of a master (that owns a bunch of tasks and
49 distributes them) , some slaves (that process tasks whenever they
50 receive one) and some forwarder agents (that simply pass the tasks
51 they receive to some slaves).
53 If you decide to go for the GRAS interface, you should definitively
54 read the \ref GRAS_tut. The first section constitutes an introduction
55 to the tool and presents the model we use. The second section
56 constitutes a complete step-by-step tutorial building a distributed
57 application from the beginning and exemplifying most of the GRAS
58 features in the process. The last section groups some HOWTOS
59 highlighting a given feature of the framework in a more concise way.
61 If you decide to go for another interface, I'm afraid your only sources
62 of information will be the source code and the mailing lists...
64 \subsection faq_visualization Visualizing and analyzing the results
66 It is sometime convenient to "see" how the agents are behaving. If you
67 like colors, you can use <tt>tools/MSG_visualization/colorize.pl </tt>
68 as a filter to your MSG outputs. It works directly with INFO. Beware,
69 INFO() prints on stderr. Do not forget to redirect if you want to
70 filter (e.g. with bash):
72 ./msg_test small_platform.xml small_deployment.xml 2>&1 | ../../tools/MSG_visualization/colorize.pl
75 We also have a more graphical output. Have a look at section \ref faq_tracing.
77 \subsection faq_C Argh! Do I really have to code in C?
79 Up until now, there is no binding for other languages. If you use C++,
80 you should be able to use the SimGrid library as a standard C library
81 and everything should work fine (simply <i>link</i> against this
82 library; recompiling SimGrid with a C++ compiler won't work and it
83 wouldn't help if you could).
85 In fact, we are currently working on Java bindings of MSG to allow
86 all the undergrad students of the world to use this tool. This is a
87 little more tricky than I would have expected, but the work is moving
88 fast forward [2006/05/13]. More languages are evaluated, but for now,
89 we do not feel a real demand for any other language. Please speak up!
91 \section faq_cmake Installing the SimGrid library with Cmake (since V3.4)
93 \subsection faq_intro Some generalitty
95 \subsubsection faq_intro1 What is Cmake?
97 CMake is a family of tools designed to build, test and package software. CMake is used to control the software compilation process using simple platform and compiler independent configuration files. CMake generates native makefiles and workspaces that can be used in the compiler environment of your choice. For more information see official web site <a href="http://www.cmake.org/">here</a>.
99 \subsubsection faq_intro2 Why cmake?
101 CMake permits to developers to compil projects on different plateforms. Then many tools are embedded like ctest for making test, a link to cdash for vizualise results but also test coverage and bug reports.
103 \subsubsection faq_intro3 What cmake need?
105 CMake needs some prerequists like :
107 \li c, c++ and java compiler regards to developers
108 \li ccmake for graphical used of CMake
109 \li cmake <a href="http://www.cmake.org/cmake/resources/software.html">(download page)</a>
112 \li Visual C++ 2010 Express <a href="http://www.microsoft.com/express/Downloads/#2010-Visual-CPP">(download page)</a>
113 \li cmake <a href="http://www.cmake.org/cmake/resources/software.html">(download page)</a>
114 \li Set CC, CXX, INCLUDE, LIB and RC to environment variables.
116 SET --> CC TO --> C:\MicrosoftVisualStudio10\VC\bin\cl
117 --> CXX --> C:\MicrosoftVisualStudio10\VC\bin\cl
118 --> INCLUDE --> C:\MicrosoftVisualStudio10\VC\include;C:\Program Files\Microsoft SDKs\Windows\v7.OA\Include
119 --> LIB --> C:\MicrosoftVisualStudio10\VC\lib;C:\Program Files\Microsoft SDKs\Windows\v7.OA\Lib
120 --> RC --> C:\Program Files\Microsoft SDKs\Windows\v7.OA\bin\RC
122 \li Add to environment variable "Path" the path where to find nmake executable and some needed files.
125 ;C\MicrosoftVisualStudio10\VC\bin
126 ;C\MicrosoftVisualStudio10\Common7\IDE
127 ;C:\Program Files\Microsoft SDKs\Windows\v7.OA\bin
128 ;C:\Program Files\Microsoft SDKs\Windows\v7.OA\Lib
129 ;C:\Program Files\Microsoft SDKs\Windows\v7.OA\bInclude
131 \subsection faq_cmakeoption Cmake options
133 \subsubsection faq_cmakeoption1 Liste of options
136 "cmake -D[name]=[value] ... ./"
138 [name] enable_gtnets [value] ON/OFF or TRUE/FALSE or 1/0
139 enable_java ON/OFF or TRUE/FALSE or 1/0
140 enable_lua ON/OFF or TRUE/FALSE or 1/0
141 enable_ruby ON/OFF or TRUE/FALSE or 1/0
142 enable_compile_optimizations ON/OFF or TRUE/FALSE or 1/0
143 enable_compile_warnings ON/OFF or TRUE/FALSE or 1/0
144 enable_smpi ON/OFF or TRUE/FALSE or 1/0
145 enable_maintainer_mode ON/OFF or TRUE/FALSE or 1/0
146 enable_supernovae ON/OFF or TRUE/FALSE or 1/0
147 enable_tracing ON/OFF or TRUE/FALSE or 1/0
148 enable_coverage ON/OFF or TRUE/FALSE or 1/0
149 enable_memcheck ON/OFF or TRUE/FALSE or 1/0
150 enable_model-checking ON/OFF or TRUE/FALSE or 1/0
151 enable_doc ON/OFF or TRUE/FALSE or 1/0
152 gtnets_path <path_to_gtnets_directory>
153 prefix <path_to_install_directory>
154 BIBTEX2HTML <path_to_bibtex2html>
155 with_context auto/ucontext/pthread/window
156 pipol_user <pipol_username>
159 \subsubsection faq_cmakeoption2 Options explaination
161 \li enable_gtnets : set to true implie that user wants to use gtnets.
163 \li enable_java : set to true implie that user wants to add java langage into simgrid compilation.
165 \li enable_lua : set to true implie that user wants to add lua langage into simgrid compilation.
167 \li enable_ruby : set to true implie that user wants to add ruby langage into simgrid compilation.
169 \li enable_compile_optimizations : add flags "-O3 -finline-functions -funroll-loops -fno-strict-aliasing"
171 \li enable_compile_warnings : add flags "-Wall -Wunused -Wmissing-prototypes -Wmissing-declarations -Wpointer-arith -Wchar-subscripts -Wcomment -Wformat -Wwrite-strings -Wno-unused-function -Wno-unused-parameter -Wno-strict-aliasing -Wno-format-nonliteral -Werror"
173 \li enable_smpi : Set to true if you want to use smpi lib. Actually on simgrid v3.4.1 Mac doesn't support lib smpi.
175 \li enable_maintainer_mode : set to true it remakes some files.
177 \li enable_supernovae : set to true make one file for each lib and compile with those generated files.
179 \li enable_tracing : To enable the generation of simulation traces for visualization
181 \li enable_coverage : When set to true this option enable code coverage by setting -fprofile-arcs -ftest-coverage flags.
183 \li enable_memcheck : When set to true this option enable tests for memcheck.
185 \li enable_model-checking : Enable the model checking when set to true.
187 \li enable_doc : Generate the documentation for simgrid with make command. (You can also make the doc manually with command : make html)
189 \li gtnets_path : Path to gtnets install directory (ex /usr)
191 \li prefix : Path where are installed lib/ doc/ and include/ directories (ex /usr/local)
193 \li BIBTEX2HTML : Path where is installed bibtex2html.
195 \li with context : specify which context the user wants to use.
197 \li pipol_user : specify your pipol username if you want to use the pipol-remote command.
199 \subsubsection faq_cmakeoption3 Initialisation
201 Those options are initialized the first time you launch "cmake ." whithout specified option.
208 enable_compile_optimizations off
209 enable_compile_warnings off
211 enable_maintainer_mode off
212 enable_supernovae off
216 enable_model-checking off
226 \subsubsection faq_cmakeoption4 Option's cache and how to reset?
228 When options have been set they are keep into a cache file named "CMakeCache.txt". So if you want
229 reset values you just delete this file located to the project directory.
231 \subsection faq_cmakecompilation Cmake compilation
233 \subsubsection faq_cmakecompilation1 With command line.
236 cmake -D[name]=[value] ... ./
240 \subsubsection faq_cmakecompilation2 With ccmake tool.
245 Then follow instructions.
247 \subsubsection faq_cmakecompilation2bis Build out of source.
249 As cmake generate many files used for compilation, we recommand to make a build directory.
250 For examples you can make :
253 "navarrop@caraja:~/Developments$ cd simgrid/"
254 "navarrop@caraja:~/Developments/simgrid$ mkdir build_directory"
255 "navarrop@caraja:~/Developments/simgrid$ cd build_directory/"
256 "navarrop@caraja:~/Developments/simgrid/build_directory$ cmake ../"
257 "navarrop@caraja:~/Developments/simgrid/build_directory$ make"
260 Or complety out of sources :
263 "navarrop@caraja:~/Developments$ mkdir build_dir"
264 "navarrop@caraja:~/Developments$ cd build_dir/"
265 "navarrop@caraja:~/Developments/build_dir$ cmake ../simgrid/"
266 "navarrop@caraja:~/Developments/build_dir$ make"
269 Those two kind of compilation permit to delete files created by compilation easier.
271 \subsubsection faq_cmakecompilation3 Resume of command line
275 cmake <path> configure the project
276 make build all targets
277 make VERBOSE=1 build all targets and print build command lines
278 make check test all targets and summarize
279 make dist make the distrib
280 make distcheck check the dist (make + make dist + make check)
281 make install install the project (doc/ lib/ include/)
282 make uninstall uninstall the project (doc/ lib/ include/)
283 make clean clean all targets
284 make java-clean clean files created by java option
285 make doc-clean clean files created for making doc
286 make supernovae-clean clean supernovae files
287 make maintainer-clean clean maintainer files
288 make all-clean execute the 5 upper clean command
289 make html Create simgrid documentation
292 When the project have been succesfully compiling and build you can make tests.
296 ctest launch only tests
298 ctest -D Continuous(Start|Update|Configure|Build)
299 ctest -D Continuous(Test|Coverage|MemCheck|Submit)
300 ctest -D Experimental
301 ctest -D Experimental(Start|Update|Configure|Build)
302 ctest -D Experimental(Test|Coverage|MemCheck|Submit)
304 ctest -D Nightly(Start|Update|Configure|Build)
305 ctest -D Nightly(Test|Coverage|MemCheck|Submit)
306 ctest -D NightlyMemoryCheck
309 If you want to test before make a commit you can simply make "ctest -D Experimental" and then you can visualize results submitted into Cdash. <a href="http://cdash.inria.fr/CDash/index.php?project=Simgrid">(Go to Cdash site)</a>.
311 \subsubsection faq_cmakecompilation5 Examples for different mode.
315 cmake -Denable_maintainer_mode=on ./
317 -- lookign for config.h
318 with_context auto change to ucontext
319 GIT_DATE : 2010-05-04~09-59-15
320 GIT_VERSION : 53ec816
321 GIT_SVN_VERSION : 7669
323 Configuration of package `simgrid' (revision 7669) on arch (=4):
325 SITE : Linux_2.6.31-21-generic_x86_64
326 Release : simgrid-3.4~rev7669
328 Compiler: c++ : /usr/bin/c++
329 version: c++ (Ubuntu 4.4.1-4ubuntu9) 4.4.1
330 Compiler: c : /usr/bin/gcc
331 version: gcc (Ubuntu 4.4.1-4ubuntu9) 4.4.1
333 CFlags : -I/usr/lib/ruby/1.8/x86_64-linux -I/usr/include/lua5.1 -g3
335 LDFlags : -L/usr/lib/
337 Context backend: ucontext
349 Simgrid dependencies: -lm -lruby1.8 -module -ldl -llua5.1 -lrt
350 Gras dependencies : -lm -lpthread -lrt
353 INSTALL_PREFIX: /usr/local
357 -- Build files have been written to: /home/navarrop/Developments/simgrid
362 cmake -Dsupernovae=on ./
364 -- lookign for config.h
365 with_context auto change to ucontext
366 GIT_DATE : 2010-05-04~09-59-15
367 GIT_VERSION : 53ec816
368 GIT_SVN_VERSION : 7669
370 Configuration of package `simgrid' (revision 7669) on arch (=4):
371 BUILDNAME : SUPERNOVAE
372 SITE : Linux_2.6.31-21-generic_x86_64
373 Release : simgrid-3.4~rev7669
375 Compiler: c++ : /usr/bin/c++
376 version: c++ (Ubuntu 4.4.1-4ubuntu9) 4.4.1
377 Compiler: c : /usr/bin/gcc
378 version: gcc (Ubuntu 4.4.1-4ubuntu9) 4.4.1
380 CFlags : -O3 -finline-functions -funroll-loops -fno-strict-aliasing -Wall -Wunused -Wmissing-prototypes -Wmissing-declarations -Wpointer-arith -Wchar-subscripts -Wcomment -Wformat -Wwrite-strings -Wno-unused-function -Wno-unused-parameter -Wno-strict-aliasing -Wno-format-nonliteral -Werror -I/usr/lib/ruby/1.8/x86_64-linux -I/usr/include/lua5.1 -g3
382 LDFlags : -L/usr/lib/
384 Context backend: ucontext
396 Simgrid dependencies: -lm -lruby1.8 -module -ldl -llua5.1 -lrt
397 Gras dependencies : -lm -lpthread -lrt
400 INSTALL_PREFIX: /usr/local
404 -- Build files have been written to: /home/navarrop/Developments/simgrid
410 cmake -Dgtnets_path=/home/navarrop/Bureau/usr/ ./
412 -- lookign for config.h
413 with_context auto change to ucontext
414 GIT_DATE : 2010-05-04~09-59-15
415 GIT_VERSION : 53ec816
416 GIT_SVN_VERSION : 7669
418 Configuration of package `simgrid' (revision 7669) on arch (=4):
420 SITE : Linux_2.6.31-21-generic_x86_64
421 Release : simgrid-3.4~rev7669
423 Compiler: c++ : /usr/bin/c++
424 version: c++ (Ubuntu 4.4.1-4ubuntu9) 4.4.1
425 Compiler: c : /usr/bin/gcc
426 version: gcc (Ubuntu 4.4.1-4ubuntu9) 4.4.1
428 CFlags : -O3 -finline-functions -funroll-loops -fno-strict-aliasing -Wall -Wunused -Wmissing-prototypes -Wmissing-declarations -Wpointer-arith -Wchar-subscripts -Wcomment -Wformat -Wwrite-strings -Wno-unused-function -Wno-unused-parameter -Wno-strict-aliasing -Wno-format-nonliteral -Werror -I/usr/lib/ruby/1.8/x86_64-linux -L/usr/lib -I/usr/include/gtnets -I/usr/include/lua5.1 -g3
429 CPPFlags: -L/usr/lib -I/usr/include/gtnets
430 LDFlags : -L/usr/lib/
432 Context backend: ucontext
444 Simgrid dependencies: -lm -lruby1.8 -module -ldl -llua5.1 -lgtnets -lrt
445 Gras dependencies : -lm -lpthread -lrt
448 INSTALL_PREFIX: /usr/local
452 -- Build files have been written to: /home/navarrop/Developments/simgrid
456 \subsection faq_cmakeinstall How to install with cmake?
