8 <object id="TOC" data="graphical-toc.svg" type="image/svg+xml"></object>
10 window.onload=function() { // Wait for the SVG to be loaded before changing it
11 var elem=document.querySelector("#TOC").contentDocument.getElementById("ConfigBox")
12 elem.style="opacity:0.93999999;fill:#ff0000;fill-opacity:0.1;stroke:#000000;stroke-width:0.35277778;stroke-linecap:round;stroke-linejoin:round;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1";
18 A number of options can be given at runtime to change the default
19 SimGrid behavior. For a complete list of all configuration options
20 accepted by the SimGrid version used in your simulator, simply pass
21 the --help configuration flag to your program. If some of the options
22 are not documented on this page, this is a bug that you should please
23 report so that we can fix it. Note that some of the options presented
24 here may not be available in your simulators, depending on the
25 :ref:`compile-time options <install_src_config>` that you used.
27 Setting Configuration Items
28 ---------------------------
30 There is several way to pass configuration options to the simulators.
31 The most common way is to use the ``--cfg`` command line argument. For
32 example, to set the item ``Item`` to the value ``Value``, simply
33 type the following on the command-line:
35 .. code-block:: console
37 $ my_simulator --cfg=Item:Value (other arguments)
39 Several ``--cfg`` command line arguments can naturally be used. If you
40 need to include spaces in the argument, don't forget to quote the
41 argument. You can even escape the included quotes (write ``@'`` for ``'`` if
42 you have your argument between simple quotes).
44 Another solution is to use the ``<config>`` tag in the platform file. The
45 only restriction is that this tag must occur before the first
46 platform element (be it ``<zone>``, ``<cluster>``, ``<peer>`` or whatever).
47 The ``<config>`` tag takes an ``id`` attribute, but it is currently
48 ignored so you don't really need to pass it. The important part is that
49 within that tag, you can pass one or several ``<prop>`` tags to specify
50 the configuration to use. For example, setting ``Item`` to ``Value``
51 can be done by adding the following to the beginning of your platform
57 <prop id="Item" value="Value"/>
60 A last solution is to pass your configuration directly in your program
61 with :cpp:func:`simgrid::s4u::Engine::set_config` or :cpp:func:`MSG_config`.
65 #include <simgrid/s4u.hpp>
67 int main(int argc, char *argv[]) {
68 simgrid::s4u::Engine e(&argc, argv);
70 simgrid::s4u::Engine::set_config("Item:Value");
77 Existing Configuration Items
78 ----------------------------
81 The full list can be retrieved by passing ``--help`` and
82 ``--help-cfg`` to an executable that uses SimGrid. Try passing
83 ``help`` as a value to get the list of values accepted by a given
84 option. For example, ``--cfg=plugin:help`` will give you the list
85 of plugins available in your installation of SimGrid.
87 - **bmf/max-iterations:** :ref:`cfg=bmf/max-iterations`
88 - **bmf/precision:** :ref:`cfg=bmf/precision`
90 - **contexts/factory:** :ref:`cfg=contexts/factory`
91 - **contexts/guard-size:** :ref:`cfg=contexts/guard-size`
92 - **contexts/nthreads:** :ref:`cfg=contexts/nthreads`
93 - **contexts/stack-size:** :ref:`cfg=contexts/stack-size`
94 - **contexts/synchro:** :ref:`cfg=contexts/synchro`
96 - **cpu/maxmin-selective-update:** :ref:`Cpu Optimization Level <options_model_optim>`
97 - **cpu/model:** :ref:`options_model_select`
98 - **cpu/optim:** :ref:`Cpu Optimization Level <options_model_optim>`
100 - **debug/breakpoint:** :ref:`cfg=debug/breakpoint`
101 - **debug/clean-atexit:** :ref:`cfg=debug/clean-atexit`
102 - **debug/verbose-exit:** :ref:`cfg=debug/verbose-exit`
104 - **exception/cutpath:** :ref:`cfg=exception/cutpath`
106 - **host/model:** :ref:`options_model_select`
108 - **maxmin/precision:** :ref:`cfg=maxmin/precision`
109 - **maxmin/concurrency-limit:** :ref:`cfg=maxmin/concurrency-limit`
111 - **msg/debug-multiple-use:** :ref:`cfg=msg/debug-multiple-use`
113 - **model-check:** :ref:`options_modelchecking`
114 - **model-check/checkpoint:** :ref:`cfg=model-check/checkpoint`
115 - **model-check/communications-determinism:** :ref:`cfg=model-check/communications-determinism`
116 - **model-check/dot-output:** :ref:`cfg=model-check/dot-output`
117 - **model-check/max-depth:** :ref:`cfg=model-check/max-depth`
118 - **model-check/property:** :ref:`cfg=model-check/property`
119 - **model-check/reduction:** :ref:`cfg=model-check/reduction`
120 - **model-check/replay:** :ref:`cfg=model-check/replay`
121 - **model-check/send-determinism:** :ref:`cfg=model-check/send-determinism`
122 - **model-check/setenv:** :ref:`cfg=model-check/setenv`
123 - **model-check/termination:** :ref:`cfg=model-check/termination`
124 - **model-check/timeout:** :ref:`cfg=model-check/timeout`
125 - **model-check/visited:** :ref:`cfg=model-check/visited`
127 - **network/bandwidth-factor:** :ref:`cfg=network/bandwidth-factor`
128 - **network/crosstraffic:** :ref:`cfg=network/crosstraffic`
129 - **network/latency-factor:** :ref:`cfg=network/latency-factor`
130 - **network/loopback-lat:** :ref:`cfg=network/loopback`
131 - **network/loopback-bw:** :ref:`cfg=network/loopback`
132 - **network/maxmin-selective-update:** :ref:`Network Optimization Level <options_model_optim>`
133 - **network/model:** :ref:`options_model_select`
134 - **network/optim:** :ref:`Network Optimization Level <options_model_optim>`
135 - **network/TCP-gamma:** :ref:`cfg=network/TCP-gamma`
136 - **network/weight-S:** :ref:`cfg=network/weight-S`
138 - **ns3/TcpModel:** :ref:`options_pls`
139 - **ns3/seed:** :ref:`options_pls`
140 - **path:** :ref:`cfg=path`
141 - **plugin:** :ref:`cfg=plugin`
143 - **storage/max_file_descriptors:** :ref:`cfg=storage/max_file_descriptors`
145 - **surf/precision:** :ref:`cfg=surf/precision`
147 - **For collective operations of SMPI,** please refer to Section :ref:`cfg=smpi/coll-selector`
148 - **smpi/auto-shared-malloc-thresh:** :ref:`cfg=smpi/auto-shared-malloc-thresh`
149 - **smpi/async-small-thresh:** :ref:`cfg=smpi/async-small-thresh`
150 - **smpi/barrier-finalization:** :ref:`cfg=smpi/barrier-finalization`
151 - **smpi/barrier-collectives:** :ref:`cfg=smpi/barrier-collectives`
152 - **smpi/buffering:** :ref:`cfg=smpi/buffering`
153 - **smpi/coll-selector:** :ref:`cfg=smpi/coll-selector`
154 - **smpi/comp-adjustment-file:** :ref:`cfg=smpi/comp-adjustment-file`
155 - **smpi/cpu-threshold:** :ref:`cfg=smpi/cpu-threshold`
156 - **smpi/display-allocs:** :ref:`cfg=smpi/display-allocs`
157 - **smpi/display-timing:** :ref:`cfg=smpi/display-timing`
158 - **smpi/errors-are-fatal:** :ref:`cfg=smpi/errors-are-fatal`
159 - **smpi/grow-injected-times:** :ref:`cfg=smpi/grow-injected-times`
160 - **smpi/host-speed:** :ref:`cfg=smpi/host-speed`
161 - **smpi/IB-penalty-factors:** :ref:`cfg=smpi/IB-penalty-factors`
162 - **smpi/iprobe:** :ref:`cfg=smpi/iprobe`
163 - **smpi/iprobe-cpu-usage:** :ref:`cfg=smpi/iprobe-cpu-usage`
164 - **smpi/init:** :ref:`cfg=smpi/init`
165 - **smpi/keep-temps:** :ref:`cfg=smpi/keep-temps`
166 - **smpi/ois:** :ref:`cfg=smpi/ois`
167 - **smpi/or:** :ref:`cfg=smpi/or`
168 - **smpi/os:** :ref:`cfg=smpi/os`
169 - **smpi/papi-events:** :ref:`cfg=smpi/papi-events`
170 - **smpi/pedantic:** :ref:`cfg=smpi/pedantic`
171 - **smpi/privatization:** :ref:`cfg=smpi/privatization`
172 - **smpi/privatize-libs:** :ref:`cfg=smpi/privatize-libs`
173 - **smpi/send-is-detached-thresh:** :ref:`cfg=smpi/send-is-detached-thresh`
174 - **smpi/shared-malloc:** :ref:`cfg=smpi/shared-malloc`
175 - **smpi/shared-malloc-hugepage:** :ref:`cfg=smpi/shared-malloc-hugepage`
176 - **smpi/simulate-computation:** :ref:`cfg=smpi/simulate-computation`
177 - **smpi/test:** :ref:`cfg=smpi/test`
178 - **smpi/wtime:** :ref:`cfg=smpi/wtime`
179 - **smpi/list-leaks** :ref:`cfg=smpi/list-leaks`
181 - **Tracing configuration options** can be found in Section :ref:`tracing_tracing_options`
183 - **storage/model:** :ref:`options_model_select`
185 - **vm/model:** :ref:`options_model_select`
189 Configuring the Platform Models
190 -------------------------------
192 .. _options_model_select:
194 Choosing the Platform Models
195 ............................
