1 This file follows the Doxygen syntax to be included in the
2 documentation, but it should remain readable directly.
5 @defgroup MSG_examples MSG examples
7 @brief Find the MSG example fitting your needs from the extensive set provided in the archive.
13 - @ref msg_ex_tracing_user_variables
18 - @ref msg_ex_full_apps
21 @section msg_ex_basic Basic examples and features
23 - <b>Ping Pong</b>: @ref examples/msg/app-pingpong/app-pingpong.c\n
24 It's hard to think of a simpler example: it is just sending one
25 message back and forth.
26 The tesh file laying in the directory show how to start the
27 simulator binary, enlighting how to pass options to the simulators
28 (as detailed in Section \ref options).
31 @ref examples/msg/app-token-ring/app-token-ring.c\n
32 Classical communication pattern, where a token is exchanged
33 along a ring to reach every participant.
34 The tesh file laying in the directory shows how to run the same
35 example on different virtual platforms.
37 - <b>Master Workers</b>.
38 @ref examples/msg/app-masterworker/app-masterworker.c\n
39 Another good old example, where one Master process has a bunch of
40 task to dispatch to a set of several Worker processes. It is fully
41 commented in @ref MSG_ex_master_worker.
43 @section msg_ex_async Asynchronous communications
45 In addition to the fully documented example of @ref
46 MSG_ex_asynchronous_communications, there are several other examples
47 shipped in the archive:
49 - <b>Basic asynchronous communications</b>.
50 @ref examples/msg/async-wait/async-wait.c \n
51 Illustrates how to have non-blocking communications, that are
52 communications running in the background leaving the process free
53 to do something else during their completion. The main functions
54 involved are @ref MSG_task_isend, @ref MSG_task_irecv, and @ref
57 - <b>Waiting for all communications in a set</b>.
58 @ref examples/msg/async-waitall/async-waitall.c\n
59 The @ref MSG_comm_waitall function is useful when you want to block
60 until all activities in a given set have completed.
62 - <b>Waiting for the first completed communication in a set</b>.
63 @ref examples/msg/async-waitall/async-waitany.c\n
64 The @ref MSG_comm_waitany function is useful when you want to block
65 until one activity of the set completes, no matter which terminates
68 @section msg_ex_process Acting on Processes
70 - <b>Suspend and Resume processes</b>.
71 @ref examples/msg/process-suspend/process-suspend.c \n
72 Processes can be suspended and resumed during their executions
73 thanks to the @ref MSG_process_suspend and @ref MSG_process_resume functions.
75 - <b>Kill processes</b>.
76 @ref examples/msg/process-kill/process-kill.c \n
77 Processes can forcefully stop other processes with the @ref MSG_process_kill function.
79 - <b>Migrating processes</b>.
80 @ref examples/msg/process-migration/process-migration.c \n
81 Processes can move or be moved from a host to another with the @ref MSG_process_migrate function.
83 - <b>Controling the process life cycle from the XML</b>.
84 @ref examples/msg/process-startkilltime/process-startkilltime.c \n
85 You can specify a start time and a kill time in the deployment
86 file. See all *_d.xml files in this directory.
88 TODO: add an example using @ref MSG_process_create()
90 @section msg_ex_tracing Tracing and visualization features
92 Tracing can be activated by various configuration options which
93 are illustrated in these example. See also the
94 @ref tracing_tracing_options "full list of options related to tracing".
96 - <b>Basic example</b>. @ref examples/msg/trace-simple/trace-simple.c \n
97 In this very simple program, each process creates, executes,
98 and destroy a task. Recommanded options:
99 @verbatim --cfg=tracing:yes --cfg=tracing/uncategorized:yes @endverbatim
101 - <b>Platform tracing</b>.
102 @ref examples/msg/trace-platform/trace-platform.c \n
103 This program is a toy example just loading the platform, so that
104 you can play with the platform visualization. Recommanded options:
105 @verbatim --cfg=tracing:yes --cfg=tracing/categorized:yes
108 - <b>Setting Categories</b>.