458 \subsubsection faq_cmakeinstall1 From svn.
461 cmake -Denable_maintainer_mode=on -Dprefix=/home/navarrop/Bureau/install_simgrid ./
466 \subsubsection faq_cmakeinstall2 From a distrib
469 cmake -Dprefix=/home/navarrop/Bureau/install_simgrid ./
474 \subsection faq_cmakehowto How to modified sources files for developers
476 \subsubsection faq_cmakehowto1 Add an executable or examples.
478 If you want make an executable you have to create a CMakeList.txt to the src directory.
479 You must specified where to create the executable, source list, dependencies and the name of the binary.
482 cmake_minimum_required(VERSION 2.6)
484 set(EXECUTABLE_OUTPUT_PATH "./")
485 set(LIBRARY_OUTPUT_PATH "${PROJECT_DIRECTORY}/lib")
487 add_executable(get_sender get_sender.c) #add_executable(<name_of_target> <src list>)
489 ### Add definitions for compile
490 target_link_libraries(get_sender simgrid m pthread -fprofile-arcs) #target_link_libraries(<name_of_targe> <dependencies>)
493 Then you have to modified <project/directory>/buildtools/Cmake/MakeExeLib.cmake and add
496 add_subdirectory(${PROJECT_DIRECTORY}/<path_where_is_CMakeList.txt>)
499 \subsubsection faq_cmakehowto2 Delete/add sources to lib.
501 If you want modified, add or delete source files from a library you have to edit <project/directory>/buildtools/Cmake/DefinePackages.cmake
505 ${PROJECT_DIRECTORY}/src/java/simgrid/msg/MsgException.java
506 ${PROJECT_DIRECTORY}/src/java/simgrid/msg/JniException.java
507 ${PROJECT_DIRECTORY}/src/java/simgrid/msg/NativeException.java
508 ${PROJECT_DIRECTORY}/src/java/simgrid/msg/HostNotFoundException.java
509 ${PROJECT_DIRECTORY}/src/java/simgrid/msg/ProcessNotFoundException.java
510 ${PROJECT_DIRECTORY}/src/java/simgrid/msg/Msg.java
511 ${PROJECT_DIRECTORY}/src/java/simgrid/msg/Process.java
512 ${PROJECT_DIRECTORY}/src/java/simgrid/msg/Host.java
513 ${PROJECT_DIRECTORY}/src/java/simgrid/msg/Task.java
514 ${PROJECT_DIRECTORY}/src/java/simgrid/msg/MsgNative.java
515 ${PROJECT_DIRECTORY}/src/java/simgrid/msg/ApplicationHandler.java
516 ${PROJECT_DIRECTORY}/src/java/simgrid/msg/Sem.java
520 \subsubsection faq_cmakehowto3 Add test
522 If you want modified, add or delete tests you have to edit <project/directory>/buildtools/Cmake/AddTests.cmake
523 with this function : ADD_TEST(<name> <bin> <ARGS>)
526 add_test(test-simdag-1 ${PROJECT_DIRECTORY}/testsuite/simdag/sd_test --cfg=path:${PROJECT_DIRECTORY}/testsuite/simdag small_platform_variable.xml)
529 \subsection faq_PIPOL Pipol-remote
531 Now we offer the possibility to test your local sources on pipol platforms before a commit. Of course you have to be user of pipol <a href="https://pipol.inria.fr/users/">(Account request)</a> cause you need to give your pipol_username to cmake. Here is a list of available systems :
533 amd64_kvm-linux-debian-lenny
534 amd64_kvm-linux-debian-testing
536 amd64-linux-centos-5.dd.gz
537 amd64-linux-debian-etch.dd.gz
538 amd64-linux-debian-lenny.dd.gz
539 amd64-linux-debian-testing.dd.gz
540 amd64-linux-fedora-core10.dd.gz
541 amd64-linux-fedora-core11.dd.gz
542 amd64-linux-fedora-core12.dd.gz
543 amd64-linux-fedora-core13.dd.gz
544 amd64-linux-fedora-core7.dd.gz
545 amd64-linux-fedora-core8.dd.gz
546 amd64-linux-fedora-core9.dd.gz
547 amd64-linux-mandriva-2007_springs_powerpack.dd.gz
548 amd64-linux-mandriva-2009_powerpack.dd.gz
549 amd64-linux-opensuse-11.dd.gz
550 amd64-linux-redhatEL-5.0.dd.gz
551 amd64-linux-suse-LES10.dd.gz
552 amd64-linux-ubuntu-feisty.dd.gz
553 amd64-linux-ubuntu-hardy.dd.gz
554 amd64-linux-ubuntu-intrepid.dd.gz
555 amd64-linux-ubuntu-jaunty.dd.gz
556 amd64-linux-ubuntu-karmic.dd.gz
557 amd64-linux-ubuntu-lucid.dd.gz
558 amd64-unix-freebsd-7.dd.gz
559 amd64-windows-server-2003-64bits.dd.gz
560 amd64-windows-server-2008-64bits.dd.gz
561 i386_kvm-linux-debian-lenny
562 i386_kvm-linux-debian-testing
563 i386_kvm-linux-fedora-core13
564 i386_kvm-windows-xp-pro-sp3
565 i386-linux-centos-5.dd.gz
566 i386-linux-debian-etch.dd.gz
567 i386-linux-debian-lenny.dd.gz
568 i386-linux-debian-testing.dd.gz
569 i386-linux-fedora-core10.dd.gz
570 i386-linux-fedora-core11.dd.gz
571 i386-linux-fedora-core12.dd.gz
572 i386-linux-fedora-core13.dd.gz
573 i386-linux-fedora-core7.dd.gz
574 i386-linux-fedora-core8.dd.gz
575 i386-linux-fedora-core9.dd.gz
576 i386-linux-mandriva-2007_springs_powerpack.dd.gz
577 i386-linux-mandriva-2009_powerpack.dd.gz
578 i386-linux-opensuse-11.dd.gz
579 i386-linux-redhatEL-5.0.dd.gz
580 i386-linux-suse-LES10.dd.gz
581 i386-linux-ubuntu-feisty.dd.gz
582 i386-linux-ubuntu-hardy.dd.gz
583 i386-linux-ubuntu-intrepid.dd.gz
584 i386-linux-ubuntu-jaunty.dd.gz
585 i386-linux-ubuntu-karmic.dd.gz
586 i386-linux-ubuntu-lucid.dd.gz
587 i386_mac-mac-osx-server-leopard.dd.gz
588 i386-unix-freebsd-7.dd.gz
589 i386-unix-opensolaris-10.dd.gz
590 i386-unix-opensolaris-11.dd.gz
591 i386-unix-solaris-10.dd.gz
592 ia64-linux-debian-lenny.dd
593 ia64-linux-fedora-core9.dd
594 ia64-linux-redhatEL-5.0.dd
595 x86_64_mac-mac-osx-server-snow-leopard.dd.gz
596 x86_mac-mac-osx-server-snow-leopard.dd.gz
599 Two kind of uses are possible :
601 This command copy your source and execute a configure then a build and finish with tests.
602 bob@caraja:~/Developments/simgrid/tmp_build$ make <name_of_image>
604 This command copy your source and execute a \"ctest -D Experimental\" and submit the result to cdash.
605 bob@caraja:~/Developments/simgrid/tmp_build$ make <name_of_image>_experimental
607 All commands are resumed with :
609 bob@caraja:~/Developments/simgrid/tmp_build$ make pipol_experimental_list_images
610 bob@caraja:~/Developments/simgrid/tmp_build$ make pipol_test_list_images
613 \subsection faq_cmakeExplain Explaination of sources files for cmake
617 Those files are the "main parts". One located at the project directory call all the cmake sources files. The others
618 are little projects called by the first for make examples.
620 \li CompleteInFiles.cmake
622 Complete all .in files and define Variables for h files
624 \li GenerateDoc.cmake
626 This file make the html documentation.
630 Here are callled all "CMakeLists.txt" for make executables and libraries.
634 This file is called at the end of the build for summarize environment variables.
636 \li DefinePackages.cmake
638 Here is defined sources packages for compiling libs.
642 Defined flags which are used for compiling sources.
646 Here are made files for the supernovae mode.
650 Here is defined packages for install simgrid and make a distribution.
652 \li MaintainerMode.cmake
654 Part where are generated source files for maintainer mode.
658 Here are defined options and initialized values.
662 All tests are listed.
664 \li CTestConfig.cmake
666 Properties which link tests with dashboard.
668 \subsection faq_cmakeList List of files added for cmake
670 Here is a list of files involved into cmake build (relative to project directory path) :
675 ./buildtools/Cmake/AddTests.cmake
676 ./buildtools/Cmake/CompleteInFiles.cmake
677 ./buildtools/Cmake/CTestConfig.cmake
678 ./buildtools/Cmake/DefinePackages.cmake
679 ./buildtools/Cmake/Distrib.cmake
680 ./buildtools/Cmake/Flags.cmake
681 ./buildtools/Cmake/GenerateDocs.cmake
682 ./buildtools/Cmake/MaintainerMode.cmake
683 ./buildtools/Cmake/MakeExeLib.cmake
684 ./buildtools/Cmake/MakeExeLibWin.cmake
685 ./buildtools/Cmake/MakeJava.cmake
686 ./buildtools/Cmake/Option.cmake
687 ./buildtools/Cmake/PrintArgs.cmake
688 ./buildtools/Cmake/Supernovae.cmake
690 CMakeLists for each binaries or examples:
693 ./teshsuite/gras/empty_main/CMakeLists.txt
694 ./teshsuite/gras/small_sleep/CMakeLists.txt
695 ./teshsuite/gras/datadesc/CMakeLists.txt
696 ./teshsuite/gras/msg_handle/CMakeLists.txt
697 ./teshsuite/simdag/CMakeLists.txt
698 ./teshsuite/simdag/partask/CMakeLists.txt
699 ./teshsuite/simdag/platforms/CMakeLists.txt
700 ./teshsuite/simdag/network/CMakeLists.txt
701 ./teshsuite/simdag/network/mxn/CMakeLists.txt
702 ./teshsuite/simdag/network/p2p/CMakeLists.txt
703 ./teshsuite/xbt/CMakeLists.txt
704 ./teshsuite/msg/CMakeLists.txt
705 ./tools/gras/CMakeLists.txt
706 ./tools/tesh/CMakeLists.txt
707 ./testsuite/simdag/CMakeLists.txt
708 ./testsuite/xbt/CMakeLists.txt
709 ./testsuite/surf/CMakeLists.txt
710 ./examples/gras/properties/CMakeLists.txt
711 ./examples/gras/ping/CMakeLists.txt
712 ./examples/gras/pmm/CMakeLists.txt
713 ./examples/gras/mmrpc/CMakeLists.txt
714 ./examples/gras/synchro/CMakeLists.txt
715 ./examples/gras/timer/CMakeLists.txt
716 ./examples/gras/mutual_exclusion/simple_token/CMakeLists.txt
717 ./examples/gras/spawn/CMakeLists.txt
718 ./examples/gras/chrono/CMakeLists.txt
719 ./examples/gras/rpc/CMakeLists.txt
720 ./examples/gras/all2all/CMakeLists.txt
721 ./examples/simdag/properties/CMakeLists.txt
722 ./examples/simdag/CMakeLists.txt
723 ./examples/simdag/metaxml/CMakeLists.txt
724 ./examples/simdag/dax/CMakeLists.txt
725 ./examples/smpi/CMakeLists.txt
726 ./examples/amok/bandwidth/CMakeLists.txt
727 ./examples/amok/saturate/CMakeLists.txt
728 ./examples/msg/priority/CMakeLists.txt
729 ./examples/msg/properties/CMakeLists.txt
730 ./examples/msg/migration/CMakeLists.txt
731 ./examples/msg/gtnets/CMakeLists.txt
732 ./examples/msg/parallel_task/CMakeLists.txt
733 ./examples/msg/trace/CMakeLists.txt
734 ./examples/msg/suspend/CMakeLists.txt
735 ./examples/msg/masterslave/CMakeLists.txt
736 ./examples/msg/actions/CMakeLists.txt
737 ./examples/msg/sendrecv/CMakeLists.txt
740 \section faq_installation Installing the SimGrid library with Autotools (valid until V3.3.4)
742 Many people have been asking me questions on how to use SimGrid. Quite
743 often, the questions were not really about SimGrid but on the
744 installation process. This section is intended to help people that are
745 not familiar with compiling C files under UNIX. If you follow these
746 instructions and still have some troubles, drop an e-mail to
747 <simgrid-user@lists.gforge.inria.fr>.
749 \subsection faq_compiling Compiling SimGrid from a stable archive
751 First of all, you need to download the latest version of SimGrid from
752 <a href="http://gforge.inria.fr/frs/?group_id=12">here</a>.
753 Suppose you have uncompressed SimGrid in some temporary location of
754 your home directory (say <tt>/home/joe/tmp/simgrid-3.0.1 </tt>). The
755 simplest way to use SimGrid is to install it in your home
756 directory. Change your directory to
757 <tt>/home/joe/tmp/simgrid-3.0.1</tt> and type
760 ./configure --prefix=$HOME
765 If at some point, something fails, check the section \ref faq_trouble_compil .
766 If it does not help, you can report this problem to the
767 list but, please, avoid sending a laconic mail like "There is a problem. Is it
768 okay?". Send the config.log file which is automatically generated by
769 configure. Try to capture both the standard output and the error output of the
770 <tt>make</tt> command with <tt>script</tt>. There is no way for us to help you
771 without the relevant bits of information.