197 SimGrid comes with several network, CPU and disk models built in,
198 and you can change the used model at runtime by changing the passed
199 configuration. The three main configuration items are given below.
200 For each of these items, passing the special ``help`` value gives you
201 a short description of all possible values (for example,
202 ``--cfg=network/model:help`` will present all provided network
203 models). Also, ``--help-models`` should provide information about all
204 models for all existing resources.
206 - ``network/model``: specify the used network model. Possible values:
208 - **LV08 (default one):** Realistic network analytic model
209 (slow-start modeled by multiplying latency by 13.01, bandwidth by
210 .97; bottleneck sharing uses a payload of S=20537 for evaluating
211 RTT). Described in `Accuracy Study and Improvement of Network
212 Simulation in the SimGrid Framework
213 <http://mescal.imag.fr/membres/arnaud.legrand/articles/simutools09.pdf>`_.
214 - **Constant:** Simplistic network model where all communication
215 take a constant time (one second). This model provides the lowest
216 realism, but is (marginally) faster.
217 - **SMPI:** Realistic network model specifically tailored for HPC
218 settings (accurate modeling of slow start with correction factors on
219 three intervals: < 1KiB, < 64 KiB, >= 64 KiB). This model can be
220 :ref:`further configured <options_model_network>`.
221 - **IB:** Realistic network model specifically tailored for HPC
222 settings with InfiniBand networks (accurate modeling contention
223 behavior, based on the model explained in `this PhD work
224 <http://mescal.imag.fr/membres/jean-marc.vincent/index.html/PhD/Vienne.pdf>`_.
225 This model can be :ref:`further configured <options_model_network>`.
226 - **CM02:** Legacy network analytic model. Very similar to LV08, but
227 without corrective factors. The timings of small messages are thus
228 poorly modeled. This model is described in `A Network Model for
229 Simulation of Grid Application
230 <https://hal.inria.fr/inria-00071989/document>`_.
231 - **ns-3** (only available if you compiled SimGrid accordingly):
232 Use the packet-level network
233 simulators as network models (see :ref:`model_ns3`).
234 This model can be :ref:`further configured <options_pls>`.
236 - ``cpu/model``: specify the used CPU model. We have only one model
239 - **Cas01:** Simplistic CPU model (time=size/speed)
241 - ``host/model``: The host concept is the aggregation of a CPU with a
242 network card. Three models exists, but actually, only 2 of them are
243 interesting. The "compound" one is simply due to the way our
244 internal code is organized, and can easily be ignored. So at the
245 end, you have two host models: The default one allows aggregation of
246 an existing CPU model with an existing network model, but does not
247 allow parallel tasks because these beasts need some collaboration
248 between the network and CPU model.
250 - **default:** Default host model. Currently, CPU:Cas01 and
251 network:LV08 (with cross traffic enabled)
252 - **compound:** Host model that is automatically chosen if
253 you change the network and CPU models
254 - **ptask_L07:** Host model somehow similar to Cas01+CM02 but
255 allowing "parallel tasks", that are intended to model the moldable
256 tasks of the grid scheduling literature.
258 - ``storage/model``: specify the used storage model. Only one model is
260 - ``vm/model``: specify the model for virtual machines. Only one model
263 .. todo: make 'compound' the default host model.
265 .. _options_model_solver:
270 The different models rely on a linear inequalities solver to share
271 the underlying resources. SimGrid allows you to change the solver, but
272 be cautious, **don't change it unless you are 100% sure**.
274 - items ``cpu/solver``, ``network/solver``, ``disk/solver`` and ``host/solver``
275 allow you to change the solver for each model:
277 - **maxmin:** The default solver for all models except ptask. Provides a
278 max-min fairness allocation.
279 - **fairbottleneck:** The default solver for ptasks. Extends max-min to
280 allow heterogeneous resources.
281 - **bmf:** More realistic solver for heterogeneous resource sharing.
282 Implements BMF (Bottleneck max fairness) fairness. To be used with
283 parallel tasks instead of fair-bottleneck.
285 .. _options_model_optim:
290 The network and CPU models that are based on linear inequalities solver (that
291 is, all our analytical models) accept specific optimization
294 - items ``network/optim`` and ``cpu/optim`` (both default to 'Lazy'):
296 - **Lazy:** Lazy action management (partial invalidation in lmm +
297 heap in action remaining).
298 - **TI:** Trace integration. Highly optimized mode when using
299 availability traces (only available for the Cas01 CPU model for
301 - **Full:** Full update of remaining and variables. Slow but may be
302 useful when debugging.
304 - items ``network/maxmin-selective-update`` and
305 ``cpu/maxmin-selective-update``: configure whether the underlying
306 should be lazily updated or not. It should have no impact on the
307 computed timings, but should speed up the computation. |br| It is
308 still possible to disable this feature because it can reveal
309 counter-productive in very specific scenarios where the
310 interaction level is high. In particular, if all your
311 communication share a given backbone link, you should disable it:
312 without it, a simple regular loop is used to update each
313 communication. With it, each of them is still updated (because of
314 the dependency induced by the backbone), but through a complicated
315 and slow pattern that follows the actual dependencies.
317 .. _cfg=bmf/precision:
318 .. _cfg=maxmin/precision:
319 .. _cfg=surf/precision:
324 **Option** ``maxmin/precision`` **Default:** 1e-5 (in flops or bytes) |br|
325 **Option** ``surf/precision`` **Default:** 1e-9 (in seconds) |br|
326 **Option** ``bmf/precision`` **Default:** 1e-12 (no unit)
328 The analytical models handle a lot of floating point values. It is
329 possible to change the epsilon used to update and compare them through
330 this configuration item. Changing it may speedup the simulation by
331 discarding very small actions, at the price of a reduced numerical
332 precision. You can modify separately the precision used to manipulate
333 timings (in seconds) and the one used to manipulate amounts of work
336 .. _cfg=maxmin/concurrency-limit:
341 **Option** ``maxmin/concurrency-limit`` **Default:** -1 (no limit)
343 The maximum number of variables per resource can be tuned through this
344 option. You can have as many simultaneous actions per resources as you
345 want. If your simulation presents a very high level of concurrency, it
346 may help to use e.g. 100 as a value here. It means that at most 100
347 actions can consume a resource at a given time. The extraneous actions
348 are queued and wait until the amount of concurrency of the considered
349 resource lowers under the given boundary.
351 Such limitations help both to the simulation speed and simulation accuracy
352 on highly constrained scenarios, but the simulation speed suffers of this
353 setting on regular (less constrained) scenarios so it is off by default.
355 .. _cfg=bmf/max-iterations:
360 **Option** ``bmf/max-iterations`` **Default:** 1000
362 It may happen in some settings that the BMF solver fails to converge to
363 a solution, so there is a hard limit on the amount of iteration count to
364 avoid infinite loops.
366 .. _options_model_network:
368 Configuring the Network Model
369 .............................
371 .. _cfg=network/TCP-gamma:
373 Maximal TCP Window Size
374 ^^^^^^^^^^^^^^^^^^^^^^^
376 **Option** ``network/TCP-gamma`` **Default:** 4194304
378 The analytical models need to know the maximal TCP window size to take
379 the TCP congestion mechanism into account. On Linux, this value can
380 be retrieved using the following commands. Both give a set of values,
381 and you should use the last one, which is the maximal size.
383 .. code-block:: console
385 $ cat /proc/sys/net/ipv4/tcp_rmem # gives the sender window
386 $ cat /proc/sys/net/ipv4/tcp_wmem # gives the receiver window
388 .. _cfg=network/bandwidth-factor:
389 .. _cfg=network/latency-factor:
390 .. _cfg=network/weight-S:
392 Manual calibration factors
393 ^^^^^^^^^^^^^^^^^^^^^^^^^^
395 SimGrid can take network irregularities such as a slow startup or changing behavior depending on the message size into account.
396 The values provided by default were computed a long time ago through data fitting one the timings of either packet-level
397 simulators or direct experiments on real platforms. These default values should be OK for most users, but if simulation realism
398 is really important to you, you probably want to recalibrate the models (i.e., devise sensible values for your specific
399 settings). This section only describes how to pass new values to the models while the calibration process involved in the
400 computation of these values is described :ref:`in the relevant chapter <models_calibration>`.
402 We found out that many networking effects can be realistically accounted for with the three following correction factors. They
403 were shown to be enough to capture slow-start effects, the different transmission modes of MPI systems (eager vs. rendez-vous
404 mode), or the non linear effects of wifi sharing.