109 @ref examples/msg/trace-categories/trace-categories.c \n
110 This example declares several tracing categories
111 to that are used to classify its tasks. When the program is executed,
112 the tracing mechanism registers the resource utilization of hosts
113 and links according to these categories. Recommanded options:
114 @verbatim --cfg=tracing:yes --cfg=tracing/categorized:yes --cfg=tracing/uncategorized:yes --cfg=viva/categorized:viva_cat.plist --cfg=viva/uncategorized:viva_uncat.plist
117 - <b>Master Workers tracing</b>.
118 @ref examples/msg/trace-masterworker/trace-masterworker.c \n
119 This is an augmented version of our basic master/worker example
120 using several tracing features. It traces resource usage, sorted
121 out in several categories; Trace marks and user variables are also
122 used. Recommanded options:
123 @verbatim --cfg=tracing/categorized:yes --cfg=tracing/uncategorized:yes --cfg=viva/categorized:viva_cat.plist --cfg=viva/uncategorized:viva_uncat.plist
126 - <b>Process migration tracing</b>.
127 @ref examples/msg/trace-process-migration/trace-process-migration.c \n
128 This version is enhanced so that the process migrations can be
129 displayed as arrows in a Gantt-chart visualization. Recommanded
130 options to that extend:
131 @verbatim -cfg=tracing:yes --cfg=tracing/msg/process:yes
134 TODO: These tracing examples should be integrated in the examples to
135 not dupplicate the C files. A full command line to see the result in
136 the right tool (viva/vite/FrameSoc) should be given along with some
139 @subsection msg_ex_tracing_user_variables Tracing user variables
141 You can also attach your own variables to a any resource described in
142 the platform file. The following examples illustrate this feature.
143 They have to be run with the following options:
144 @verbatim --cfg=tracing:yes --cfg=tracing/platform:yes
147 - <b>Attaching variables to Hosts</b>.
148 @ref examples/msg/trace-user-variables/trace-user-variables.c
150 - <b>Attaching variables to Links</b>.
151 @ref examples/msg/trace-link-user-variables/trace-link-user-variables.c \n
152 The tricky part is that you have to know the name of the link you
153 want to enhance with a variable.
155 - <b>Attaching variables to network Routes</b>
156 @ref examples/msg/trace-link-srcdst-user-variables/trace-link-srcdst-user-variables.c \n
157 It is often easier to update a given variable for all links of a
158 given network path (identified by its source and destination
159 hosts) instead of knowing the name of each specific link.
161 TODO: rename trace-user-variables to trace-host-user-variables
163 TODO: rename trace-link-srcdst-user-variables to trace-route-user-variables
165 @section msg_ex_models Models-related examples
167 @subsection msg_ex_ns3 NS3 as a SimGrid Network Model
169 This example demonstrates how to use the bindings to the Network
170 Simulator, as explained in @ref pls. The most
171 interesting is probably not the C files since they are unchanged from
172 the other simulations, but the associated files, such as the platform
173 file to see how to declare a platform to be used with the PLS bindings
174 of SimGrid and the tesh file to see how to actually start a simulation
177 - @ref examples/msg/network-ns3/network-ns3.c. Simple ping-pong using
178 ns-3 instead of the SimGrid network models.
180 TODO: merge the C files
182 TODO: show the XML files instead if it's what is interesting. On a "XML example files" page that does not exist yet.
184 @subsection msg_ex_io Simulating disks and files
186 The examples of this section demonstrate how to interact with the
189 - <b>Basic example</b>.
190 @ref examples/msg/io-storage/io-storage.c \n
191 All main storage and file functions are demoed.
193 - <b>File Management</b>. @ref examples/msg/io-file/io-file.c \n
194 This example illustrates the use of operations on file
195 (@ref MSG_file_open, @ref MSG_file_read, @ref MSG_file_write,
196 or @ref MSG_file_close).
198 - <b>Remote I/O</b>. @ref examples/msg/io-remote/io-remote.c \n
199 I/O operations can also be done in a remote, i.e. when the
200 accessed disk is not mounted on the caller's host.