773 Now, the following directory should have been created :
775 \li <tt>/home/joe/doc/simgrid/html/</tt>
776 \li <tt>/home/joe/lib/</tt>
777 \li <tt>/home/joe/include/</tt>
779 SimGrid is not a binary, it is a library. Both a static and a dynamic
780 version are available. Here is what you can find if you try a <tt>ls
783 \verbatim libsimgrid.a libsimgrid.la libsimgrid.so libsimgrid.so.0 libsimgrid.so.0.0.1
786 Thus, there is two ways to link your program with SimGrid:
787 \li Either you use the static version, e.g
788 \verbatim gcc libsimgrid.a -o MainProgram MainProgram.c
790 In this case, all the SimGrid functions are directly
791 included in <tt>MainProgram</tt> (hence a bigger binary).
792 \li Either you use the dynamic version (the preferred method)
793 \verbatim gcc -lsimgrid -o MainProgram MainProgram.c
795 In this case, the SimGrid functions are not included in
796 <tt>MainProgram</tt> and you need to set your environment
797 variable in such a way that <tt>libsimgrid.so</tt> will be
798 found at runtime. This can be done by adding the following
799 line in your .bashrc (if you use bash and if you have
800 installed the SimGrid libraries in your home directory):
801 \verbatim export LD_LIBRARY_PATH=$HOME/lib/:$LD_LIBRARY_PATH
804 \subsection faq_compiling_java Java bindings don't get compiled
806 The configure script detects automatically whether you have the
807 softwares needed to use the Java bindings or not. At the end of the
808 configure, you can see the configuration picked by the script, which
809 should look similar to
810 \verbatim Configuration of package simgrid' (version 3.3.4-svn) on
813 Compiler: gcc (version: )
815 CFlags: -O3 -finline-functions -funroll-loops -fno-strict-aliasing -Wall -Wunused -Wmissing-prototypes -Wmissing-declarations -Wpointer-arith -Wchar-subscripts -Wcomment -Wformat -Wwrite-strings -Wno-unused-function -Wno-unused-parameter -Wno-strict-aliasing -Wno-format-nonliteral -Werror -g3
819 Context backend: ucontext
826 In this example, Java backends won't be compiled.
828 On Debian-like systems (which includes ubuntu), you need the following
829 packages: sun-java6-jdk libgcj10-dev. If you cannot find the
830 libgcj10-dev, try another version, like libgcj9-dev (on Ubuntu before
831 9.10) or libgcj11-dev (not released yet, but certainly one day).
832 Please note that you need to activate the contrib and non-free
833 repositories in Debian, and the universe ones in Ubuntu. Java comes at
836 \subsection faq_compiling_snapshoot SimGrid development snapshots
838 We have very high standards on software quality, and we are reluctant releasing
839 a stable release as long as there is still some known bug in the code base. In
840 addition, we added quite an extensive test base, making sure that we correctly
841 test the most important parts of the tool.
843 As an unfortunate conclusion, there may be some time between the stable
844 releases. If you want to benefit from the most recent features we introduced,
845 but don't want to take the risk of an untested version from the SVN, then
846 development snapshots are done for you.
848 These are pre-releases of SimGrid that still fail some tests about features
849 that almost nobody use, or on platforms not being in our core target (which is
850 Linux, Mac, other Unixes and Windows, from the most important to the less
851 one). That means that using this development releases should be safe for most
854 These archives can be found on
855 <a href="http://www.loria.fr/~quinson/simgrid.html">this web page</a>. Once you
856 got the lastest archive, you can compile it just like any archive (see above).
858 \subsection faq_compiling_svn Compiling SimGrid from the SVN
860 The project development takes place in the SVN, where all changes are
861 committed when they happen. Then every once in a while, we make sure that the
862 code quality meets our standard and release an archive from the code in the
863 SVN. We afterward go back to the development in the SVN. So, if you need a
864 recently added feature and can afford some little problem with the stability
865 of the lastest features, you may want to use the SVN version instead of a
868 For that, you first need to get the "simgrid" module from
869 <a href="http://gforge.inria.fr/scm/?group_id=12">here</a>.
871 You won't find any <tt>configure</tt> and a few other things
872 (<tt>Makefile.in</tt>'s, documentation, ...) will be missing as well. The
873 reason for that is that all these files have to be regenerated using the
874 latest versions of <tt>autoconf</tt>, <tt>libtool</tt>, <tt>automake</tt>
875 (>1.9) and <tt>doxygen</tt> (>1.4). To generate the <tt>configure</tt> and
876 the <tt>Makefile.in</tt>'s, you just have to launch the <tt>bootstrap</tt>
877 command that resides in the top of the source tree. Then just follow the
878 instructions of Section \ref faq_compiling.
880 We insist on the fact that you really need the latest versions of
881 autoconf, automake and libtool. Doing this step on exotic architectures/systems
882 (i.e. anything different from a recent linux distribution) may be
883 ... uncertain. If you need to compile the SVN version on a machine where all these
884 dependencies are not met, the easiest is to do <tt>make dist</tt> in the SVN
885 directory of another machine where all dependencies are met. It will create an
886 archive you may deploy on other sites just as a regular stable release.
888 In summary, the following commands will checkout the SVN, regenerate the
889 configure script and friends, configure SimGrid and build it.
891 \verbatim svn checkout svn://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk simgrid
894 ./configure --enable-maintainer-mode --prefix=<where to install SimGrid>
897 Then, if you want to install SimGrid on the current box, just do:
898 \verbatim make install \endverbatim
900 If you want to build an snapshot of the SVN to deploy it on another box (for
901 example because the other machine don't have the autotools), do:
902 \verbatim make dist \endverbatim
904 Moreover, you should never call the autotools manually since you must run
905 them in a specific order with specific arguments. Most of the times, the
906 makefiles will automatically call the tools for you. When it's not possible
907 (such as the first time you checkout the SVN), use the ./bootstrap command
908 to call them explicitly.
911 \subsection faq_setting_MSG Setting up your own MSG code
913 Do not build your simulator by modifying the SimGrid examples. Go
914 outside the SimGrid source tree and create your own working directory
915 (say <tt>/home/joe/SimGrid/MyFirstScheduler/</tt>).
917 Suppose your simulation has the following structure (remember it is
918 just an example to illustrate a possible way to compile everything;
919 feel free to organize it as you want).
921 \li <tt>sched.h</tt>: a description of the core of the
922 scheduler (i.e. which functions are can be used by the
923 agents). For example we could find the following functions
924 (master, forwarder, slave).
926 \li <tt>sched.c</tt>: a C file including <tt>sched.h</tt> and
927 implementing the core of the scheduler. Most of these
928 functions use the MSG functions defined in section \ref
931 \li <tt>masterslave.c</tt>: a C file with the main function, i.e.
932 the MSG initialization (MSG_global_init()), the platform
933 creation (e.g. with MSG_create_environment()), the
934 deployment phase (e.g. with MSG_function_register() and
935 MSG_launch_application()) and the call to
938 To compile such a program, we suggest to use the following
939 Makefile. It is a generic Makefile that we have used many times with
940 our students when we teach the C language.
944 masterslave: masterslave.o sched.o
946 INSTALL_PATH = $$HOME
948 PEDANTIC_PARANOID_FREAK = -O0 -Wshadow -Wcast-align \
949 -Waggregate-return -Wmissing-prototypes -Wmissing-declarations \
950 -Wstrict-prototypes -Wmissing-prototypes -Wmissing-declarations \
951 -Wmissing-noreturn -Wredundant-decls -Wnested-externs \
952 -Wpointer-arith -Wwrite-strings -finline-functions
953 REASONABLY_CAREFUL_DUDE = -Wall
954 NO_PRAYER_FOR_THE_WICKED = -w -O2
955 WARNINGS = $(REASONABLY_CAREFUL_DUDE)
956 CFLAGS = -g $(WARNINGS)
958 INCLUDES = -I$(INSTALL_PATH)/include
959 DEFS = -L$(INSTALL_PATH)/lib/
960 LDADD = -lm -lsimgrid
964 $(CC) $(INCLUDES) $(DEFS) $(CFLAGS) $^ $(LIBS) $(LDADD) -o $@
967 $(CC) $(INCLUDES) $(DEFS) $(CFLAGS) -c -o $@ $<
970 rm -f $(BIN_FILES) *.o *~
976 The first two lines indicates what should be build when typing make
977 (<tt>masterslave</tt>) and of which files it is to be made of
978 (<tt>masterslave.o</tt> and <tt>sched.o</tt>). This makefile assumes
979 that you have set up correctly your <tt>LD_LIBRARY_PATH</tt> variable
980 (look, there is a <tt>LDADD = -lm -lsimgrid</tt>). If you prefer using
981 the static version, remove the <tt>-lsimgrid</tt> and add a
982 <tt>$(INSTALL_PATH)/lib/libsimgrid.a</tt> on the next line, right
983 after the <tt>LIBS = </tt>.
985 More generally, if you have never written a Makefile by yourself, type
986 in a terminal : <tt>info make</tt> and read the introduction. The
987 previous example should be enough for a first try but you may want to
988 perform some more complex compilations...
990 \subsection faq_setting_GRAS Setting up your own GRAS code
992 If you use the GRAS interface instead of the MSG one, then previous section
993 is not the better source of information. Instead, you should check the GRAS
994 tutorial in general, and the \ref GRAS_tut_tour_setup in particular.
996 \section faq_howto Feature related questions
998 \subsection faq_MIA "Could you please add (your favorite feature here) to SimGrid?"
1000 Here is the deal. The whole SimGrid project (MSG, SURF, GRAS, ...) is
1001 meant to be kept as simple and generic as possible. We cannot add
1002 functions for everybody's needs when these functions can easily be
1003 built from the ones already in the API. Most of the time, it is
1004 possible and when it was not possible we always have upgraded the API
1005 accordingly. When somebody asks us a question like "How to do that?
1006 Is there a function in the API to simply do this?", we're always glad
1007 to answer and help. However if we don't need this code for our own
1008 need, there is no chance we're going to write it... it's your job! :)
1009 The counterpart to our answers is that once you come up with a neat
1010 implementation of this feature (task duplication, RPC, thread
1011 synchronization, ...), you should send it to us and we will be glad to
1012 add it to the distribution. Thus, other people will take advantage of
1013 it (and we don't have to answer this question again and again ;).
1015 You'll find in this section a few "Missing In Action" features. Many
1016 people have asked about it and we have given hints on how to simply do
1017 it with MSG. Feel free to contribute...
1019 \subsection faq_MIA_MSG MSG features
1021 \subsubsection faq_MIA_examples I want some more complex MSG examples!
1023 Many people have come to ask me a more complex example and each time,
1024 they have realized afterward that the basics were in the previous three
1027 Of course they have often been needing more complex functions like
1028 MSG_process_suspend(), MSG_process_resume() and
1029 MSG_process_isSuspended() (to perform synchronization), or
1030 MSG_task_Iprobe() and MSG_process_sleep() (to avoid blocking
1031 receptions), or even MSG_process_create() (to design asynchronous
1032 communications or computations). But the examples are sufficient to
1035 We know. We should add some more examples, but not really some more
1036 complex ones... We should add some examples that illustrate some other
1037 functionalists (like how to simply encode asynchronous
1038 communications, RPC, process migrations, thread synchronization, ...)
1039 and we will do it when we will have a little bit more time. We have
1040 tried to document the examples so that they are understandable. Tell
1041 us if something is not clear and once again feel free to participate!
1044 \subsubsection faq_MIA_taskdup Missing in action: MSG Task duplication/replication
1046 There is no task duplication in MSG. When you create a task, you can
1047 process it or send it somewhere else. As soon as a process has sent
1048 this task, he doesn't have this task anymore. It's gone. The receiver
1049 process has got the task. However, you could decide upon receiving to
1050 create a "copy" of a task but you have to handle by yourself the
1051 semantic associated to this "duplication".
1053 As we already told, we prefer keeping the API as simple as
1054 possible. This kind of feature is rather easy to implement by users
1055 and the semantic you associate really depends on people. Having a
1056 *generic* task duplication mechanism is not that trivial (in
1057 particular because of the data field). That is why I would recommand
1058 that you write it by yourself even if I can give you advice on how to
1061 You have the following functions to get informations about a task:
1062 MSG_task_get_name(), MSG_task_get_compute_duration(),
1063 MSG_task_get_remaining_computation(), MSG_task_get_data_size(),
1064 and MSG_task_get_data().
1066 You could use a dictionary (#xbt_dict_t) of dynars (#xbt_dynar_t). If
1067 you still don't see how to do it, please come back to us...
1069 \subsubsection faq_MIA_asynchronous I want to do asynchronous communications in MSG
1071 In the past (version <= 3.4), there was no function to perform asynchronous communications.
1072 It could easily be implemented by creating new process when needed though. Since version 3.5,
1073 we have introduced the following functions:
1078 - MSG_comm_waitall()
1079 - MSG_comm_waitany()
1080 - MSG_comm_destroy()
1082 We refer you to the description of these functions for more details on their usage as well
1083 as to the exemple section on \ref MSG_ex_asynchronous_communications.
1085 \subsubsection faq_MIA_thread_synchronization I need to synchronize my MSG processes
1087 You obviously cannot use pthread_mutexes of pthread_conds since we handle every
1088 scheduling related decision within SimGrid.
1090 In the past (version <=3.3.4) you could do it by playing with
1091 MSG_process_suspend() and MSG_process_resume() or with fake communications (using MSG_task_get(),
1092 MSG_task_put() and MSG_task_Iprobe()).
1094 Since version 3.4, you can use classical synchronization structures. See page \ref XBT_synchro or simply check in
1095 include/xbt/synchro_core.h.
1097 \subsubsection faq_MIA_host_load Where is the get_host_load function hidden in MSG?
1099 There is no such thing because its semantic wouldn't be really
1100 clear. Of course, it is something about the amount of host throughput,
1101 but there is as many definition of "host load" as people asking for
1102 this function. First, you have to remember that resource availability
1103 may vary over time, which make any load notion harder to define.
1105 It may be instantaneous value or an average one. Moreover it may be only the
1106 power of the computer, or may take the background load into account, or may
1107 even take the currently running tasks into account. In some SURF models,
1108 communications have an influence on computational power. Should it be taken
1111 First of all, it's near to impossible to predict the load beforehands in the
1112 simulator since it depends on too much parameters (background load
1113 variation, bandwidth sharing algorithmic complexity) some of them even being
1114 not known beforehands (other task starting at the same time). So, getting
1115 this information is really hard (just like in real life). It's not just that
1116 we want MSG to be as painful as real life. But as it is in some way
1117 realistic, we face some of the same problems as we would face in real life.