406 **Option** ``network/latency-factor`` **Default:** 1.0, but overridden by most models
408 This option specifies a multiplier to apply to the *physical* latency (i.e., the one described in the platform) of the set of
409 links involved in a communication. The factor can either be a constant to apply to any communication, or it may depend on the
410 message size. The ``CM02`` model does not use any correction factor, so the latency-factor remains to 1. The ``LV08`` model sets
411 it to 13.01 to model slow-start, while the ``SMPI`` model has several possible values depending on the interval in which the
412 message size falls. The default SMPI setting given below specifies for example that a message smaller than 257 bytes will get a
413 latency multiplier of 2.01467 while a message whose size is in [15424, 65472] will get a latency multiplier of 3.48845. The
414 ``wifi`` model goes further and uses a callback in the program to compute the factor that must be non-linear in this case.
416 This multiplier is applied to the latency computed from the platform, that is the sum of all link *physical* latencies over the
417 :ref:`network path <platform_routing>` used by the considered communication, to derive the *effective* end-to-end latency.
419 Constant factors are easy to express, but the interval-based syntax used in SMPI is somewhat complex. It expects a set of
420 factors separated by semicolons, each of the form ``boundary:factor``. For example if your specification is
421 ``0:1;1000:2;5000:3``, it means that on [0, 1000) the factor is 1. On [1000,5000), the factor is 2 while the factor is 3 for
422 5000 and beyond. If your first interval does include size=0, then the default value of 1 is used before. Changing the factor
423 callback is not possible from the command line and must be done from your code, as shown in `this example
424 <https://framagit.org/simgrid/simgrid/tree/master/examples/cpp/network-factors/s4u-network-factors.cpp>`_. Note that the chosen
425 model only provides some default settings. You may pick a ``LV08`` model to get some of the settings, and override the latency
426 with interval-based values.
428 SMPI default value: 65472:11.6436; 15424:3.48845; 9376:2.59299; 5776:2.18796; 3484:1.88101; 1426:1.61075; 732:1.9503;
429 257:1.95341;0:2.01467 (interval boundaries are sorted automatically). These values were computed by data fitting on the Stampede
430 Supercomputer at TACC, with optimal deployment of processes on nodes. To accurately model your settings, you should redo the
431 :ref:`calibration <models_calibration>`.
433 **Option** ``network/bandwidth-factor`` **Default:** 1.0, but overridden by most models
435 Setting this option automatically adjusts the *effective* bandwidth (i.e., the one perceived by the application) used by any
436 given communication. As with latency-factor above, the value can be a constant (``CM02`` uses 1 -- no correction -- while
437 ``LV08`` uses 0.97 to discount TCP headers while computing the payload bandwidth), interval-based (as the default provided by
438 the ``SMPI``), or using in-program callbacks (as with ``wifi``).
440 SMPI default value: 65472:0.940694;15424:0.697866;9376:0.58729;5776:1.08739;3484:0.77493;1426:0.608902;732:0.341987;257:0.338112;0:0.812084
441 This was also computed on the Stampede Supercomputer.
443 **Option** ``network/weight-S`` **Default:** depends on the model
445 Value used to account for RTT-unfairness when sharing a bottleneck (network connections with a large RTT are generally penalized
446 against those with a small one). Described in `Accuracy Study and Improvement of Network Simulation in the SimGrid Framework
447 <http://mescal.imag.fr/membres/arnaud.legrand/articles/simutools09.pdf>`_
449 Default values for ``CM02`` is 0. ``LV08`` sets it to 20537 while both ``SMPI`` and ``IB`` set it to 8775.
451 .. _cfg=network/loopback:
453 Configuring loopback link
454 ^^^^^^^^^^^^^^^^^^^^^^^^^
456 Several network models provide an implicit loopback link to account for local
457 communication on a host. By default it has a 10GBps bandwidth and a null latency.
458 This can be changed with ``network/loopback-lat`` and ``network/loopback-bw``
459 items. Note that this loopback is conveniently modeled with a :ref:`single FATPIPE link <pf_loopback>`
460 for the whole platform. If modeling contention inside nodes is important then you should
461 rather add such loopback links (one for each host) yourself.
463 .. _cfg=smpi/IB-penalty-factors:
468 InfiniBand network behavior can be modeled through 3 parameters
469 ``smpi/IB-penalty-factors:"βe;βs;γs"``, as explained in `the PhD
470 thesis of Jean-Marc Vincent
471 <http://mescal.imag.fr/membres/jean-marc.vincent/index.html/PhD/Vienne.pdf>`_ (in French)
472 or more concisely in `this paper <https://hal.inria.fr/hal-00953618/document>`_,
473 even if that paper does only describe models for myrinet and ethernet.
474 You can see in Fig 2 some results for Infiniband, for example. This model
475 may be outdated by now for modern infiniband, anyway, so a new
476 validation would be good.
478 The three paramaters are defined as follows:
480 - βs: penalty factor for outgoing messages, computed by running a simple send to
481 two nodes and checking slowdown compared to a single send to one node,
483 - βe: penalty factor for ingoing messages, same computation method but with one
484 node receiving several messages
485 - γr: slowdown factor when communication buffer memory is saturated. It needs a
486 more complicated pattern to run in order to be computed (5.3 in the thesis,
487 page 107), and formula in the end is γr = time(c)/(3×βe×time(ref)), where
488 time(ref) is the time of a single comm with no contention).
490 Once these values are computed, a penalty is assessed for each message (this is
491 the part implemented in the simulator) as shown page 106 of the thesis. Here is
492 a simple translation of this text. First, some notations:
494 - ∆e(e) which corresponds to the incoming degree of node e, that is to say the number of communications having as destination node e.
495 - ∆s (s) which corresponds to the degree outgoing from node s, that is to say the number of communications sent by node s.
496 - Φ (e) which corresponds to the number of communications destined for the node e but coming from a different node.
497 - Ω (s, e) which corresponds to the number of messages coming from node s to node e. If node e only receives communications from different nodes then Φ (e) = ∆e (e). On the other hand if, for example, there are three messages coming from node s and going from node e then Φ (e) 6 = ∆e (e) and Ω (s, e) = 3
499 To determine the penalty for a communication, two values need to be calculated. First, the penalty caused by the conflict in transmission, noted ps.
502 - if ∆s (i) = 1 then ps = 1.
503 - if ∆s (i) ≥ 2 and ∆e (i) ≥ 3 then ps = ∆s (i) × βs × γr
504 - else, ps = ∆s (i) × βs
507 Then, the penalty caused by the conflict in reception (noted pe) should be computed as follows:
509 - if ∆e (i) = 1 then pe = 1
510 - else, pe = Φ (e) × βe × Ω (s, e)
512 Finally, the penalty associated with the communication is:
513 p = max (ps ∈ s, pe)
515 .. _cfg=network/crosstraffic:
517 Simulating Cross-Traffic
518 ^^^^^^^^^^^^^^^^^^^^^^^^
520 Since SimGrid v3.7, cross-traffic effects can be taken into account in
521 analytical simulations. It means that ongoing and incoming
522 communication flows are treated independently. In addition, the LV08
523 model adds 0.05 of usage on the opposite direction for each new
524 created flow. This can be useful to simulate some important TCP
525 phenomena such as ack compression.
527 For that to work, your platform must have two links for each
528 pair of interconnected hosts. An example of usable platform is
529 available in ``examples/platforms/crosstraffic.xml``.
531 This is activated through the ``network/crosstraffic`` item, that
532 can be set to 0 (disable this feature) or 1 (enable it).
534 Note that with the default host model this option is activated by default.
536 .. _cfg=smpi/async-small-thresh:
538 Simulating Asynchronous Send
539 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
541 (this configuration item is experimental and may change or disappear)
543 It is possible to specify that messages below a certain size (in bytes) will be
544 sent as soon as the call to MPI_Send is issued, without waiting for
545 the correspondent receive. This threshold can be configured through
546 the ``smpi/async-small-thresh`` item. The default value is 0. This
547 behavior can also be manually set for mailboxes, by setting the
548 receiving mode of the mailbox with a call to
549 :cpp:func:`MSG_mailbox_set_async`. After this, all messages sent to
550 this mailbox will have this behavior regardless of the message size.
552 This value needs to be smaller than or equals to the threshold set at
553 :ref:`cfg=smpi/send-is-detached-thresh`, because asynchronous messages
554 are meant to be detached as well.
561 **Option** ``ns3/TcpModel`` **Default:** "default" (ns-3 default)
563 When using ns-3, there is an extra item ``ns3/TcpModel``, corresponding
564 to the ``ns3::TcpL4Protocol::SocketType`` configuration item in
565 ns-3. The only valid values (enforced on the SimGrid side) are
566 'default' (no change to the ns-3 configuration), 'NewReno' or 'Reno' or
569 **Option** ``ns3/seed`` **Default:** "" (don't set the seed in ns-3)
571 This option is the random seed to provide to ns-3 with
572 ``ns3::RngSeedManager::SetSeed`` and ``ns3::RngSeedManager::SetRun``.
574 If left blank, no seed is set in ns-3. If the value 'time' is
575 provided, the current amount of seconds since epoch is used as a seed.