202 - @ref examples/msg/actions-comm/actions-comm.c \n
203 - @ref examples/msg/actions-storage/actions-storage.c \n
204 - @ref examples/msg/app-pmm/app-pmm.c \n
205 - @ref examples/msg/dht-chord \n
206 - @ref examples/msg/task-priority/task-priority.c \n
207 - @ref examples/msg/properties/properties.c \n
211 As a human, you can stop reading at this point. The rest is garbage:
213 Every example must be listed in the following, but it's not possible
214 to move this content upper as each @example directive seems to eat the
215 next doxygen commands (and the content is placed at the top of the
220 @defgroup MSG_ex_examples ignored
221 @example examples/msg/app-pingpong/app-pingpong.c
222 @example examples/msg/app-token-ring/app-token-ring.c
223 @example examples/msg/app-masterworker/app-masterworker.c
225 @example examples/msg/async-wait/async-wait.c
226 @example examples/msg/async-waitall/async-waitall.c
227 @example examples/msg/async-waitall/async-waitany.c
229 @example examples/msg/process-suspend/process-suspend.c
230 @example examples/msg/process-kill/process-kill.c
231 @example examples/msg/process-migration/process-migration.c
232 @example examples/msg/process-startkilltime/process-startkilltime.c
234 @example examples/msg/trace-simple/trace-simple.c
235 @example examples/msg/trace-platform/trace-platform.c
236 @example examples/msg/trace-categories/trace-categories.c
237 @example examples/msg/trace-masterworker/trace-masterworker.c
238 @example examples/msg/trace-process-migration/trace-process-migration.c
239 @example examples/msg/trace-user-variables/trace-user-variables.c
240 @example examples/msg/trace-link-user-variables/trace-link-user-variables.c
241 @example examples/msg/trace-link-srcdst-user-variables/trace-link-srcdst-user-variables.c
243 @example examples/msg/network-ns3/network-ns3.c
245 @example examples/msg/io-storage/io-storage.c
246 @example examples/msg/io-file/io-file.c
247 @example examples/msg/io-remote/io-remote.c
248 @example examples/msg/actions-comm/actions-comm.c
249 @example examples/msg/actions-storage/actions-storage.c
250 @example examples/msg/app-pmm/app-pmm.c
251 @example examples/msg/dht-chord
252 @example examples/msg/task-priority/task-priority.c
253 @example examples/msg/properties/properties.c
257 Basic examples and features
258 ===========================
260 * properties/msg_prop.c Attaching arbitrary information to host,
261 processes and such, and retrieving them with
262 MSG_host_get_properties(), MSG_host_get_property_value(),
263 MSG_process_get_properties() and MSG_process_get_property_value().
264 Also make sure to read the platform and deployment XML files to see
265 how to declare these data.
267 * parallel_task/parallel_task.c: Demonstrates the use of
268 MSG_parallel_task_create(), to create special tasks that run on
269 several hosts at the same time. The resulting simulations are very
270 close to what can be achieved in SimDag, but still allows to use
271 the other features of MSG (it'd be cool to be able to mix
272 interfaces, but it's not possible ATM).
274 * priority/priority.c: Demonstrates the use of
275 MSG_task_set_priority() to change the computation priority of a
278 Trace driven simulations
279 ========================
281 The actions/actions.c example demonstrates how to run trace-driven
282 simulations. It is very handy when you want to test an algorithm or
283 protocol that does nothing unless it receives some events from
284 outside. For example, a P2P protocol reacts to requests from the user,
285 but does nothing if there is no such event.
287 In such situations, SimGrid allows to write your protocol in your C
288 file, and the events to react to in a separate text file. Declare a
289 function handling each of the events that you want to accept in your
290 trace files, register them using MSG_action_register in your main, and
291 then use MSG_action_trace_run to launch the simulation. You can either
292 have one trace file containing all your events, or a file per
293 simulated process. Check the tesh files in the example directory for
294 details on how to do it.
296 This example uses this approach to replay MPI-like traces. It comes
297 with a set of event handlers reproducing MPI events. This is somehow
298 similar to SMPI, yet differently implemented. This code should
299 probably be changed to use SMPI internals instead, but wasn't, so far.
301 Examples of full applications
302 =============================
304 * chord/chord.c: Classical Chord P2P protocol This example implements
305 the well known Chord P2P protocol. Its main advantage is that it
306 constitute a fully working non-trivial example. In addition, its
307 implementation is rather efficient, as demonstrated in
308 [57]http://hal.inria.fr/inria-00602216/