1119 How would you do it for real? The most common option is to use something
1120 like NWS that performs active probes. The best solution is probably to do
1121 the same within MSG, as in next code snippet. It is very close from what you
1122 would have to do out of the simulator, and thus gives you information that
1123 you could also get in real settings to not hinder the realism of your
1127 double get_host_load() {
1128 m_task_t task = MSG_task_create("test", 0.001, 0, NULL);
1129 double date = MSG_get_clock();
1131 MSG_task_execute(task);
1132 date = MSG_get_clock() - date;
1133 MSG_task_destroy(task);
1134 return (0.001/date);
1138 Of course, it may not match your personal definition of "host load". In this
1139 case, please detail what you mean on the mailing list, and we will extend
1140 this FAQ section to fit your taste if possible.
1142 \subsubsection faq_MIA_communication_time How can I get the *real* communication time?
1144 Communications are synchronous and thus if you simply get the time
1145 before and after a communication, you'll only get the transmission
1146 time and the time spent to really communicate (it will also take into
1147 account the time spent waiting for the other party to be
1148 ready). However, getting the *real* communication time is not really
1149 hard either. The following solution is a good starting point.
1154 m_task_t task = MSG_task_create("Task", task_comp_size, task_comm_size,
1155 calloc(1,sizeof(double)));
1156 *((double*) task->data) = MSG_get_clock();
1157 MSG_task_put(task, slaves[i % slaves_count], PORT_22);
1158 INFO0("Send completed");
1163 m_task_t task = NULL;
1166 time1 = MSG_get_clock();
1167 a = MSG_task_get(&(task), PORT_22);
1168 time2 = MSG_get_clock();
1169 if(time1<*((double *)task->data))
1170 time1 = *((double *) task->data);
1171 INFO1("Communication time : \"%f\" ", time2-time1);
1173 MSG_task_destroy(task);
1178 \subsection faq_MIA_SimDag SimDag related questions
1180 \subsubsection faq_SG_comm Implementing communication delays between tasks.
1182 A classic question of SimDag newcomers is about how to express a
1183 communication delay between tasks. The thing is that in SimDag, both
1184 computation and communication are seen as tasks. So, if you want to
1185 model a data dependency between two DAG tasks t1 and t2, you have to
1186 create 3 SD_tasks: t1, t2 and c and add dependencies in the following
1190 SD_task_dependency_add(NULL, NULL, t1, c);
1191 SD_task_dependency_add(NULL, NULL, c, t2);
1194 This way task t2 cannot start before the termination of communication c
1195 which in turn cannot start before t1 ends.
1197 When creating task c, you have to associate an amount of data (in bytes)
1198 corresponding to what has to be sent by t1 to t2.
1200 Finally to schedule the communication task c, you have to build a list
1201 comprising the workstations on which t1 and t2 are scheduled (w1 and w2
1202 for example) and build a communication matrix that should look like
1205 \subsubsection faq_SG_DAG How to implement a distributed dynamic scheduler of DAGs.
1207 Distributed is somehow "contagious". If you start making distributed
1208 decisions, there is no way to handle DAGs directly anymore (unless I
1209 am missing something). You have to encode your DAGs in term of
1210 communicating process to make the whole scheduling process
1211 distributed. Here is an example of how you could do that. Assume T1
1212 has to be done before T2.
1215 int your_agent(int argc, char *argv[] {
1217 T1 = MSG_task_create(...);
1218 T2 = MSG_task_create(...);
1222 if(cond) MSG_task_execute(T1);
1224 if((MSG_task_get_remaining_computation(T1)=0.0) && (you_re_in_a_good_mood))
1225 MSG_task_execute(T2)
1227 /* do something else */
1233 If you decide that the distributed part is not that much important and that
1234 DAG is really the level of abstraction you want to work with, then you should
1235 give a try to \ref SD_API.
1237 \subsection faq_MIA_generic Generic features
1239 \subsubsection faq_more_processes Increasing the amount of simulated processes
1241 Here are a few tricks you can apply if you want to increase the amount
1242 of processes in your simulations.
1244 - <b>A few thousands of simulated processes</b> (soft tricks)\n
1245 SimGrid can use either pthreads library or the UNIX98 contextes. On
1246 most systems, the number of pthreads is limited and then your
1247 simulation may be limited for a stupid reason. This is especially
1248 true with the current linux pthreads, and I cannot get more than
1249 2000 simulated processes with pthreads on my box. The UNIX98
1250 contexts allow me to raise the limit to 25,000 simulated processes
1252 The <tt>--with-context</tt> option of the <tt>./configure</tt>
1253 script allows you to choose between UNIX98 contextes
1254 (<tt>--with-context=ucontext</tt>) and the pthread version
1255 (<tt>--with-context=pthread</tt>). The default value is ucontext
1256 when the script detect a working UNIX98 context implementation. On
1257 Windows boxes, the provided value is discarded and an adapted
1258 version is picked up.\n\n
1259 We experienced some issues with contextes on some rare systems
1260 (solaris 8 and lower or old alpha linuxes comes to mind). The main
1261 problem is that the configure script detect the contextes as being
1262 functional when it's not true. If you happen to use such a system,
1263 switch manually to the pthread version, and provide us with a good
1264 patch for the configure script so that it is done automatically ;)
1266 - <b>Hundred thousands of simulated processes</b> (hard-core tricks)\n
1267 As explained above, SimGrid can use UNIX98 contextes to represent
1268 and handle the simulated processes. Thanks to this, the main
1269 limitation to the number of simulated processes becomes the
1270 available memory.\n\n
1271 Here are some tricks I had to use in order to run a token ring
1272 between 25,000 processes on my laptop (1Gb memory, 1.5Gb swap).\n
1273 - First of all, make sure your code runs for a few hundreds
1274 processes before trying to push the limit. Make sure it's
1275 valgrind-clean, ie that valgrind does not report neither memory
1276 error nor memory leaks. Indeed, numerous simulated processes
1277 result in *fat* simulation hindering debugging.
1278 - It was really boring to write 25,000 entries in the deployment
1279 file, so I wrote a little script
1280 <tt>examples/gras/mutual_exclusion/simple_token/make_deployment.pl</tt>, which you may
1281 want to adapt to your case. You could also think about hijacking
1282 the SURFXML parser (have look at \ref faq_flexml_bypassing).
1283 - The deployment file became quite big, so I had to do what is in
1284 the FAQ entry \ref faq_flexml_limit
1285 - Each UNIX98 context has its own stack entry. As debugging this is
1286 quite hairly, the default value is a bit overestimated so that
1287 user don't get into trouble about this. You want to tune this
1288 size to increse the number of processes. This is the
1289 <tt>STACK_SIZE</tt> define in
1290 <tt>src/xbt/xbt_context_sysv.c</tt>, which is 128kb by default.
1291 Reduce this as much as you can, but be warned that if this value
1292 is too low, you'll get a segfault. The token ring example, which
1293 is quite simple, runs with 40kb stacks.
1294 - You may tweak the logs to reduce the stack size further. When
1295 logging something, we try to build the string to display in a
1296 char array on the stack. The size of this array is constant (and
1297 equal to XBT_LOG_BUFF_SIZE, defined in include/xbt/log/h). If the
1298 string is too large to fit this buffer, we move to a dynamically
1299 sized buffer. In which case, we have to traverse one time the log
1300 event arguments to compute the size we need for the buffer,
1301 malloc it, and traverse the argument list again to do the actual
1303 The idea here is to move XBT_LOG_BUFF_SIZE to 1, forcing the logs
1304 to use a dynamic array each time. This allows us to lower further
1305 the stack size at the price of some performance loss...\n
1306 This allowed me to run the reduce the stack size to ... 4k. Ie,
1307 on my 1Gb laptop, I can run more than 250,000 processes!
1309 \subsubsection faq_MIA_batch_scheduler Is there a native support for batch schedulers in SimGrid?
1311 No, there is no native support for batch schedulers and none is
1312 planned because this is a very specific need (and doing it in a
1313 generic way is thus very hard). However some people have implemented
1314 their own batch schedulers. Vincent Garonne wrote one during his PhD
1315 and put his code in the contrib directory of our SVN so that other can
1316 keep working on it. You may find inspiring ideas in it.
1318 \subsubsection faq_MIA_checkpointing I need a checkpointing thing
1320 Actually, it depends on whether you want to checkpoint the simulation, or to
1321 simulate checkpoints.
1323 The first one could help if your simulation is a long standing process you
1324 want to keep running even on hardware issues. It could also help to
1325 <i>rewind</i> the simulation by jumping sometimes on an old checkpoint to
1326 cancel recent calculations.\n
1327 Unfortunately, such thing will probably never exist in SG. One would have to
1328 duplicate all data structures because doing a rewind at the simulator level
1329 is very very hard (not talking about the malloc free operations that might
1330 have been done in between). Instead, you may be interested in the Libckpt
1331 library (http://www.cs.utk.edu/~plank/plank/www/libckpt.html). This is the
1332 checkpointing solution used in the condor project, for example. It makes it
1333 easy to create checkpoints (at the OS level, creating something like core
1334 files), and rerunning them on need.
1336 If you want to simulate checkpoints instead, it means that you want the
1337 state of an executing task (in particular, the progress made towards
1338 completion) to be saved somewhere. So if a host (and the task executing on
1339 it) fails (cf. #MSG_HOST_FAILURE), then the task can be restarted
1340 from the last checkpoint.\n
1342 Actually, such a thing does not exists in SimGrid either, but it's just
1343 because we don't think it is fundamental and it may be done in the user code
1344 at relatively low cost. You could for example use a watcher that
1345 periodically get the remaining amount of things to do (using
1346 MSG_task_get_remaining_computation()), or fragment the task in smaller
1349 \subsection faq_platform Platform building and Dynamic resources
1351 \subsubsection faq_platform_example Where can I find SimGrid platform files?
1353 There is several little examples in the archive, in the examples/msg
1354 directory. From time to time, we are asked for other files, but we
1355 don't have much at hand right now.
1357 You should refer to the Platform Description Archive
1358 (http://pda.gforge.inria.fr) project to see the other platform file we
1359 have available, as well as the Simulacrum simulator, meant to generate
1360 SimGrid platforms using all classical generation algorithms.
1362 \subsubsection faq_platform_alnem How can I automatically map an existing platform?
1364 We are working on a project called ALNeM (Application-Level Network
1365 Mapper) which goal is to automatically discover the topology of an
1366 existing network. Its output will be a platform description file
1367 following the SimGrid syntax, so everybody will get the ability to map
1368 their own lab network (and contribute them to the catalog project).
1369 This tool is not ready yet, but it move quite fast forward. Just stay
1372 \subsubsection faq_platform_synthetic Generating synthetic but realistic platforms
1374 The third possibility to get a platform file (after manual or
1375 automatic mapping of real platforms) is to generate synthetic
1376 platforms. Getting a realistic result is not a trivial task, and
1377 moreover, nobody is really able to define what "realistic" means when
1378 speaking of topology files. You can find some more thoughts on this
1380 <a href="http://graal.ens-lyon.fr/~alegrand/articles/Simgrid-Introduction.pdf">slides</a>.
1382 If you are looking for an actual tool, there we have a little tool to
1383 annotate Tiers-generated topologies. This perl-script is in
1384 <tt>tools/platform_generation/</tt> directory of the SVN. Dinda et Al.
1385 released a very comparable tool, and called it GridG.
1387 \subsubsection faq_SURF_dynamic Expressing dynamic resource availability in platform files
1389 A nice feature of SimGrid is that it enables you to seamlessly have
1390 resources whose availability change over time. When you build a
1391 platform, you generally declare hosts like that:
1394 <host id="host A" power="100.00"/>
1397 If you want the availability of "host A" to change over time, the only
1398 thing you have to do is change this definition like that:
1401 <host id="host A" power="100.00" availability_file="trace_A.txt" state_file="trace_A_failure.txt"/>
1404 For hosts, availability files are expressed in fraction of available
1405 power. Let's have a look at what "trace_A.txt" may look like:
1414 At time 0, our host will deliver 100 flop/s. At time 11.0, it will
1415 deliver only 50 flop/s until time 20.0 where it will will start
1416 delivering 90 flop/s. Last at time 21.0 (20.0 plus the periodicity
1417 1.0), we'll be back to the beginning and it will deliver 100 flop/s.
1419 Now let's look at the state file:
1426 A negative value means "off" while a positive one means "on". At time
1427 1.0, the host is on. At time 1.0, it is turned off and at time 2.0, it
1428 is turned on again until time 12 (2.0 plus the periodicity 10.0). It
1429 will be turned on again at time 13.0 until time 23.0, and so on.
1431 Now, let's look how the same kind of thing can be done for network
1432 links. A usual declaration looks like:
1435 <link id="LinkA" bandwidth="10.0" latency="0.2"/>
1438 You have at your disposal the following options: bandwidth_file,
1439 latency_file and state_file. The only difference with hosts is that
1440 bandwidth_file and latency_file do not express fraction of available
1441 power but are expressed directly in bytes per seconds and seconds.
1443 \subsubsection faq_platform_multipath How to express multipath routing in platform files?
1445 It is unfortunately impossible to express the fact that there is more
1446 than one routing path between two given hosts. Let's consider the
1447 following platform file:
1450 <route src="A" dst="B">
1453 <route src="B" dst="C">
1456 <route src="A" dst="C">
1461 Although it is perfectly valid, it does not mean that data traveling
1462 from A to C can either go directly (using link 3) or through B (using
1463 links 1 and 2). It simply means that the routing on the graph is not
1464 trivial, and that data do not following the shortest path in number of
1465 hops on this graph. Another way to say it is that there is no implicit
1466 in these routing descriptions. The system will only use the routes you
1467 declare (such as <route src="A" dst="C"><link:ctn
1468 id="3"/></route>), without trying to build new routes by aggregating
1471 You are also free to declare platform where the routing is not
1472 symmetric. For example, add the following to the previous file:
1475 <route src="C" dst="A">
1481 This makes sure that data from C to A go through B where data from A
1482 to C go directly. Don't worry about realism of such settings since
1483 we've seen ways more weird situation in real settings (in fact, that's
1484 the realism of very regular platforms which is questionable, but
1485 that's another story).