576 Otherwise, the provided value must be a number to use as a seed.
578 Configuring the Storage model
579 .............................
581 .. _cfg=storage/max_file_descriptors:
583 File Descriptor Count per Host
584 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
586 **Option** ``storage/max_file_descriptors`` **Default:** 1024
588 Each host maintains a fixed-size array of its file descriptors. You
589 can change its size through this item to either enlarge it if your
590 application requires it or to reduce it to save memory space.
597 SimGrid plugins allow one to extend the framework without changing its
598 source code directly. Read the source code of the existing plugins to
599 learn how to do so (in ``src/plugins``), and ask your questions to the
600 usual channels (Stack Overflow, Mailing list, IRC). The basic idea is
601 that plugins usually register callbacks to some signals of interest.
602 If they need to store some information about a given object (Link, CPU
603 or Actor), they do so through the use of a dedicated object extension.
605 Some of the existing plugins can be activated from the command line,
606 meaning that you can activate them from the command line without any
607 modification to your simulation code. For example, you can activate
608 the host energy plugin by adding ``--cfg=plugin:host_energy`` to your
611 Here is a partial list of plugins that can be activated this way. You can get
612 the full list by passing ``--cfg=plugin:help`` to your simulator.
614 - :ref:`Host Energy <plugin_host_energy>`: models the energy dissipation of the compute units.
615 - :ref:`Link Energy <plugin_link_energy>`: models the energy dissipation of the network.
616 - :ref:`Host Load <plugin_host_load>`: monitors the load of the compute units.
618 .. _options_modelchecking:
620 Configuring the Model-Checking
621 ------------------------------
623 To enable SimGrid's model-checking support, the program should
624 be executed using the simgrid-mc wrapper:
626 .. code-block:: console
628 $ simgrid-mc ./my_program
630 Safety properties are expressed as assertions using the function
631 :cpp:func:`void MC_assert(int prop)`.
633 .. _cfg=smpi/buffering:
635 Specifying the MPI buffering behavior
636 .....................................
638 **Option** ``smpi/buffering`` **Default:** infty
640 Buffering in MPI has a huge impact on the communication semantic. For example,
641 standard blocking sends are synchronous calls when the system buffers are full
642 while these calls can complete immediately without even requiring a matching
643 receive call for small messages sent when the system buffers are empty.
645 In SMPI, this depends on the message size, that is compared against two thresholds:
647 - if (size < :ref:`smpi/async-small-thresh <cfg=smpi/async-small-thresh>`) then
648 MPI_Send returns immediately, even if the corresponding receive has not be issued yet.
649 - if (:ref:`smpi/async-small-thresh <cfg=smpi/async-small-thresh>` < size < :ref:`smpi/send-is-detached-thresh <cfg=smpi/send-is-detached-thresh>`) then
650 MPI_Send returns as soon as the corresponding receive has been issued. This is known as the eager mode.
651 - if (:ref:`smpi/send-is-detached-thresh <cfg=smpi/send-is-detached-thresh>` < size) then
652 MPI_Send returns only when the message has actually been sent over the network. This is known as the rendez-vous mode.
654 The ``smpi/buffering`` (only valid with MC) option gives an easier interface to choose between these semantics. It can take two values:
656 - **zero:** means that buffering should be disabled. All communications are actually blocking.
657 - **infty:** means that buffering should be made infinite. All communications are non-blocking.
659 .. _cfg=model-check/property:
661 Specifying a liveness property
662 ..............................
664 **Option** ``model-check/property`` **Default:** unset
666 If you want to specify liveness properties, you have to pass them on
667 the command line, specifying the name of the file containing the
668 property, as formatted by the `ltl2ba <https://github.com/utwente-fmt/ltl2ba>`_ program.
669 Note that ltl2ba is not part of SimGrid and must be installed separately.
671 .. code-block:: console
673 $ simgrid-mc ./my_program --cfg=model-check/property:<filename>
675 .. _cfg=model-check/checkpoint:
677 Going for Stateful Verification
678 ...............................
680 By default, the system is backtracked to its initial state to explore
681 another path, instead of backtracking to the exact step before the fork
682 that we want to explore (this is called stateless verification). This
683 is done this way because saving intermediate states can rapidly
684 exhaust the available memory. If you want, you can change the value of
685 the ``model-check/checkpoint`` item. For example,
686 ``--cfg=model-check/checkpoint:1`` asks to take a checkpoint every
687 step. Beware, this will certainly explode your memory. Larger values
688 are probably better, make sure to experiment a bit to find the right
689 setting for your specific system.
691 .. _cfg=model-check/reduction:
693 Specifying the kind of reduction
694 ................................
696 The main issue when using the model-checking is the state space
697 explosion. You can activate some reduction technique with
698 ``--cfg=model-check/reduction:<technique>``. For now, this
699 configuration variable can take 2 values:
701 - **none:** Do not apply any kind of reduction (mandatory for
702 liveness properties, as our current DPOR algorithm breaks cycles)
703 - **dpor:** Apply Dynamic Partial Ordering Reduction. Only valid if
704 you verify local safety properties (default value for safety
707 Another way to mitigate the state space explosion is to search for
708 cycles in the exploration with the :ref:`cfg=model-check/visited`
709 configuration. Note that DPOR and state-equality reduction may not
710 play well together. You should choose between them.
712 Our current DPOR implementation could be improved in may ways. We are
713 currently improving its efficiency (both in term of reduction ability
714 and computational speed), and future work could make it compatible
715 with liveness properties.
717 .. _cfg=model-check/visited:
719 Size of Cycle Detection Set (state equality reduction)
720 ......................................................
722 Mc SimGrid can be asked to search for cycles during the exploration,
723 i.e. situations where a new explored state is in fact the same state
724 than a previous one.. This can prove useful to mitigate the state
725 space explosion with safety properties, and this is the crux when
726 searching for counter-examples to the liveness properties.
728 Note that this feature may break the current implementation of the
729 DPOR reduction technique.
731 The ``model-check/visited`` item is the maximum number of states, which
732 are stored in memory. If the maximum number of snapshotted state is
733 reached, some states will be removed from the memory and some cycles
734 might be missed. Small values can lead to incorrect verifications, but
735 large values can exhaust your memory and be CPU intensive as each new
736 state must be compared to that amount of older saved states.
738 The default settings depend on the kind of exploration. With safety
739 checking, no state is snapshotted and cycles cannot be detected. With
740 liveness checking, all states are snapshotted because missing a cycle
741 could hinder the exploration soundness.
743 .. _cfg=model-check/termination:
745 Non-Termination Detection
746 .........................
748 The ``model-check/termination`` configuration item can be used to
749 report if a non-termination execution path has been found. This is a
750 path with a cycle, which means that the program might never terminate.
752 This only works in safety mode, not in liveness mode.
754 This options is disabled by default.
756 .. _cfg=model-check/dot-output:
761 If set, the ``model-check/dot-output`` configuration item is the name
762 of a file in which to write a dot file of the path leading to the
763 property violation discovered (safety or liveness violation), as well
764 as the cycle for liveness properties. This dot file can then be fed to the
765 graphviz dot tool to generate a corresponding graphical representation.
767 .. _cfg=model-check/max-depth:
769 Exploration Depth Limit
770 .......................
772 The ``model-check/max-depth`` can set the maximum depth of the
773 exploration graph of the model checker. If this limit is reached, a
774 logging message is sent and the results might not be exact.
776 By default, the exploration is limited to the depth of 1000.
778 .. _cfg=model-check/timeout:
783 By default, the model checker does not handle timeout conditions: the `wait`
784 operations never time out. With the ``model-check/timeout`` configuration item
785 set to **yes**, the model checker will explore timeouts of `wait` operations.
787 .. _cfg=model-check/communications-determinism:
788 .. _cfg=model-check/send-determinism:
790 Communication Determinism
791 .........................
793 The ``model-check/communications-determinism`` and
794 ``model-check/send-determinism`` items can be used to select the
795 communication determinism mode of the model checker, which checks
796 determinism properties of the communications of an application.
798 .. _cfg=model-check/setenv:
800 Passing environment variables
801 .............................
803 You can specify extra environment variables to be set in the verified application
804 with ``model-check/setenv``. For example, you can preload a library as follows:
805 ``-cfg=model-check/setenv:LD_PRELOAD=toto;LD_LIBRARY_PATH=/tmp``.
809 Verification Performance Considerations
810 .......................................
812 The size of the stacks can have a huge impact on the memory
813 consumption when using model-checking. By default, each snapshot will
814 save a copy of the whole stacks and not only of the part that is
815 really meaningful: you should expect the contribution of the memory
816 consumption of the snapshots to be:
817 :math:`\text{number of processes} \times \text{stack size} \times \text{number of states}`.
819 When compiled against the model checker, the stacks are not
820 protected with guards: if the stack size is too small for your
821 application, the stack will silently overflow into other parts of the
822 memory (see :ref:`contexts/guard-size <cfg=contexts/guard-size>`).
824 .. _cfg=model-check/replay:
826 Replaying buggy execution paths from the model checker
827 ......................................................
829 Debugging the problems reported by the model checker is challenging:
830 First, the application under verification cannot be debugged with gdb
831 because the model checker already traces it. Then, the model checker may
832 explore several execution paths before encountering the issue, making it
833 very difficult to understand the output. Fortunately, SimGrid provides
834 the execution path leading to any reported issue so that you can replay
835 this path reported by the model checker, enabling the usage of classical
838 When the model checker finds an interesting path in the application
839 execution graph (where a safety or liveness property is violated), it
840 generates an identifier for this path. Here is an example of the output:
842 .. code-block:: console
844 [ 0.000000] (0:@) Check a safety property
845 [ 0.000000] (0:@) **************************
846 [ 0.000000] (0:@) *** PROPERTY NOT VALID ***
847 [ 0.000000] (0:@) **************************
848 [ 0.000000] (0:@) Counter-example execution trace:
849 [ 0.000000] (0:@) [(1)Tremblay (app)] MC_RANDOM(3)
850 [ 0.000000] (0:@) [(1)Tremblay (app)] MC_RANDOM(4)
851 [ 0.000000] (0:@) Path = 1/3;1/4
852 [ 0.000000] (0:@) Expanded states = 27
853 [ 0.000000] (0:@) Visited states = 68
854 [ 0.000000] (0:@) Executed transitions = 46
856 The interesting line is ``Path = 1/3;1/4``, which means that you should use
857 ``--cfg=model-check/replay:1/3;1/4`` to replay your application on the buggy
858 execution path. All options (but the model checker related ones) must
859 remain the same. In particular, if you ran your application with
860 ``smpirun -wrapper simgrid-mc``, then do it again. Remove all
861 MC-related options, keep non-MC-related ones and add
862 ``--cfg=model-check/replay:???``.
864 Currently, if the path is of the form ``X;Y;Z``, each number denotes
865 the actor's pid that is selected at each indecision point. If it's of
866 the form ``X/a;Y/b``, the X and Y are the selected pids while the a
867 and b are the return values of their simcalls. In the previous
868 example, ``1/3;1/4``, you can see from the full output that the actor
869 1 is doing MC_RANDOM simcalls, so the 3 and 4 simply denote the values
870 that these simcall return on the execution branch leading to the
873 Configuring the User Code Virtualization
874 ----------------------------------------
876 .. _cfg=contexts/factory:
878 Selecting the Virtualization Factory
879 ....................................
881 **Option** contexts/factory **Default:** "raw"
883 In SimGrid, the user code is virtualized in a specific mechanism that
884 allows the simulation kernel to control its execution: when a user
885 process requires a blocking action (such as sending a message), it is
886 interrupted, and only gets released when the simulated clock reaches
887 the point where the blocking operation is done. This is explained
888 graphically in the `relevant tutorial, available online
889 <https://simgrid.org/tutorials/simgrid-simix-101.pdf>`_.
891 In SimGrid, the containers in which user processes are virtualized are
892 called contexts. Several context factory are provided, and you can
893 select the one you want to use with the ``contexts/factory``
894 configuration item. Some of the following may not exist on your
895 machine because of portability issues. In any case, the default one
896 should be the most effcient one (please report bugs if the
897 auto-detection fails for you). They are approximately sorted here from
898 the slowest to the most efficient:
900 - **thread:** very slow factory using full featured threads (either
901 pthreads or windows native threads). They are slow but very
902 standard. Some debuggers or profilers only work with this factory.
903 - **java:** Java applications are virtualized onto java threads (that
904 are regular pthreads registered to the JVM)
905 - **ucontext:** fast factory using System V contexts (Linux and FreeBSD only)
906 - **boost:** This uses the `context
907 implementation <http://www.boost.org/doc/libs/1_59_0/libs/context/doc/html/index.html>`_
908 of the boost library for a performance that is comparable to our
910 |br| Install the relevant library (e.g. with the
911 libboost-contexts-dev package on Debian/Ubuntu) and recompile
913 - **raw:** amazingly fast factory using a context switching mechanism
914 of our own, directly implemented in assembly (only available for x86
915 and amd64 platforms for now) and without any unneeded system call.
917 The main reason to change this setting is when the debugging tools become
918 fooled by the optimized context factories. Threads are the most
919 debugging-friendly contexts, as they allow one to set breakpoints
920 anywhere with gdb and visualize backtraces for all processes, in order
921 to debug concurrency issues. Valgrind is also more comfortable with
922 threads, but it should be usable with all factories (Exception: the
923 callgrind tool really dislikes raw and ucontext factories).
925 .. _cfg=contexts/stack-size:
927 Adapting the Stack Size
928 .......................
930 **Option** ``contexts/stack-size`` **Default:** 8192 KiB
932 Each virtualized used process is executed using a specific system
933 stack. The size of this stack has a huge impact on the simulation
934 scalability, but its default value is rather large. This is because
935 the error messages that you get when the stack size is too small are
936 rather disturbing: this leads to stack overflow (overwriting other
937 stacks), leading to segfaults with corrupted stack traces.
939 If you want to push the scalability limits of your code, you might
940 want to reduce the ``contexts/stack-size`` item. Its default value is
941 8192 (in KiB), while our Chord simulation works with stacks as small
942 as 16 KiB, for example. You can ensure that some actors have a specific
943 size by simply changing the value of this configuration item before
944 creating these actors. The :cpp:func:`simgrid::s4u::Engine::set_config`
945 functions are handy for that.
947 This *setting is ignored* when using the thread factory (because there
948 is no way to modify the stack size with C++ system threads). Instead,
949 you should compile SimGrid and your application with
950 ``-fsplit-stack``. Note that this compilation flag is not compatible
951 with the model checker right now.
953 The operating system should only allocate memory for the pages of the
954 stack which are actually used and you might not need to use this in
955 most cases. However, this setting is very important when using the
956 model checker (see :ref:`options_mc_perf`).
958 .. _cfg=contexts/guard-size:
960 Disabling Stack Guard Pages
961 ...........................
963 **Option** ``contexts/guard-size`` **Default** 1 page in most case (0 pages on Windows or with MC)
965 Unless you use the threads context factory (see
966 :ref:`cfg=contexts/factory`), a stack guard page is usually used
967 which prevents the stack of a given actor from overflowing on another
968 stack. But the performance impact may become prohibitive when the
969 amount of actors increases. The option ``contexts/guard-size`` is the
970 number of stack guard pages used. By setting it to 0, no guard pages
971 will be used: in this case, you should avoid using small stacks (with
972 :ref:`contexts/stack-size <cfg=contexts/stack-size>`) as the stack
973 will silently overflow on other parts of the memory.
975 When no stack guard page is created, stacks may then silently overflow
976 on other parts of the memory if their size is too small for the
979 .. _cfg=contexts/nthreads:
980 .. _cfg=contexts/synchro:
982 Running User Code in Parallel
983 .............................
985 Parallel execution of the user code is only considered stable in
986 SimGrid v3.7 and higher, and mostly for MSG simulations. SMPI
987 simulations may well fail in parallel mode. It is described in
988 `INRIA RR-7653 <http://hal.inria.fr/inria-00602216/>`_.
990 If you are using the **ucontext** or **raw** context factories, you can
991 request to execute the user code in parallel. Several threads are
992 launched, each of them handling the same number of user contexts at each
993 run. To activate this, set the ``contexts/nthreads`` item to the amount
994 of cores that you have in your computer (or lower than 1 to have the
995 amount of cores auto-detected).
997 When parallel execution is activated, you can choose the
998 synchronization schema used with the ``contexts/synchro`` item,
999 which value is either:
1001 - **futex:** ultra optimized synchronisation schema, based on futexes
1002 (fast user-mode mutexes), and thus only available on Linux systems.
1003 This is the default mode when available.
1004 - **posix:** slow but portable synchronisation using only POSIX
1006 - **busy_wait:** not really a synchronisation: the worker threads
1007 constantly request new contexts to execute. It should be the most
1008 efficient synchronisation schema, but it loads all the cores of
1009 your machine for no good reason. You probably prefer the other less
1012 Configuring the Tracing
1013 -----------------------
1015 The :ref:`tracing subsystem <outcome_vizu>` can be configured in
1016 several different ways depending on the used interface (S4U, SMPI)
1017 and the kind of traces that needs to be obtained. See the
1018 :ref:`Tracing Configuration Options subsection
1019 <tracing_tracing_options>` for a full description of each
1020 configuration option.
1022 We detail here a simple way to get the traces working for you, even if
1023 you never used the tracing API.
1026 - Any SimGrid-based simulator (MSG, SMPI, ...) and raw traces:
1028 .. code-block:: none
1030 --cfg=tracing:yes --cfg=tracing/uncategorized:yes
1032 The first parameter activates the tracing subsystem, and the second
1033 tells it to trace host and link utilization (without any
1036 - MSG-based simulator and categorized traces (you need to
1037 declare categories and classify your tasks according to them)
1039 .. code-block:: none
1041 --cfg=tracing:yes --cfg=tracing/categorized:yes
1043 The first parameter activates the tracing subsystem, and the second
1044 tells it to trace host and link categorized utilization.