1487 \subsubsection faq_flexml_bypassing Bypassing the XML parser with your own C functions
1489 So you want to bypass the XML files parser, uh? Maybe doing some parameter
1490 sweep experiments on your simulations or so? This is possible, and
1491 it's not even really difficult (well. Such a brutal idea could be
1492 harder to implement). Here is how it goes.
1494 For this, you have to first remember that the XML parsing in SimGrid is done
1495 using a tool called FleXML. Given a DTD, this gives a flex-based parser. If
1496 you want to bypass the parser, you need to provide some code mimicking what
1497 it does and replacing it in its interactions with the SURF code. So, let's
1498 have a look at these interactions.
1500 FleXML parser are close to classical SAX parsers. It means that a
1501 well-formed SimGrid platform XML file might result in the following
1504 - start "platform_description" with attribute version="2"
1505 - start "host" with attributes id="host1" power="1.0"
1507 - start "host" with attributes id="host2" power="2.0"
1509 - start "link" with ...
1511 - start "route" with ...
1512 - start "link:ctn" with ...
1515 - end "platform_description"
1517 The communication from the parser to the SURF code uses two means:
1518 Attributes get copied into some global variables, and a surf-provided
1519 function gets called by the parser for each event. For example, the event
1520 - start "host" with attributes id="host1" power="1.0"
1522 let the parser do something roughly equivalent to:
1524 strcpy(A_host_id,"host1");
1529 In SURF, we attach callbacks to the different events by initializing the
1530 pointer functions to some the right surf functions. Since there can be
1531 more than one callback attached to the same event (if more than one
1532 model is in use, for example), they are stored in a dynar. Example in
1533 workstation_ptask_L07.c:
1535 /* Adding callback functions */
1536 surf_parse_reset_parser();
1537 surfxml_add_callback(STag_surfxml_host_cb_list, &parse_cpu_init);
1538 surfxml_add_callback(STag_surfxml_prop_cb_list, &parse_properties);
1539 surfxml_add_callback(STag_surfxml_link_cb_list, &parse_link_init);
1540 surfxml_add_callback(STag_surfxml_route_cb_list, &parse_route_set_endpoints);
1541 surfxml_add_callback(ETag_surfxml_link_c_ctn_cb_list, &parse_route_elem);
1542 surfxml_add_callback(ETag_surfxml_route_cb_list, &parse_route_set_route);
1544 /* Parse the file */
1545 surf_parse_open(file);
1546 xbt_assert1((!surf_parse()), "Parse error in %s", file);
1550 So, to bypass the FleXML parser, you need to write your own version of the
1551 surf_parse function, which should do the following:
1552 - Fill the A_<tag>_<attribute> variables with the wanted values
1553 - Call the corresponding STag_<tag>_fun function to simulate tag start
1554 - Call the corresponding ETag_<tag>_fun function to simulate tag end
1555 - (do the same for the next set of values, and loop)
1557 Then, tell SimGrid that you want to use your own "parser" instead of the stock one:
1559 surf_parse = surf_parse_bypass_environment;
1560 MSG_create_environment(NULL);
1561 surf_parse = surf_parse_bypass_application;
1562 MSG_launch_application(NULL);
1565 A set of macros are provided at the end of
1566 include/surf/surfxml_parse.h to ease the writing of the bypass
1567 functions. An example of this trick is distributed in the file
1568 examples/msg/masterslave/masterslave_bypass.c
1570 \subsection faq_simgrid_configuration Changing SimGrid's behavior
1572 A number of options can be given at runtime to change the default
1573 SimGrid behavior. In particular, you can change the default cpu and
1576 \subsubsection faq_simgrid_configuration_fullduplex Using Fullduplex
1578 Experimental fullduplex support is now available on the svn branch. In order to fullduple to work your platform must have two links for each pair
1579 of interconnected hosts, see an example here:
1581 simgrid_svn_sources/exemples/msg/gtnets/fullduplex-p.xml
1584 Using fullduplex support ongoing and incoming communication flows are
1585 treated independently for most models. The exception is the LV08 model which
1586 adds 0.05 of usage on the opposite direction for each new created flow. This
1587 can be useful to simulate some important TCP phenomena such as ack compression.
1589 Running a fullduplex example:
1591 cd simgrid_svn_sources/exemples/msg/gtnets
1592 ./gtnets fullduplex-p.xml fullduplex-d.xml --cfg=fullduplex:1
1599 \subsubsection faq_simgrid_configuration_gtnets Using GTNetS
1601 It is possible to use a packet-level network simulator
1602 instead of the default flow-based simulation. You may want to use such
1603 an approach if you have doubts about the validity of the default model
1604 or if you want to perform some validation experiments. At the moment,
1605 we support the GTNetS simulator (it is still rather experimental
1606 though, so leave us a message if you play with it).
1610 To enable GTNetS model inside SimGrid it is needed to patch the GTNetS simulator source code
1611 and build/install it from scratch
1614 - <b>Download and enter the recent downloaded GTNetS directory</b>
1617 svn checkout svn://scm.gforge.inria.fr/svn/simgrid/contrib/trunk/GTNetS/
1622 - <b>Use the following commands to unzip and patch GTNetS package to work within SimGrid.</b>
1625 unzip gtnets-current.zip
1626 tar zxvf gtnets-current-patch.tgz
1628 cat ../00*.patch | patch -p1
1631 - <b>OPTIONALLY</b> you can use a patch for itanium 64bit processor family.
1634 cat ../AMD64-FATAL-Removed-DUL_SIZE_DIFF-Added-fPIC-compillin.patch | patch -p1
1637 - <b>Compile GTNetS</b>
1639 Due to portability issues it is possible that GTNetS does not compile in your architecture. The patches furnished in SimGrid SVN repository are intended for use in Linux architecture only. Unfortunately, we do not have the time, the money, neither the manpower to guarantee GTNetS portability. We advice you to use one of GTNetS communication channel to get more help in compiling GTNetS.
1643 ln -sf Makefile.linux Makefile
1649 - <b>NOTE</b> A lot of warnings are expected but the application should compile
1650 just fine. If the makefile insists in compiling some QT libraries
1651 please try a make clean before asking for help.
1654 - <b>To compile optimized version</b>
1661 - <b>Installing GTNetS</b>
1663 It is important to put the full path of your libgtsim-xxxx.so file when creating the symbolic link. Replace < userhome > by some path you have write access to.
1666 ln -sf /<absolute_path>/gtnets_current/libgtsim-debug.so /<userhome>/usr/lib/libgtnets.so
1667 export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/<userhome>/usr/lib/libgtnets.so
1668 mkdir /<userhome>/usr/include/gtnets
1669 cp -fr SRC/*.h /<userhome>/usr/include/gtnets
1673 - <b>Enable GTNetS support in SimGrid</b>
1675 In order to enable gtnets with simgrid you have to give where is gtnets. (path to <gtnets_path>/lib and <gtnets_path>/include)
1678 Since v3.4 (with cmake)
1679 cmake . -Dgtnets_path=/<userhome>/usr
1681 Until v3.4 (with autotools)
1682 ./configure --with-gtnets=/<userhome>/usr
1685 - <b>Once you have followed all the instructions for compiling and
1686 installing successfully you can activate this feature at
1687 runntime with the following options:</b>
1690 Since v3.4 (with cmake)
1695 Until v3.4 (with autotools)
1696 cd simgrid/example/msg/
1702 - <b>Or try the GTNetS model dogbone example with</b>
1705 gtnets/gtnets gtnets/onelink-p.xml gtnets/onelink-d.xml --cfg=network_model:GTNets
1709 A long version of this <a href="http://gforge.inria.fr/docman/view.php/12/6283/GTNetS HowTo.html">HowTo</a> it is available
1712 More about GTNetS simulator at <a href="http://www.ece.gatech.edu/research/labs/MANIACS/GTNetS/index.html">GTNetS Website</a>
1716 The patches provided by us worked successfully with GTNetS found
1717 <a href="http://www.ece.gatech.edu/research/labs/MANIACS/GTNetS/software/gtnets-current.zip">here</a>,
1718 dated from 12th June 2008. Due to the discontinuing development of
1719 GTNetS it is impossible to precise a version number. We STRONGLY recommend you
1720 to download and install the GTNetS version found in SimGrid repository as explained above.
1725 \subsubsection faq_simgrid_configuration_alternate_network Using alternative flow models
1727 The default simgrid network model uses a max-min based approach as
1728 explained in the research report
1729 <a href="ftp://ftp.ens-lyon.fr/pub/LIP/Rapports/RR/RR2002/RR2002-40.ps.gz">A Network Model for Simulation of Grid Application</a>.
1730 Other models have been proposed and implemented since then (see for example
1731 <a href="http://mescal.imag.fr/membres/arnaud.legrand/articles/simutools09.pdf">Accuracy Study and Improvement of Network Simulation in the SimGrid Framework</a>)
1732 and can be activated at runtime. For example:
1734 ./mycode platform.xml deployment.xml --cfg=workstation/model:compound --cfg=network/model:LV08 -cfg=cpu/model:Cas01
1737 Possible models for the network are currently "Constant", "CM02",
1738 "LegrandVelho", "GTNets", Reno", "Reno2", "Vegas". Others will
1739 probably be added in the future and many of the previous ones are
1740 experimental and are likely to disappear without notice... To know the
1741 list of the currently implemented models, you should use the
1742 --help-models command line option.
1745 ./masterslave_forwarder ../small_platform.xml deployment_masterslave.xml --help-models
1746 Long description of the workstation models accepted by this simulator:
1747 CLM03: Default workstation model, using LV08 and CM02 as network and CPU
1748 compound: Workstation model allowing you to use other network and CPU models
1749 ptask_L07: Workstation model with better parallel task modeling
1750 Long description of the CPU models accepted by this simulator:
1751 Cas01_fullupdate: CPU classical model time=size/power
1752 Cas01: Variation of Cas01_fullupdate with partial invalidation optimization of lmm system. Should produce the same values, only faster
1753 CpuTI: Variation of Cas01 with also trace integration. Should produce the same values, only faster if you use availability traces
1754 Long description of the network models accepted by this simulator:
1755 Constant: Simplistic network model where all communication take a constant time (one second)
1756 CM02: Realistic network model with lmm_solve and no correction factors
1757 LV08: Realistic network model with lmm_solve and these correction factors: latency*=10.4, bandwidth*=.92, S=8775
1758 Reno: Model using lagrange_solve instead of lmm_solve (experts only)
1759 Reno2: Model using lagrange_solve instead of lmm_solve (experts only)
1760 Vegas: Model using lagrange_solve instead of lmm_solve (experts only)
1763 \subsection faq_tracing Tracing Simulations for Visualization
1765 The trace visualization is widely used to observe and understand the behavior
1766 of parallel applications and distributed algorithms. Usually, this is done in a
1767 two-step fashion: the user instruments the application and the traces are
1768 analyzed after the end of the execution. The visualization itself can highlights
1769 unexpected behaviors, bottlenecks and sometimes can be used to correct
1770 distributed algorithms. The SimGrid team is currently instrumenting the library
1771 in order to let users trace their simulations and analyze them. This part of the
1772 user manual explains how the tracing-related features can be enabled and used
1773 during the development of simulators using the SimGrid library.
1775 \subsubsection faq_tracing_howitworks How it works
1777 For now, the SimGrid library is instrumented so users can trace the <b>platform
1778 utilization</b> using the MSG interface. This means that the tracing will
1779 register how much power is used for each host and how much bandwidth is used for
1780 each link of the platform. The idea with this type of tracing is to observe the
1781 overall view of resources utilization in the first place, especially the
1782 identification of bottlenecks, load-balancing among hosts, and so on.
1784 The idea of the instrumentation is to classify the MSG tasks by category,
1786 the platform utilization (hosts and links) for each of the categories. For that,
1787 the tracing interface enables the declaration of categories and a function to
1788 mark a task with a previously declared category. <em>The tasks that are not
1789 classified according to a category are not traced</em>.
1791 \subsubsection faq_tracing_enabling Enabling using CMake
1793 With the sources of SimGrid, it is possible to enable the tracing
1794 using the parameter <b>-Dtracing=on</b> when the cmake is executed.
1795 The section \ref faq_tracing_functions describes all the functions available
1796 when this Cmake options is activated. These functions will have no effect
1797 if SimGrid is configured without this option (they are wiped-out by the
1801 $ cmake -Dtracing=on .
1805 \subsubsection faq_tracing_functions Tracing Functions
1807 \subsubsubsection Mandatory Functions
1809 \li <b>\c TRACE_start ()</b>: This is the first function to
1810 be called. It returns 0 if everything was properly initialized, 1 otherwise.
1811 All trace functions called before TRACE_start do nothing.
1813 \li <b>\c TRACE_category (const char *category)</b>: This function should be used
1814 to define a user category. The category can be used to differentiate the tasks
1815 that are created during the simulation (for example, tasks from server1,
1816 server2, or request tasks, computation tasks, communication tasks).
1817 All resource utilization (host power and link bandwidth) will be
1818 classified according to the task category. Tasks that do not belong to a
1819 category are not traced.
1821 \li <b>\c TRACE_msg_set_task_category (m_task_t task, const char *category)</b>:
1822 This function should be called after the creation of a task, to define the
1823 category of that task. The first parameter \c task must contain a task that was
1824 created with the function \c MSG_task_create. The second parameter
1825 \c category must contain a category that was previously defined by the function
1828 \li <b>\c TRACE_end ()</b>: This is the last function to be called. It closes
1829 the trace file and stops the tracing of the simulation. All tracing will be
1830 completely disabled after the calling this function. Although we recommend
1831 the use of this function somewhere in the end of program, it can be used
1832 anywhere in the code. This function returns 0 if everything is ok, 1 otherwise.
1834 \subsubsubsection Optional Functions
1836 \li <b>\c TRACE_host_variable_declare (const char *variable)</b>:
1837 Declare a user variable that will be associated to hosts. A variable can
1838 be used to trace user variables such as the number of tasks in a server,
1839 the number of clients in an application, and so on.
1841 \li <b>\c TRACE_host_variable_[set|add|sub] (const char *variable, double
1843 Set the value of a given user variable. It is important to remind that
1844 the value of this variable is always associated to the host. The host
1845 that will be used when these functions are called is the one returned by
1846 the function \c MSG_host_self().