1046 - SMPI simulator and traces for a space/time view:
1048 .. code-block:: console
1050 $ smpirun -trace ...
1052 The `-trace` parameter for the smpirun script runs the simulation
1053 with ``--cfg=tracing:yes --cfg=tracing/smpi:yes``. Check the
1054 smpirun's `-help` parameter for additional tracing options.
1056 Sometimes you might want to put additional information on the trace to
1057 correctly identify them later, or to provide data that can be used to
1058 reproduce an experiment. You have two ways to do that:
1060 - Add a string on top of the trace file as comment:
1062 .. code-block:: none
1064 --cfg=tracing/comment:my_simulation_identifier
1066 - Add the contents of a textual file on top of the trace file as comment:
1068 .. code-block:: none
1070 --cfg=tracing/comment-file:my_file_with_additional_information.txt
1072 Please, use these two parameters (for comments) to make reproducible
1073 simulations. For additional details about this and all tracing
1074 options, check See the :ref:`tracing_tracing_options`.
1079 .. _cfg=msg/debug-multiple-use:
1084 **Option** ``msg/debug-multiple-use`` **Default:** off
1086 Sometimes your application may try to send a task that is still being
1087 executed somewhere else, making it impossible to send this task. However,
1088 for debugging purposes, one may want to know what the other host is/was
1089 doing. This option shows a backtrace of the other process.
1094 The SMPI interface provides several specific configuration items.
1095 These are not easy to see with ``--help-cfg``, since SMPI binaries are usually launched through the ``smiprun`` script.
1097 .. _cfg=smpi/host-speed:
1098 .. _cfg=smpi/cpu-threshold:
1099 .. _cfg=smpi/simulate-computation:
1101 Automatic Benchmarking of SMPI Code
1102 ...................................
1104 In SMPI, the sequential code is automatically benchmarked, and these
1105 computations are automatically reported to the simulator. That is to
1106 say that if you have a large computation between a ``MPI_Recv()`` and
1107 a ``MPI_Send()``, SMPI will automatically benchmark the duration of
1108 this code, and create an execution task within the simulator to take
1109 this into account. For that, the actual duration is measured on the
1110 host machine and then scaled to the power of the corresponding
1111 simulated machine. The variable ``smpi/host-speed`` allows one to
1112 specify the computational speed of the host machine (in flop/s by
1113 default) to use when scaling the execution times.
1115 The default value is ``smpi/host-speed=20kf`` (= 20,000 flop/s). This
1116 is probably underestimated for most machines, leading SimGrid to
1117 overestimate the amount of flops in the execution blocks that are
1118 automatically injected in the simulator. As a result, the execution
1119 time of the whole application will probably be overestimated until you
1120 use a realistic value.
1122 When the code consists of numerous consecutive MPI calls, the
1123 previous mechanism feeds the simulation kernel with numerous tiny
1124 computations. The ``smpi/cpu-threshold`` item becomes handy when this
1125 impacts badly on the simulation performance. It specifies a threshold (in
1126 seconds) below which the execution chunks are not reported to the
1127 simulation kernel (default value: 1e-6).
1129 .. note:: The option ``smpi/cpu-threshold`` ignores any computation
1130 time spent below this threshold. SMPI does not consider the
1131 `amount of time` of these computations; there is no offset for
1132 this. Hence, a value that is too small, may lead to unreliable
1135 In some cases, however, one may wish to disable simulation of
1136 the computation of an application. This is the case when SMPI is used not to
1137 simulate an MPI application, but instead an MPI code that performs
1138 "live replay" of another MPI app (e.g., ScalaTrace's replay tool, or
1139 various on-line simulators that run an app at scale). In this case the
1140 computation of the replay/simulation logic should not be simulated by
1141 SMPI. Instead, the replay tool or on-line simulator will issue
1142 "computation events", which correspond to the actual MPI simulation
1143 being replayed/simulated. At the moment, these computation events can
1144 be simulated using SMPI by calling internal smpi_execute*() functions.
1146 To disable the benchmarking/simulation of a computation in the simulated
1147 application, the variable ``smpi/simulate-computation`` should be set
1148 to **no**. This option just ignores the timings in your simulation; it
1149 still executes the computations itself. If you want to stop SMPI from
1150 doing that, you should check the SMPI_SAMPLE macros, documented in
1151 Section :ref:`SMPI_use_faster`.
1153 +------------------------------------+-------------------------+-----------------------------+
1154 | Solution | Computations executed? | Computations simulated? |
1155 +====================================+=========================+=============================+
1156 | --cfg=smpi/simulate-computation:no | Yes | Never |
1157 +------------------------------------+-------------------------+-----------------------------+
1158 | --cfg=smpi/cpu-threshold:42 | Yes, in all cases | If it lasts over 42 seconds |
1159 +------------------------------------+-------------------------+-----------------------------+
1160 | SMPI_SAMPLE() macro | Only once per loop nest | Always |
1161 +------------------------------------+-------------------------+-----------------------------+
1163 .. _cfg=smpi/comp-adjustment-file:
1165 Slow-down or speed-up parts of your code
1166 ........................................
1168 **Option** ``smpi/comp-adjustment-file:`` **Default:** unset
1170 This option allows you to pass a file that contains two columns: The
1171 first column defines the section that will be subject to a speedup;
1172 the second column is the speedup. For instance:
1174 .. code-block:: none
1176 "start:stop","ratio"
1177 "exchange_1.f:30:exchange_1.f:130",1.18244559422142
1179 The first line is the header - you must include it. The following
1180 line means that the code between two consecutive MPI calls on line 30
1181 in exchange_1.f and line 130 in exchange_1.f should receive a speedup
1182 of 1.18244559422142. The value for the second column is therefore a
1183 speedup, if it is larger than 1 and a slowdown if it is smaller
1184 than 1. Nothing will be changed if it is equal to 1.
1186 Of course, you can set any arbitrary filenames you want (so the start
1187 and end don't have to be in the same file), but be aware that this
1188 mechanism only supports `consecutive calls!`
1190 Please note that you must pass the ``-trace-call-location`` flag to
1191 smpicc or smpiff, respectively. This flag activates some internal
1192 macro definitions that help with obtaining the call location.
1194 Bandwidth and latency factors
1195 .............................
1197 Adapting the bandwidth and latency acurately to the network conditions is of a paramount importance to get realistic results.
1198 This is done through the :ref:`network/bandwidth-factor <cfg=network/bandwidth-factor>` and :ref:`network/latency-factor
1199 <cfg=network/latency-factor>` items. You probably also want to read the following section: :ref:`models_calibration`.
1201 .. _cfg=smpi/display-timing:
1203 Reporting Simulation Time
1204 .........................
1206 **Option** ``smpi/display-timing`` **Default:** 0 (false)
1208 Most of the time, you run MPI code with SMPI to compute the time it
1209 would take to run it on a platform. But since the code is run through
1210 the ``smpirun`` script, you don't have any control on the launcher
1211 code, making it difficult to report the simulated time when the
1212 simulation ends. If you enable the ``smpi/display-timing`` item,
1213 ``smpirun`` will display this information when the simulation
1215 SMPI will also display information about the amout of real time spent
1216 in application code and in SMPI internals, to provide hints about the
1217 need to use sampling to reduce simulation time.
1219 .. _cfg=smpi/display-allocs:
1221 Reporting memory allocations
1222 ............................
1224 **Option** ``smpi/display-allocs`` **Default:** 0 (false)
1226 SMPI intercepts malloc and calloc calls performed inside the running
1227 application, if it wasn't compiled with SMPI_NO_OVERRIDE_MALLOC.
1228 With this option, SMPI will show at the end of execution the amount of
1229 memory allocated through these calls, and locate the most expensive one.
1230 This helps finding the targets for manual memory sharing, or the threshold
1231 to use for smpi/auto-shared-malloc-thresh option (see :ref:`cfg=smpi/auto-shared-malloc-thresh`).
1233 .. _cfg=smpi/keep-temps:
1235 Keeping temporary files after simulation
1236 ........................................
1238 **Option** ``smpi/keep-temps`` **default:** 0 (false)
1240 SMPI usually generates a lot of temporary files that are cleaned after
1241 use. This option requests to preserve them, for example to debug or
1242 profile your code. Indeed, the binary files are removed very early
1243 under the dlopen privatization schema, which tends to fool the
1246 .. _cfg=smpi/papi-events:
1248 Trace hardware counters with PAPI
1249 .................................
1251 **Option** ``smpi/papi-events`` **default:** unset
1253 When the PAPI support is compiled into SimGrid, this option takes the
1254 names of PAPI counters and adds their respective values to the trace
1255 files (See Section :ref:`tracing_tracing_options`).
1259 This feature currently requires superuser privileges, as registers
1260 are queried. Only use this feature with code you trust! Call
1261 smpirun for instance via ``smpirun -wrapper "sudo "
1262 <your-parameters>`` or run ``sudo sh -c "echo 0 >
1263 /proc/sys/kernel/perf_event_paranoid"`` In the later case, sudo
1264 will not be required.
1266 It is planned to make this feature available on a per-process (or per-thread?) basis.
1267 The first draft, however, just implements a "global" (i.e., for all processes) set
1268 of counters, the "default" set.
1270 .. code-block:: none
1272 --cfg=smpi/papi-events:"default:PAPI_L3_LDM:PAPI_L2_LDM"
1274 .. _cfg=smpi/privatization:
1276 Automatic Privatization of Global Variables
1277 ...........................................
1279 **Option** ``smpi/privatization`` **default:** "dlopen" (when using smpirun)
1281 MPI executables are usually meant to be executed in separate
1282 processes, but SMPI is executed in only one process. Global variables
1283 from executables will be placed in the same memory region and shared
1284 between processes, causing intricate bugs. Several options are
1285 possible to avoid this, as described in the main `SMPI publication
1286 <https://hal.inria.fr/hal-01415484>`_ and in the :ref:`SMPI
1287 documentation <SMPI_what_globals>`. SimGrid provides two ways of
1288 automatically privatizing the globals, and this option allows one to
1289 choose between them.
1291 - **no** (default when not using smpirun): Do not automatically
1292 privatize variables. Pass ``-no-privatize`` to smpirun to disable
1294 - **dlopen** or **yes** (default when using smpirun): Link multiple
1295 times against the binary.
1296 - **mmap** (slower, but maybe somewhat more stable):
1297 Runtime automatic switching of the data segments.
1300 This configuration option cannot be set in your platform file. You can only
1301 pass it as an argument to smpirun.
1303 .. _cfg=smpi/privatize-libs:
1305 Automatic privatization of global variables inside external libraries
1306 .....................................................................
1308 **Option** ``smpi/privatize-libs`` **default:** unset
1310 **Linux/BSD only:** When using dlopen (default) privatization,
1311 privatize specific shared libraries with internal global variables, if
1312 they can't be linked statically. For example libgfortran is usually
1313 used for Fortran I/O and indexes in files can be mixed up.
1315 Multiple libraries can be given, semicolon separated.
1317 This configuration option can only use either full paths to libraries,
1318 or full names. Check with ldd the name of the library you want to
1321 .. code-block:: console
1325 libgfortran.so.3 => /usr/lib/x86_64-linux-gnu/libgfortran.so.3 (0x00007fbb4d91b000)
1328 Then you can use ``--cfg=smpi/privatize-libs:libgfortran.so.3``
1329 or ``--cfg=smpi/privatize-libs:/usr/lib/x86_64-linux-gnu/libgfortran.so.3``,
1330 but not ``libgfortran`` nor ``libgfortran.so``.
1332 .. _cfg=smpi/send-is-detached-thresh:
1334 Simulating MPI detached send
1335 ............................
1337 **Option** ``smpi/send-is-detached-thresh`` **default:** 65536
1339 This threshold specifies the size in bytes under which the send will
1340 return immediately. This is different from the threshold detailed in
1341 :ref:`cfg=smpi/async-small-thresh` because the message is not
1342 really sent when the send is posted. SMPI still waits for the
1343 corresponding receive to be posted, in order to perform the communication
1346 .. _cfg=smpi/coll-selector:
1348 Simulating MPI collective algorithms
1349 ....................................
1351 **Option** ``smpi/coll-selector`` **Possible values:** naive (default), ompi, mpich
1353 SMPI implements more than 100 different algorithms for MPI collective
1354 communication, to accurately simulate the behavior of most of the
1355 existing MPI libraries. The ``smpi/coll-selector`` item can be used to
1356 select the decision logic either of the OpenMPI or the MPICH libraries. (By
1357 default SMPI uses naive version of collective operations.)
1359 Each collective operation can be manually selected with a
1360 ``smpi/collective_name:algo_name``. Available algorithms are listed in
1361 :ref:`SMPI_use_colls`.
1363 .. TODO:: All available collective algorithms will be made available
1364 via the ``smpirun --help-coll`` command.
1366 .. _cfg=smpi/barrier-collectives:
1368 Add a barrier in all collectives
1369 ................................
1371 **Option** ``smpi/barrier-collectives`` **default:** off
1373 This option adds a simple barrier in some collective operations to catch dangerous
1374 code that may or may not work depending on the MPI implementation: Bcast, Exscan,
1375 Gather, Gatherv, Scan, Scatter, Scatterv and Reduce.
1377 For example, the following code works with OpenMPI while it deadlocks in MPICH and
1378 Intel MPI. Broadcast seem to be "fire and forget" in OpenMPI while other
1379 implementations expect to receive a message.
1384 MPI_Bcast(buf1, buff_size, MPI_CHAR, 0, newcom);
1385 MPI_Send(&buf2, buff_size, MPI_CHAR, 1, tag, newcom);
1386 } else if (rank==1) {
1387 MPI_Recv(&buf2, buff_size, MPI_CHAR, 0, tag, newcom, MPI_STATUS_IGNORE);
1388 MPI_Bcast(buf1, buff_size, MPI_CHAR, 0, newcom);
1391 The barrier is only simulated and does not involve any additional message (it is a S4U barrier).
1392 This option is disabled by default, and activated by the `-analyze` flag of smpirun.
1394 .. _cfg=smpi/barrier-finalization:
1396 Add a barrier in MPI_Finalize
1397 .............................
1399 **Option** ``smpi/finalization-barrier`` **default:** off
1401 By default, SMPI processes are destroyed as soon as soon as their code ends,
1402 so after a successful MPI_Finalize call returns. In some rare cases, some data
1403 might have been attached to MPI objects still active in the remaining processes,
1404 and can be destroyed eagerly by the finished process.
1405 If your code shows issues at finalization, such as segmentation fault, triggering
1406 this option will add an explicit MPI_Barrier(MPI_COMM_WORLD) call inside the
1407 MPI_Finalize, so that all processes will terminate at almost the same point.
1408 It might affect the total timing by the cost of a barrier.
1410 .. _cfg=smpi/errors-are-fatal:
1412 Disable MPI fatal errors
1413 ........................
1415 **Option** ``smpi/errors-are-fatal`` **default:** on
1417 By default, SMPI processes will crash if a MPI error code is returned. MPI allows
1418 to explicitely set MPI_ERRORS_RETURN errhandler to avoid this behaviour. This flag
1419 will turn on this behaviour by default (for all concerned types and errhandlers).
1420 This can ease debugging by going after the first reported error.
1422 .. _cfg=smpi/pedantic:
1424 Disable pedantic MPI errors
1425 ...........................
1427 **Option** ``smpi/pedantic`` **default:** on
1429 By default, SMPI will report all errors it finds in MPI codes. Some of these errors
1430 may not be considered as errors by all developers. This flag can be turned off to
1431 avoid reporting some usually harmless mistakes.
1432 Concerned errors list (will be expanded in the future):
1434 - Calling MPI_Win_fence only once in a program, hence just opening an epoch without
1437 .. _cfg=smpi/iprobe:
1439 Inject constant times for MPI_Iprobe
1440 ....................................
1442 **Option** ``smpi/iprobe`` **default:** 0.0001
1444 The behavior and motivation for this configuration option is identical
1445 with :ref:`smpi/test <cfg=smpi/test>`, but for the function
1448 .. _cfg=smpi/iprobe-cpu-usage:
1450 Reduce speed for iprobe calls
1451 .............................
1453 **Option** ``smpi/iprobe-cpu-usage`` **default:** 1 (no change)
1455 MPI_Iprobe calls can be heavily used in applications. To account
1456 correctly for the energy that cores spend probing, it is necessary to
1457 reduce the load that these calls cause inside SimGrid.
1459 For instance, we measured a maximum power consumption of 220 W for a
1460 particular application but only 180 W while this application was
1461 probing. Hence, the correct factor that should be passed to this
1462 option would be 180/220 = 0.81.
1466 Inject constant times for MPI_Init
1467 ..................................
1469 **Option** ``smpi/init`` **default:** 0
1471 The behavior and motivation for this configuration option is identical
1472 with :ref:`smpi/test <cfg=smpi/test>`, but for the function ``MPI_Init()``.
1476 Inject constant times for MPI_Isend()
1477 .....................................
1479 **Option** ``smpi/ois``
1481 The behavior and motivation for this configuration option is identical
1482 with :ref:`smpi/os <cfg=smpi/os>`, but for the function ``MPI_Isend()``.
1486 Inject constant times for MPI_send()
1487 ....................................
1489 **Option** ``smpi/os``
1491 In several network models such as LogP, send (MPI_Send, MPI_Isend) and
1492 receive (MPI_Recv) operations incur costs (i.e., they consume CPU
1493 time). SMPI can factor these costs in as well, but the user has to
1494 configure SMPI accordingly as these values may vary by machine. This
1495 can be done by using ``smpi/os`` for MPI_Send operations; for MPI_Isend
1496 and MPI_Recv, use ``smpi/ois`` and ``smpi/or``, respectively. These work
1497 exactly as ``smpi/ois``.
1499 This item can consist of multiple sections; each section takes three
1500 values, for example ``1:3:2;10:5:1``. The sections are divided by ";"
1501 so this example contains two sections. Furthermore, each section
1502 consists of three values.