1848 \subsubsection faq_tracing_options Tracing configuration Options
1850 These are the options accepted by the tracing system of SimGrid:
1852 \li <b>\c tracing/filename</b>: use this to specify the name of the trace file
1853 that will be created during the simulation. For example, after the binary
1854 of your simulator, you can pass as parameter this:
1856 --cfg=tracing/filename:mytracefile.trace
1858 in order to trace the behavior of the simulation in a file with the name
1861 \li <b>\c tracing/platform</b>: use this to activate the tracing of the
1862 platform. For example, you can pass as parameter to your simulator:
1864 --cfg=tracing/platform:1
1866 to trace the platform utilization by the categories you declared in your
1867 simulator. By default, this options is set to 0.
1869 \subsubsection faq_tracing_example Example of Instrumentation
1871 A simplified example using the tracing mandatory functions.
1874 int main (int argc, char **argv)
1876 MSG_global_init (&argc, &argv);
1878 //note that TRACE_start must be called after MSG_global_init
1880 TRACE_category ("request");
1881 TRACE_category ("computation");
1882 TRACE_category ("finalize");
1884 //(... after deployment ...)
1886 m_task_t req1 = MSG_task_create("1st_request_task", 10, 10, NULL);
1887 m_task_t req2 = MSG_task_create("2nd_request_task", 10, 10, NULL);
1888 m_task_t req3 = MSG_task_create("3rd_request_task", 10, 10, NULL);
1889 m_task_t req4 = MSG_task_create("4th_request_task", 10, 10, NULL);
1890 TRACE_msg_set_task_category (req1, "request");
1891 TRACE_msg_set_task_category (req2, "request");
1892 TRACE_msg_set_task_category (req3, "request");
1893 TRACE_msg_set_task_category (req4, "request");
1895 m_task_t comp = MSG_task_create ("comp_task", 100, 100, NULL);
1896 TRACE_msg_set_task_category (comp, "computation");
1898 m_task_t finalize = MSG_task_create ("finalize", 0, 0, NULL);
1899 TRACE_msg_set_task_category (finalize, "finalize");
1910 \subsubsection faq_tracing_analyzing Analyzing the SimGrid Traces
1912 The SimGrid library, during an instrumented simulation, creates a trace file in
1913 the Paje file format that contains the platform utilization for the simulation
1914 that was executed. The visualization analysis of this file is performed with the
1915 visualization tool <a href="http://triva.gforge.inria.fr">Triva</a>, with
1916 special configurations tunned to SimGrid needs. This part of the documentation
1917 explains how to configure and use Triva to analyse a SimGrid trace file.
1919 - <b>Installing Triva</b>: the tool is available in the INRIAGforge,
1920 at <a href="http://triva.gforge.inria.fr">http://triva.gforge.inria.fr</a>.
1921 Use the following command to get the sources, and then check the file
1922 <i>INSTALL.simplified</i>. This file contains instructions to install
1923 the tool's dependencies in a Ubuntu/Debian Linux.
1925 $ svn checkout svn://scm.gforge.inria.fr/svn/triva
1927 $ cat INSTALL.simplified
1930 - <b>Executing Triva</b>: a binary called <i>Triva</i> is available after the
1931 installation (you can execute it passing <em>--help</em> to check its
1932 options). If the triva binary is not available after following the
1933 installation instructions, you may want to execute the following command to
1934 initialize the GNUstep environment variables (note that the location of the
1935 <i>GNUstep.sh</i> file may vary depending on your GNUstep installation - the
1936 command is known to work in Ubuntu and Debian Linux):
1938 $ source /usr/share/GNUstep/Makefiles/GNUstep.sh
1940 You should be able to see this output after the installation of triva:
1942 $ ./Triva.app/Triva --help
1943 Usage: Triva [OPTION...] TRACEFILE
1944 Trace Analysis through Visualization
1946 You need to use one of the following options:
1947 -g, --graph Graph Analysis
1948 -t, --treemap Treemap Analysis
1950 Other auxiliary options to check the trace file:
1951 -c, --check Check the integrity of trace file
1952 -h, --hierarchy Export the trace type hierarchy
1953 -l, --list List entity types
1955 -?, --help Give this help list
1956 --usage Give a short usage message
1958 Triva expects that the user choose one of the available options
1959 (currently <em>--graph</em> or <em>--treemap</em> for a visualization analysis)
1960 and the trace file from the simulation.
1962 - <b>Understanding Triva - time-slice</b>: the analysis of a trace file using
1963 the tool always takes into account the concept of the <em>time-slice</em>.
1964 This concept means that what is being visualized in the screen is always
1965 calculated considering a specific time frame, with its beggining and end
1966 timestamp. The time-slice is configured by the user and can be changed
1967 dynamically through the window called <em>Time Interval</em> that is opened
1968 whenever a trace file is being analyzed. The next figure depicts the time-slice
1969 configuration window.
1970 In the top of the window, in the space named <i>Trace Time</i>,
1971 the two fields show the beggining of the trace (which usually starts in 0) and
1972 the end (that depends on the time simulated by SimGrid). The middle of the
1973 window, in the square named <i>Time Slice Configuration</i>, contains the
1974 aspects related to the time-slice, including its <i>start</i> and its
1975 <i>size</i>. The gray rectangle in the bottom of this part indicates the
1976 <i>current time-slice</i> that is considered for the drawings. If the checkbox
1977 <i>Update Drawings on Sliders Change</i> is not selected, the button
1978 <i>Apply</i> must be clicked in order to inform triva that the
1979 new time-slice must be considered. The bottom part of the window, in the space
1980 indicated by the square <i>Time Slice Animation</i> can be used to advance
1981 the time-frame automatically. The user configures the amount of time that the
1982 time-frame will forward and how frequent this update will happen. Once this is
1983 configured, the user clicks the <i>Play</i> button in order to see the dynamic
1984 changes on the drawings.
1987 <a href="triva-time_interval.png" border=0><img src="triva-time_interval.png" width="50%" border=0></a>
1990 <b>Remarks:</b> when the trace has too many hosts or links, the computation to
1991 take into account a new time-slice can be expensive. When this happens, the
1992 <i>Frequency</i> parameter, but also updates caused by change on configurations
1993 when the checkbox <i>Update Drawings on Sliders
1994 Change</i> is selected will not be followed.
1996 - <b>Understanding Triva - graph</b>: this part of the documention explains how
1997 to analyze the traces using the graph view of Triva, when the user executes
1998 the tool passing <em>--graph</em> as parameter. Triva opens three windows when
1999 this parameter is used: the <i>Time Interval</i> window (previously described),
2000 the <i>Graph Representation</i> window, and the <em>Graph Configuration</em>
2001 window. The Graph Representation is the window where drawings take place.
2002 Initially, it is completely white waiting for a proper graph configuration input
2003 by the user. We start the description of this type of analysis by describing the
2004 <i>Graph Configuration</i> window (depicted below). By using a particular
2005 configuration, triva
2006 can be used to customize the graph drawing according to
2007 the SimGrid trace that was created with user-specific categories. Before delving
2008 into the details of this customization, let us first explain the major parts of
2009 the graph configuration window. The buttons located in the top-right corner can
2010 be used to delete, copy and create a new configuration. The checkbox in the
2011 top-middle part of the window indicates if the configuration typed in the
2012 textfield is syntactically correct (we are using the non-XML
2013 <a href="http://en.wikipedia.org/wiki/Property_list">Property List Format</a> to
2014 describe the configuration). The pop-up button located on the top-left corner
2015 indicates the selected configuration (the user can have multiple graph
2016 configurations). The bottom-left text field contains the name of the current
2017 configuration (updates on this field must be followed by typing enter on the
2018 keyboard to take into account the name change). The bottom-right <em>Apply</em>
2019 button activates the current configuration, resulting on an update on the graph
2023 <a href="triva-graph_configuration.png" border=0><img src="triva-graph_configuration.png" width="50%" border=0></a>
2026 <b>Basic SimGrid Configuration</b>: The figure shows in the big textfield the
2027 basic configuration that should be used during the analysis of a SimGrid trace
2028 file. The basic logic of the configuration is as follows:
2034 The nodes of the graph will be created based on the <i>node</i> parameter, which
2035 in this case is the different <em>"HOST"</em>s of the platform
2036 used to simulate. The <i>edge</i> parameter indicates that the edges of the
2037 graph will be created based on the <em>"LINK"</em>s of the platform. After the
2038 definition of these two parameters, the configuration must detail how
2039 <em>HOST</em>s and <em>LINK</em>s should be drawn. For that, the configuration
2040 must have an entry for each of the types used. For <em>HOST</em>, as basic
2041 configuration, we have:
2048 The parameter <em>size</em> indicates which variable from the trace file will be
2049 used to define the size of the node HOST in the visualization. If the simulation
2050 was executed with availability traces, the size of the nodes will be changed
2051 according to these traces. The parameter <em>scale</em> indicates if the value
2052 of the variable is <em>global</em> or <em>local</em>. If it is global, the value
2053 will be relative to the power of all other hosts, if it is local, the value will
2054 be relative locally.
2055 For <em>LINK</em> we have:
2065 For the types specified in the <em>edge</em> parameter (such as <em>LINK</em>),
2066 the configuration must contain two additional parameters: <em>src</em> and
2067 <em>dst</em> that are used to properly identify which nodes this edge is
2068 connecting. The values <em>SrcHost</em> and <em>DstHost</em> are always present
2069 in the SimGrid trace file and should not be changed in the configuration. The
2070 parameter <em>size</em> for the LINK, in this case, is configured as the
2071 variable <em>bandwidth</em>, with a <em>global</em> scale. The scale meaning
2072 here is exactly the same used for nodes. The last parameter is the GraphViz
2073 algorithm used to calculate the position of the nodes in the graph
2076 graphviz-algorithm = neato;
2079 <b>Customizing the Graph Representation</b>: triva is capable to handle
2080 a customized graph representation based on the variables present in the trace
2081 file. In the case of SimGrid, every time a category is created for tasks, two
2082 variables in the trace file are defined: one to indicate node utilization (how
2083 much power was used by that task category), and another to indicate link
2084 utilization (how much bandwidth was used by that category). For instance, if the
2085 user declares a category named <i>request</i>, there will be variables named
2086 <b>p</b><i>request</i> and a <b>b</b><i>request</i> (<b>p</b> for power and
2087 <b>b</b> for bandwidth). It is important to notice that the variable
2088 <i>prequest</i> in this case is only available for HOST, and
2089 <i>brequest</i> is only available for LINK. <b>Example</b>: suppose there are
2090 two categories for tasks: request and compute. To create a customized graph
2091 representation with a proportional separation of host and link utilization, use
2092 as configuration for HOST and LINK this:
2101 values = (prequest, pcomputation);
2114 values = (brequest, bcomputation);
2118 Where <i>sep_host</i> contains a composition of type <i>separation</i> where
2119 its max size is the <i>power</i> of the host and the variables <i>prequest</i>
2120 and <i>pcomputation</i> are drawn proportionally to the size of the HOST. And
2121 <i>sep_link</i> is also a separation where max is defined as the
2122 <i>bandwidth</i> of the link, and the variables <i>brequest</i> and
2123 <i>bcomputation</i> are drawn proportionally within a LINK.
2124 <i>This configuration enables the analysis of resource utilization by MSG tasks,
2125 and the identification of load-balancing issues, network bottlenecks, for
2127 <b>Other compositions</b>: besides <i>separation</i>, it is possible to use
2128 other types of compositions, such as gradients, and colors, like this:
2133 values = (numberOfTasks);
2137 values = (is_server);
2140 Where <i>gra_host</i> creates a gradient within a node of the graph, using a
2141 global scale and using as value a variable called <i>numberOfTasks</i>, that
2142 could be declared by the user using the optional tracing functions of SimGrid.
2143 If scale is global, the max and min value for the gradient will be equal to the
2144 max and min numberOfTasks among all hosts, and if scale is local, the max and
2145 min value based on the value of numberOfTasks locally in each host.
2146 And <i>color_host</i> composition draws a square based on a positive value of
2147 the variable <i>is_server</i>, that could also be defined by the user using the
2148 SimGrid tracing functions. \n
2149 <b>The Graph Visualization</b>: The next figure shows a graph visualization of a
2150 given time-slice of the masterslave_forwarder example (present in the SimGrid
2151 sources). The red color indicates tasks from the <i>compute</i> category. This
2152 visualization was generated with the following configuration:
2165 values = (pcompute, pfinalize);
2177 values = (bcompute, bfinalize);
2180 graphviz-algorithm = neato;
2185 <a href="triva-graph_visualization.png" border=0><img src="triva-graph_visualization.png" width="50%" border=0></a>
2189 - <b>Understading Triva - colors</b>: An important issue when using Triva is how
2190 to define colors. To do that, we have to know which variables are defined in
2191 the trace file generated by the SimGrid library. The parameter <em>--list</em>
2192 lists the variables for a given trace file:
2194 $ Triva -l masterslave_forwarder.trace
2212 We can see that HOST has seven variables (from power to pfinalize) and LINK has
2213 four (from bandwidth to bfinalize). To define a red color for the
2214 <i>pcompute</i> and <i>bcompute</i> (which are defined based on user category
2215 <i>compute</i>), execute:
2217 $ defaults write Triva 'pcompute Color' '1 0 0'
2218 $ defaults write Triva 'bcompute Color' '1 0 0'
2220 Where the three numbers in each line are the RGB color with values from 0 to 1.
2222 \subsection faq_modelchecking Model-Checking
2223 \subsubsection faq_modelchecking_howto How to use it
2224 To enable the experimental SimGrid model-checking support the program should
2225 be executed with the command line argument
2229 Properties are expressed as assertions using the function
2231 void MC_assert(int prop);
2234 \subsection faq_binding_lua Lua Binding
2235 Most of Simgrid modules require a good level in C programming ,
2236 since simgrid is used to be as standard C library .
2237 Sometime ( for some reason or another ) developers prefer using some kind of « easy scripts »
2238 (something like … lua ? Ruby ? ...?) or a language easier to code with ( Java ? ) for their works,
2239 which avoid dealing with C errors , and sometime an important gain of time (coding-time?) .
2240 Besides Java Binding, Lua and Ruby bindings are available now( since version 3.4 of Simgrid )
2241 for MSG Module, and we are currenlty working on bindings for other modules .
2244 \subsubsection faq_binding_lua_about What is lua ?
2245 Lua (Moon for portuguese !) is a lightweight, reflective, imperative and functional programming language,
2246 designed as a scripting language with extensible semantics as a primary goal.(see official web site <a href="http://www.lua.org">here</a>)
2247 \subsubsection faq_binding_lua_why Why lua ?
2248 Lua is a fast,portable and powerful script language, quite simple to use for developpers .
2249 it combines procedural features with powerful data description facilities,
2250 by using a simple, yet powerful, mechanism of tables.
2251 Lua has a relatively simple C API compared to other scripting languages,
2252 and accordingly it provides a robust, easy to use it.
2253 \subsubsection faq_binding_lua_simgrid How to use lua in Simgrid ?
2254 Actually , the use of lua in Simgrid is quite simple, you have just to follow the same steps as coding with C in Simgird,
2255 but this time, code with Lua ;) :
2256 - Coding functions coresponding to each process
2257 - loading the platforme/deployment XML file that describe the environment of simulation
2258 - and … Running the Simulation !!!
2260 \dontinclude lua/master_slave.lua
2261 \subsubsection faq_binding_lua_example_master_slave Master/Slave Example
2264 \until end_of_master
2265 we mainly use simgrid.Task.new(task_name,computation_size,communication_size) to create our MSG Task,
2266 then simgrid.Task.send(task,alias) to send it.
2267 we use also simgrid.Task.name(task), to get the task's name .
2271 Here, we could see how we use simgrid.Task.recv(alias) to receive a task with a specific alias,
2272 this function return directly the task recevied .
2274 \li Set Environmenet and run application
2275 \until simgrid.clean()
2277 \subsubsection faq_binding_lua_example_data Exchanging Data
2278 You can also exchange data between Process using lua. for that, you have to deal with lua task as a table,
2279 since lua is based itself on a mechanism of tables,
2280 so you can exchange any kind of data ( tables, matrix, strings … ) between process via tasks.
2284 task = simgrid.Task.new("data_task",task_comp,task_comm);
2285 task['matrix'] = my_matrix;
2286 task['table'] = my_table;
2287 task['message'] = "Hello from (Lua || Simgrid ) !! "
2289 simgrid.Task.send(task,alias)
2291 After creating task, we associate to it various kind of data with a specific key,( string in this case)
2292 to distinguish between data variables. Via this key the receiver could access easily to datas.
2295 \li Receiver processe
2297 task = simgrid.Task.recv(alias);
2298 sender_matrix = task['matrix'];
2299 sender_table = task['table'];
2300 sender_message = task['message']
2303 Note that in lua, both sender and receiver share the same lua task!
2304 So that the receiver could joint data directly on the received task without sending it back.
2305 You can find a complet example ( matrix multiplication case ) in the file example/lua/mult_matrix.lua
2308 \subsubsection faq_binding_lua_example_bypass Bypass XML
2309 maybe you wonder if there is a way to bypass the XML files,
2310 and describe your platform directly from the code, with lua bindings it's Possible !! how ?
2311 We provide some additional (tricky?) functions in lua that allows you to set up your own platform without using the XML files
2312 ( this can be useful for large platforms, so a simple for loop will avoid you to deal with an annoying XML File ;) )
2317 simgrid.Host.new{id="Tremblay",power=98095000};
2318 simgrid.Host.new{id="Jupiter",power=76296000};
2319 simgrid.Host.new{id="Fafard",power=76296000};
2320 simgrid.Host.new{id="Ginette",power=48492000};
2321 simgrid.Host.new{id="Bourassa",power=48492000};
2323 we use simgrid.Host.new{id=id_host,power=power_host} to instanciate our hosts.
2328 simgrid.Link.new{id=i,bandwidth=252750+ i*768,latency=0.000270544+i*0.087}; -- some crazy values ;)
2331 we used simgrid.Link.new{id=link_id,bandwidth=bw,latency=lat} with a simple for loop to create all links we need (much easier than XML hein ?)
2335 -- simgrid.Route.new(src_id,des_id,links_nb,links_list)
2336 simgrid.Route.new("Tremblay","Jupiter",1,{"1"});
2337 simgrid.Route.new("Tremblay","Fafard",6,{"0","1","2","3","4","8"});
2338 simgrid.Route.new("Tremblay","Ginette",3,{"3","4","5"});
2339 simgrid.Route.new("Tremblay","Bourassa",7,{"0","1","3","2","4","6","7"});
2341 simgrid.Route.new("Jupiter","Tremblay",1,{"1"});
2342 simgrid.Route.new("Jupiter","Fafard",7,{"0","1","2","3","4","8","9"});
2343 simgrid.Route.new("Jupiter","Ginette",4,{"3","4","5","9"});
2344 simgrid.Route.new("Jupiter","Bourassa",8,{"0","1","2","3","4","6","7","9"});
2347 for each host you have to specify which route to choose to access to the rest of hosts connected in the grid.
2351 simgrid.register_platform();
2353 Don't forget to register your platform, that SURF callbacks starts their work ;)
2357 simgrid.Host.setFunction("Tremblay","Master",4,{"20","550000000","1000000","4"});
2358 simgrid.Host.setFunction("Bourassa","Slave",1,{"0"});
2359 simgrid.Host.setFunction("Jupiter","Slave",1,{"1"});
2360 simgrid.Host.setFunction("Fafard","Slave",1,{"2"});
2361 simgrid.Host.setFunction("Ginette","Slave",1,{"3"});
2363 you don't need to use a deployment XML file, thanks to simgrid.Host.setFunction(host_id,function,args_number,args_list)
2364 you can associate functions for each host with arguments if needed .
2368 simgrid.register_application();
2370 Yes, Here too you have to resgiter your application before running the simulation.
2372 the full example is distributed in the file examples/lua/master_slave_bypass.lua
2374 \subsection faq_binding_ruby Ruby Binding
2377 \subsubsection faq_binding_ruby_simgrid Use Ruby in Simgrid
2378 Since v3.4, the use of <a href="http://ruby-lang.org">ruby</a> in simgrid is available for the MSG Module.
2379 you can find almost all MSG functionalities in Ruby code, that allows you to set up your environment, manage tasks between hosts and run the simulation.
2381 \dontinclude ruby/MasterSlave.rb
2382 \subsubsection faq_binding_ruby_example Master/Slave Ruby Application
2383 for each process method(master and slave in this example), you have to associate a ruby class, that should inherit from <i>MSG::Process</i> ruby class,
2384 with a 'main' function that describe the behaviour of the process during the simulation.
2392 \until end_of_master
2394 the class MSG::Task contains methods that allows the management of the native MSG tasks.
2395 in master ruby code we used :
2396 - <i>MSG::Task.new(task_name,compute_size,communication_size)</i> : to instanciate a new task.
2397 - <i>MSG::Task.send(mailbox)</i> : to send the task via a mailbox alias.
2398 - <i>MSG::Task.name</i> : to get the task's name.
2402 to receive a task, we use the method <i>MSG::Task.receive(mailbox)</i> that return a MSG:Task object (received task).
2407 - <i>MSG.createEnvironment(platform_file)</i> : set up the environment
2408 - <i>MSG.deployApplication(deployment_file)</i> : load the deployment file description.
2409 - <i>MSG.run</i> : run the simulation
2411 \subsubsection faq_binding_ruby_data Exchanging data
2412 ruby bindings provides two ways to exchange data between ruby processes.
2413 \li MSG::Task.join & MSG::Task.data \br
2415 the MSG::Task class contains 2 methods that allows a data exchange between 2 process.
2417 -<i>MSG::Task.join</i> : makes possible to join any kind of ruby data within a task.
2421 myTable <<1<<-2<<45<<67<<87<<76<<89<<56<<78<<3<<-4<<99
2422 # Creates and send Task With the Table inside
2423 task = MSG::Task.new("quicksort_task",taskComputeSize, taskCommunicationSize);
2428 -<i>MSG::Task.data</i> : to access to the data contained into the task.
2431 task = MSG::Task.receive(recv_mailbox.to_s)
2433 quicksort(table,0,table.size-1)
2436 you can find a complet example illustrating the use of those methods in file /example/ruby/Quicksort.rb
2440 another 'object-oriented' way to do it, is to make your own 'task' class that inherit from MSG::Task ,
2441 and contains data you want to deal with, the only 'tricky' thing is that "the initializer" method has no effect !
2443 the use of some getter/setter methods would be the simple way to manage your data :)
2445 class PingPongTask < MSG::Task
2446 # The initialize method has no effect
2456 you can find an example of use in file example/ruby/PingPong.rb
2458 \section faq_troubleshooting Troubleshooting
2460 \subsection faq_trouble_lib_compil SimGrid compilation and installation problems
2462 \subsubsection faq_trouble_lib_config ./configure fails!
2464 We know only one reason for the configure to fail:
2466 - <b>You are using a broken build environment</b>\n
2467 If symptom is that configure complains about gcc not being able to build
2468 executables, you are probably missing the libc6-dev package. Damn Ubuntu.
2470 If you experience other kind of issue, please get in touch with us. We are
2471 always interested in improving our portability to new systems.
2473 \subsubsection faq_trouble_distcheck Dude! "make check" fails on my machine!
2475 Don't assume we never run this target, because we do. Check
2476 http://bob.loria.fr:8010 if you don't believe us.
2478 There is several reasons which may cause the make check to fail on your
2481 - <b>You are using a broken libc (probably concerning the contextes)</b>.\n
2482 The symptom is that the "make check" fails within the examples/msg directory.\n
2483 By default, SimGrid uses something called ucontexts. This is part of the
2484 libc, but it's quite undertested. For example, some (old) versions of the
2485 glibc on alpha do not implement these functions, but provide the stubs
2486 (which return ENOSYS: not implemented). It may fool our detection mechanism
2487 and leads to segfaults. There is not much we can do to fix the bug.
2488 A workaround is to compile with --with-context=pthread to avoid
2489 ucontext completely. You'll be a bit more limited in the number
2490 of simulated processes you can start concurrently, but 5000
2491 processes is still enough for most purposes, isn't it?\n
2492 This limitation is the reason why we insist on using this piece of ...
2493 software even if it's so troublesome.\n
2494 <b>=> use --with-pthread on AMD64 architecture that do not have an
2495 ultra-recent libc.</b>
2497 - <b>There is a bug in SimGrid we aren't aware of</b>.\n
2498 If none of the above apply, please drop us a mail on the mailing list so
2499 that we can check it out. Make sure to read \ref faq_bugrepport
2502 \subsection faq_trouble_compil User code compilation problems
2504 \subsubsection faq_trouble_err_logcat "gcc: _simgrid_this_log_category_does_not_exist__??? undeclared (first use in this function)"
2506 This is because you are using the log mecanism, but you didn't created
2507 any default category in this file. You should refer to \ref XBT_log
2508 for all the details, but you simply forgot to call one of
2509 XBT_LOG_NEW_DEFAULT_CATEGORY() or XBT_LOG_NEW_DEFAULT_SUBCATEGORY().
2511 \subsubsection faq_trouble_pthreadstatic "gcc: undefined reference to pthread_key_create"
2513 This indicates that one of the library SimGrid depends on (libpthread
2514 here) was missing on the linking command line. Dependencies of
2515 libsimgrid are expressed directly in the dynamic library, so it's
2516 quite impossible that you see this message when doing dynamic linking.
2518 If you compile your code statically (and if you use a pthread version
2519 of SimGrid -- see \ref faq_more_processes), you must absolutely
2520 specify <tt>-lpthread</tt> on the linker command line. As usual, this should
2521 come after <tt>-lsimgrid</tt> on this command line.
2523 \subsection faq_trouble_errors Runtime error messages
2525 \subsubsection faq_flexml_limit "surf_parse_lex: Assertion `next limit' failed."
2527 This is because your platform file is too big for the parser.
2529 Actually, the message comes directly from FleXML, the technology on top of
2530 which the parser is built. FleXML has the bad idea of fetching the whole
2531 document in memory before parsing it. And moreover, the memory buffer size
2532 must be determined at compilation time.
2534 We use a value which seems big enough for our need without bloating the
2535 simulators footprints. But of course your mileage may vary. In this case,
2536 just edit src/surf/surfxml.l modify the definition of
2537 FLEXML_BUFFERSTACKSIZE. E.g.
2540 #define FLEXML_BUFFERSTACKSIZE 1000000000
2543 Then recompile and everything should be fine, provided that your version of
2544 Flex is recent enough (>= 2.5.31). If not the compilation process should
2547 A while ago, we worked on FleXML to reduce a bit its memory consumption, but
2548 these issues remain. There is two things we should do:
2550 - use a dynamic buffer instead of a static one so that the only limit
2551 becomes your memory, not a stupid constant fixed at compilation time
2552 (maybe not so difficult).
2553 - change the parser so that it does not need to get the whole file in
2554 memory before parsing
2555 (seems quite difficult, but I'm a complete newbe wrt flex stuff).
2557 These are changes to FleXML itself, not SimGrid. But since we kinda hijacked
2558 the development of FleXML, I can grant you that any patches would be really
2559 welcome and quickly integrated.
2561 <b>Update:</b> A new version of FleXML (1.7) was released. Most of the work
2562 was done by William Dowling, who use it in his own work. The good point is
2563 that it now use a dynamic buffer, and that the memory usage was greatly
2564 improved. The downside is that William also changed some things internally,
2565 and it breaks the hack we devised to bypass the parser, as explained in
2566 \ref faq_flexml_bypassing. Indeed, this is not a classical usage of the
2567 parser, and Will didn't imagine that we may have used (and even documented)
2568 such a crude usage of FleXML. So, we now have to repair the bypassing
2569 functionality to use the lastest FleXML version and fix the memory usage in
2572 \subsubsection faq_trouble_gras_transport GRAS spits networking error messages
2574 Gras, on real platforms, naturally use regular sockets to communicate. They
2575 are deeply hidden in the gras abstraction, but when things go wrong, you may
2576 get some weird error messages. Here are some example, with the probable
2579 - <b>Transport endpoint is not connected</b>: several processes try to open
2580 a server socket on the same port number of the same machine. This is
2581 naturally bad and each process should pick its own port number for this.\n
2582 Maybe, you just have some processes remaining from a previous experiment
2584 Killing them may help, but again if you kill -KILL them, you'll have to
2585 wait for a while: they didn't close there sockets properly and the system
2586 needs a while to notice that this port is free again.
2588 - <b>Socket closed by remote side</b>: if the remote process is not
2589 supposed to close the socket at this point, it may be dead.
2591 - <b>Connection reset by peer</b>: I found this on Internet about this
2592 error. I think it's what's happening here, too:\n
2593 <i>This basically means that a network error occurred while the client was
2594 receiving data from the server. But what is really happening is that the
2595 server actually accepts the connection, processes the request, and sends
2596 a reply to the client. However, when the server closes the socket, the
2597 client believes that the connection has been terminated abnormally
2598 because the socket implementation sends a TCP reset segment telling the
2599 client to throw away the data and report an error.\n
2600 Sometimes, this problem is caused by not properly closing the
2601 input/output streams and the socket connection. Make sure you close the
2602 input/output streams and socket connection properly. If everything is
2603 closed properly, however, and the problem persists, you can work around
2604 it by adding a one-second sleep before closing the streams and the
2605 socket. This technique, however, is not reliable and may not work on all
2607 Since GRAS sockets are closed properly (repeat after me: there is no bug
2608 in GRAS), it is either that you are closing your sockets on server side
2609 before the client get a chance to read them (use gras_os_sleep() to delay
2610 the server), or the server died awfully before the client got the data.
2612 \subsubsection faq_trouble_errors_big_fat_warning I'm told that my XML files are too old.
2614 The format of the XML platform description files is sometimes
2615 improved. For example, we decided to change the units used in SimGrid
2616 from MBytes, MFlops and seconds to Bytes, Flops and seconds to ease
2617 people exchanging small messages. We also reworked the route
2618 descriptions to allow more compact descriptions.
2620 That is why the XML files are versionned using the 'version' attribute
2621 of the root tag. Currently, it should read:
2623 <platform version="2">
2626 If your files are too old, you can use the simgrid_update_xml.pl
2627 script which can be found in the tools directory of the archive.
2629 \subsection faq_trouble_valgrind Valgrind-related and other debugger issues
2631 If you don't, you really should use valgrind to debug your code, it's
2634 \subsubsection faq_trouble_vg_longjmp longjmp madness in valgrind
2636 This is when valgrind starts complaining about longjmp things, just like:
2638 \verbatim ==21434== Conditional jump or move depends on uninitialised value(s)
2639 ==21434== at 0x420DBE5: longjmp (longjmp.c:33)
2641 ==21434== Use of uninitialised value of size 4
2642 ==21434== at 0x420DC3A: __longjmp (__longjmp.S:48)
2645 This is the sign that you didn't used the exception mecanism well. Most
2646 probably, you have a <tt>return;</tt> somewhere within a <tt>TRY{}</tt>
2647 block. This is <b>evil</b>, and you must not do this. Did you read the section
2650 \subsubsection faq_trouble_vg_libc Valgrind spits tons of errors about backtraces!
2652 It may happen that valgrind, the memory debugger beloved by any decent C
2653 programmer, spits tons of warnings like the following :
2654 \verbatim ==8414== Conditional jump or move depends on uninitialised value(s)
2655 ==8414== at 0x400882D: (within /lib/ld-2.3.6.so)
2656 ==8414== by 0x414EDE9: (within /lib/tls/i686/cmov/libc-2.3.6.so)
2657 ==8414== by 0x400B105: (within /lib/ld-2.3.6.so)
2658 ==8414== by 0x414F937: _dl_open (in /lib/tls/i686/cmov/libc-2.3.6.so)
2659 ==8414== by 0x4150F4C: (within /lib/tls/i686/cmov/libc-2.3.6.so)
2660 ==8414== by 0x400B105: (within /lib/ld-2.3.6.so)
2661 ==8414== by 0x415102D: __libc_dlopen_mode (in /lib/tls/i686/cmov/libc-2.3.6.so)
2662 ==8414== by 0x412D6B9: backtrace (in /lib/tls/i686/cmov/libc-2.3.6.so)
2663 ==8414== by 0x8076446: xbt_dictelm_get_ext (dict_elm.c:714)
2664 ==8414== by 0x80764C1: xbt_dictelm_get (dict_elm.c:732)
2665 ==8414== by 0x8079010: xbt_cfg_register (config.c:208)
2666 ==8414== by 0x806821B: MSG_config (msg_config.c:42)
2669 This problem is somewhere in the libc when using the backtraces and there is
2670 very few things we can do ourselves to fix it. Instead, here is how to tell
2671 valgrind to ignore the error. Add the following to your ~/.valgrind.supp (or
2672 create this file on need). Make sure to change the obj line according to
2673 your personnal mileage (change 2.3.6 to the actual version you are using,
2674 which you can retrieve with a simple "ls /lib/ld*.so").
2677 name: Backtrace madness
2679 obj:/lib/ld-2.3.6.so
2684 fun:__libc_dlopen_mode
2687 Then, you have to specify valgrind to use this suppression file by passing
2688 the <tt>--suppressions=$HOME/.valgrind.supp</tt> option on the command line.
2689 You can also add the following to your ~/.bashrc so that it gets passed
2690 automatically. Actually, it passes a bit more options to valgrind, and this
2691 happen to be my personnal settings. Check the valgrind documentation for
2694 \verbatim export VALGRIND_OPTS="--leak-check=yes --leak-resolution=high --num-callers=40 --tool=memcheck --suppressions=$HOME/.valgrind.supp" \endverbatim
2696 \subsubsection faq_trouble_backtraces Truncated backtraces
2698 When debugging SimGrid, it's easier to pass the
2699 --disable-compiler-optimization flag to the configure if valgrind or
2700 gdb get fooled by the optimization done by the compiler. But you
2701 should remove these flag when everything works before going in
2702 production (before launching your 1252135 experiments), or everything
2703 will run only one half of the true SimGrid potential.
2705 \subsection faq_deadlock There is a deadlock in my code!!!
2707 Unfortunately, we cannot debug every code written in SimGrid. We
2708 furthermore believe that the framework provides ways enough
2709 information to debug such informations yourself. If the textual output
2710 is not enough, Make sure to check the \ref faq_visualization FAQ entry to see
2711 how to get a graphical one.
2713 Now, if you come up with a really simple example that deadlocks and
2714 you're absolutely convinced that it should not, you can ask on the
2715 list. Just be aware that you'll be severely punished if the mistake is
2716 on your side... We have plenty of FAQ entries to redact and new
2717 features to implement for the impenitents! ;)
2719 \subsection faq_surf_network_latency I get weird timings when I play with the latencies.
2721 OK, first of all, remember that units should be Bytes, Flops and
2722 Seconds. If you don't use such units, some SimGrid constants (e.g. the
2723 SG_TCP_CTE_GAMMA constant used in most network models) won't have the
2724 right unit and you'll end up with weird results.
2726 Here is what happens with a single transfer of size L on a link
2727 (bw,lat) when nothing else happens.
2730 0-----lat--------------------------------------------------t
2731 |-----|**** real_bw =min(bw,SG_TCP_CTE_GAMMA/(2*lat)) *****|
2734 In more complex situations, this min is the solution of a complex
2735 max-min linear system. Have a look
2736 <a href="http://lists.gforge.inria.fr/pipermail/simgrid-devel/2006-April/thread.html">here</a>
2737 and read the two threads "Bug in SURF?" and "Surf bug not
2738 fixed?". You'll have a few other examples of such computations. You
2739 can also read "A Network Model for Simulation of Grid Application" by
2740 Henri Casanova and Loris Marchal to have all the details. The fact
2741 that the real_bw is smaller than bw is easy to understand. The fact
2742 that real_bw is smaller than SG_TCP_CTE_GAMMA/(2*lat) is due to the
2743 window-based congestion mechanism of TCP. With TCP, you can't exploit
2744 your huge network capacity if you don't have a good round-trip-time
2745 because of the acks...
2747 Anyway, what you get is t=lat + L/min(bw,SG_TCP_CTE_GAMMA/(2*lat)).
2749 * if I you set (bw,lat)=(100 000 000, 0.00001), you get t = 1.00001 (you fully
2751 * if I you set (bw,lat)=(100 000 000, 0.0001), you get t = 1.0001 (you're on the
2753 * if I you set (bw,lat)=(100 000 000, 0.001), you get t = 10.001 (ouch!)
2755 This bound on the effective bandwidth of a flow is not the only thing
2756 that may make your result be unexpected. For example, two flows
2757 competing on a saturated link receive an amount of bandwidth inversely
2758 proportional to their round trip time.
2760 \subsection faq_bugrepport So I've found a bug in SimGrid. How to report it?
2762 We do our best to make sure to hammer away any bugs of SimGrid, but this is
2763 still an academic project so please be patient if/when you find bugs in it.
2764 If you do, the best solution is to drop an email either on the simgrid-user
2765 or the simgrid-devel mailing list and explain us about the issue. You can
2766 also decide to open a formal bug report using the
2767 <a href="https://gforge.inria.fr/tracker/?atid=165&group_id=12&func=browse">relevant
2768 interface</a>. You need to login on the server to get the ability to submit
2771 We will do our best to solve any problem repported, but you need to help us
2772 finding the issue. Just telling "it segfault" isn't enough. Telling "It
2773 segfaults when running the attached simulator" doesn't really help either.
2774 You may find the following article interesting to see how to repport
2775 informative bug repports:
2776 http://www.chiark.greenend.org.uk/~sgtatham/bugs.html (it is not SimGrid
2777 specific at all, but it's full of good advices).
2779 \author Arnaud Legrand (arnaud.legrand::imag.fr)
2780 \author Martin Quinson (martin.quinson::loria.fr)
2785 ******************************************************************
2786 * OLD CRUFT NOT USED ANYMORE *
2787 ******************************************************************
2790 \subsection faq_crosscompile Cross-compiling a Windows DLL of SimGrid from linux
2792 At the moment, we do not distribute Windows pre-compiled version of SimGrid
2793 because the support for this platform is still experimental. We know that
2794 some parts of the GRAS environment do not work, and we think that the others
2795 environments (MSG and SD) have good chances to work, but we didn't test
2796 ourselves. This section explains how we generate the SimGrid DLL so that you
2797 can build it for yourself. First of all, you need to have a version more
2798 recent than 3.1 (ie, a SVN version as time of writting).
2800 In order to cross-compile the package to windows from linux, you need to
2801 install mingw32 (minimalist gnu win32). On Debian, you can do so by
2802 installing the packages mingw32 (compiler), mingw32-binutils (linker and
2803 so), mingw32-runtime.
2805 You can use the VPATH support of configure to compile at the same time for
2806 linux and windows without dupplicating the source nor cleaning the tree
2807 between each. Just run bootstrap (if you use the SVN) to run the autotools.
2808 Then, create a linux and a win directories. Then, type:
2809 \verbatim cd linux; ../configure --srcdir=.. <usual configure flags>; make; cd ..
2810 cd win; ../configure --srcdir=.. --host=i586-mingw32msvc <flags>; make; cd ..
2812 The trick to VPATH builds is to call configure from another directory,
2813 passing it an extra --srcdir argument to tell it where all the sources are.
2814 It will understand you want to use VPATH. Then, the trick to cross-compile
2815 is simply to add a --host argument specifying the target you want to build
2816 for. The i586-mingw32msvc string is what you have to pass to use the mingw32
2817 environment as distributed in Debian.
2819 After that, you can run all make targets from both directories, and test
2820 easily that what you change for one arch does not break the other one.
2822 It is possible that this VPATH build thing breaks from time to time in the
2823 SVN since it's quite fragile, but it's granted to work in any released
2824 version. If you experience problems, drop us a mail.
2826 Another possible source of issue is that at the moment, building the
2827 examples request to use the gras_stub_generator tool, which is a compiled
2828 program, not a script. In cross-compilation, you need to cross-execute with
2829 wine for example, which is not really pleasant. We are working on this, but
2830 in the meanwhile, simply don't build the examples in cross-compilation
2831 (<tt>cd src</tt> before running make).
2833 Program (cross-)compiled with mingw32 do request an extra DLL at run-time to be
2834 usable. For example, if you want to test your build with wine, you should do
2835 the following to put this library where wine looks for DLLs.
2837 cp /usr/share/doc/mingw32-runtime/mingwm10.dll.gz ~/.wine/c/windows/system/
2838 gunzip ~/.wine/c/windows/system/mingwm10.dll.gz
2841 The DLL is built in src/.libs, and installed in the <i>prefix</i>/bin directory
2842 when you run make install.
2844 If you want to use it in a native project on windows, you need to use
2845 simgrid.dll and mingwm10.dll. For each DLL, you need to build .def file
2846 under linux (listing the defined symbols), and convert it into a .lib file
2847 under windows (specifying this in a way that windows compilers like). To
2848 generate the def files, run (under linux):
2849 \verbatim echo "LIBRARY libsimgrid-0.dll" > simgrid.def
2850 echo EXPORTS >> simgrid.def
2851 nm libsimgrid-0.dll | grep ' T _' | sed 's/.* T _//' >> simgrid.def
2852 nm libsimgrid-0.dll | grep ' D _' | sed 's/.* D _//' | sed 's/$/ DATA/' >> simgrid.def
2854 echo "LIBRARY mingwm10.dll" > mingwm10.def
2855 echo EXPORTS >> mingwm10.def
2856 nm mingwm10.dll | grep ' T _' | sed 's/.* T _//' >> mingwm10.def
2857 nm mingwm10.dll | grep ' D _' | sed 's/.* D _//' | sed 's/$/ DATA/' >> mingwm10.def
2860 To create the import .lib files, use the <tt>lib</tt> windows tool (from
2861 MSVC) the following way to produce simgrid.lib and mingwm10.lib
2862 \verbatim lib /def:simgrid.def
2863 lib /def:mingwm10.def
2866 If you happen to use Borland C Builder, the right command line is the
2867 following (note that you don't need any file.def to get this working).
2868 \verbatim implib simgrid.lib libsimgrid-0.dll
2869 implib mingwm10.lib mingwm10.dll
2872 Then, set the following parameters in Visual C++ 2005:
2873 Linker -> Input -> Additional dependencies = simgrid.lib mingwm10.lib
2875 Just in case you wonder how to generate a DLL from libtool in another
2876 project, we added -no-undefined to any lib*_la_LDFLAGS variables so that
2877 libtool accepts to generate a dynamic library under windows. Then, to make
2878 it true, we pass any dependencies (such as -lws2 under windows or -lpthread
2879 on need) on the linking line. Passing such deps is a good idea anyway so
2880 that they get noted in the library itself, avoiding the users to know about
2881 our dependencies and put them manually on their compilation line. Then we
2882 added the AC_LIBTOOL_WIN32_DLL macro just before AC_PROG_LIBTOOL in the
2883 configure.ac. It means that we exported any symbols which need to be.
2884 Nowadays, functions get automatically exported, so we don't need to load our
2885 header files with tons of __declspec(dllexport) cruft. We only need to do so
2886 for data, but there is no public data in SimGrid so we are good.