1504 1. The first value denotes the minimum size in bytes for this section to take effect;
1505 read it as "if message size is greater than this value (and other section has a larger
1506 first value that is also smaller than the message size), use this".
1507 In the first section above, this value is "1".
1509 2. The second value is the startup time; this is a constant value that will always
1510 be charged, no matter what the size of the message. In the first section above,
1513 3. The third value is the `per-byte` cost. That is, it is charged for every
1514 byte of the message (incurring cost messageSize*cost_per_byte)
1515 and hence accounts also for larger messages. In the first
1516 section of the example above, this value is "2".
1518 Now, SMPI always checks which section it should use for a given
1519 message; that is, if a message of size 11 is sent with the
1520 configuration of the example above, only the second section will be
1521 used, not the first, as the first value of the second section is
1522 closer to the message size. Hence, when ``smpi/os=1:3:2;10:5:1``, a
1523 message of size 11 incurs the following cost inside MPI_Send:
1524 ``5+11*1`` because 5 is the startup cost and 1 is the cost per byte.
1526 Note that the order of sections can be arbitrary; they will be ordered internally.
1530 Inject constant times for MPI_Recv()
1531 ....................................
1533 **Option** ``smpi/or``
1535 The behavior and motivation for this configuration option is identical
1536 with :ref:`smpi/os <cfg=smpi/os>`, but for the function ``MPI_Recv()``.
1539 .. _cfg=smpi/grow-injected-times:
1541 Inject constant times for MPI_Test
1542 ..................................
1544 **Option** ``smpi/test`` **default:** 0.0001
1546 By setting this option, you can control the amount of time a process
1547 sleeps when MPI_Test() is called; this is important, because SimGrid
1548 normally only advances the time while communication is happening and
1549 thus, MPI_Test will not add to the time, resulting in deadlock if it is
1550 used as a break-condition as in the following example:
1555 MPI_Test(request, flag, status);
1559 To speed up execution, we use a counter to keep track of how often we
1560 checked if the handle is now valid or not. Hence, we actually
1561 use counter*SLEEP_TIME, that is, the time MPI_Test() causes the
1562 process to sleep increases linearly with the number of previously
1563 failed tests. This behavior can be disabled by setting
1564 ``smpi/grow-injected-times`` to **no**. This will also disable this
1565 behavior for MPI_Iprobe.
1567 .. _cfg=smpi/shared-malloc:
1568 .. _cfg=smpi/shared-malloc-hugepage:
1573 **Option** ``smpi/shared-malloc`` **Possible values:** global (default), local
1575 If your simulation consumes too much memory, you may want to modify
1576 your code so that the working areas are shared by all MPI ranks. For
1577 example, in a block-cyclic matrix multiplication, you will only
1578 allocate one set of blocks, and all processes will share them.
1579 Naturally, this will lead to very wrong results, but this will save a
1580 lot of memory. So this is still desirable for some studies. For more on
1581 the motivation for that feature, please refer to the `relevant section
1582 <https://simgrid.github.io/SMPI_CourseWare/topic_understanding_performance/matrixmultiplication>`_
1583 of the SMPI CourseWare (see Activity #2.2 of the pointed
1584 assignment). In practice, change the calls for malloc() and free() into
1585 SMPI_SHARED_MALLOC() and SMPI_SHARED_FREE().
1587 SMPI provides two algorithms for this feature. The first one, called
1588 ``local``, allocates one block per call to SMPI_SHARED_MALLOC()
1589 (each call site gets its own block) ,and this block is shared
1590 among all MPI ranks. This is implemented with the shm_* functions
1591 to create a new POSIX shared memory object (kept in RAM, in /dev/shm)
1592 for each shared block.
1594 With the ``global`` algorithm, each call to SMPI_SHARED_MALLOC()
1595 returns a new address, but it only points to a shadow block: its memory
1596 area is mapped on a 1 MiB file on disk. If the returned block is of size
1597 N MiB, then the same file is mapped N times to cover the whole block.
1598 At the end, no matter how many times you call SMPI_SHARED_MALLOC, this will
1599 only consume 1 MiB in memory.
1601 You can disable this behavior and come back to regular mallocs (for
1602 example for debugging purposes) using ``no`` as a value.
1604 If you want to keep private some parts of the buffer, for instance if these
1605 parts are used by the application logic and should not be corrupted, you
1606 can use SMPI_PARTIAL_SHARED_MALLOC(size, offsets, offsets_count). For example:
1610 mem = SMPI_PARTIAL_SHARED_MALLOC(500, {27,42 , 100,200}, 2);
1612 This will allocate 500 bytes to mem, such that mem[27..41] and
1613 mem[100..199] are shared while other area remain private.
1615 Then, it can be deallocated by calling SMPI_SHARED_FREE(mem).
1617 When smpi/shared-malloc:global is used, the memory consumption problem
1618 is solved, but it may induce too much load on the kernel's pages table.
1619 In this case, you should use huge pages so that the kernel creates only one
1620 entry per MB of malloced data instead of one entry per 4 kB.
1621 To activate this, you must mount a hugetlbfs on your system and allocate
1622 at least one huge page:
1624 .. code-block:: console
1627 $ sudo mount none /home/huge -t hugetlbfs -o rw,mode=0777
1628 $ sudo sh -c 'echo 1 > /proc/sys/vm/nr_hugepages' # echo more if you need more
1630 Then, you can pass the option
1631 ``--cfg=smpi/shared-malloc-hugepage:/home/huge`` to smpirun to
1632 actually activate the huge page support in shared mallocs.
1634 .. _cfg=smpi/auto-shared-malloc-thresh:
1636 Automatically share allocations
1637 ...............................
1639 **Option** ``smpi/auto-shared-malloc-thresh:`` **Default:** 0 (false)
1640 This value in bytes represents the size above which all allocations
1641 will be "shared" by default (as if they were performed through
1642 SMPI_SHARED_MALLOC macros). Default = 0 = disabled feature.
1643 The value must be carefully chosen to only select data buffers which
1644 will not modify execution path or cause crash if their content is false.
1645 Option :ref:`cfg=smpi/display-allocs` can be used to locate the largest
1646 allocation detected in a run, and provide a good starting threshold.
1647 Note : malloc, calloc and free are overridden by smpicc/cxx by default.
1648 This can cause some troubles if codes are already overriding these. If this
1649 is the case, defining SMPI_NO_OVERRIDE_MALLOC in the compilation flags can
1650 help, but will make this feature unusable.
1654 Inject constant times for MPI_Wtime, gettimeofday and clock_gettime
1655 ...................................................................
1657 **Option** ``smpi/wtime`` **default:** 10 ns
1659 This option controls the amount of (simulated) time spent in calls to
1660 MPI_Wtime(), gettimeofday() and clock_gettime(). If you set this value
1661 to 0, the simulated clock is not advanced in these calls, which leads
1662 to issues if your application contains such a loop:
1666 while(MPI_Wtime() < some_time_bound) {
1667 /* some tests, with no communication nor computation */
1670 When the option smpi/wtime is set to 0, the time advances only on
1671 communications and computations. So the previous code results in an
1672 infinite loop: the current [simulated] time will never reach
1673 ``some_time_bound``. This infinite loop is avoided when that option
1674 is set to a small value, as it is by default since SimGrid v3.21.
1676 Note that if your application does not contain any loop depending on
1677 the current time only, then setting this option to a non-zero value
1678 will slow down your simulations by a tiny bit: the simulation loop has
1679 to be broken out of and reset each time your code asks for the current time.
1680 If the simulation speed really matters to you, you can avoid this
1681 extra delay by setting smpi/wtime to 0.
1683 .. _cfg=smpi/list-leaks:
1685 Report leaked MPI objects
1686 .........................
1688 **Option** ``smpi/list-leaks`` **default:** 0
1690 This option controls whether to report leaked MPI objects.
1691 The parameter is the number of leaks to report.
1693 Other Configurations
1694 --------------------
1696 .. _cfg=debug/clean-atexit:
1698 Cleanup at Termination
1699 ......................
1701 **Option** ``debug/clean-atexit`` **default:** on
1703 If your code is segfaulting during its finalization, it may help to
1704 disable this option to request that SimGrid not attempt any cleanups at
1705 the end of the simulation. Since the Unix process is ending anyway,
1706 the operating system will wipe it all.
1713 **Option** ``path`` **default:** . (current dir)
1715 It is possible to specify a list of directories to search in for the
1716 trace files (see :ref:`pf_trace`) by using this configuration
1717 item. To add several directory to the path, set the configuration
1718 item several times, as in ``--cfg=path:toto --cfg=path:tutu``
1720 .. _cfg=debug/breakpoint:
1725 **Option** ``debug/breakpoint`` **default:** unset
1727 This configuration option sets a breakpoint: when the simulated clock
1728 reaches the given time, a SIGTRAP is raised. This can be used to stop
1729 the execution and get a backtrace with a debugger.
1731 It is also possible to set the breakpoint from inside the debugger, by
1732 writing in global variable simgrid::kernel::cfg_breakpoint. For example,
1735 .. code-block:: none
1737 set variable simgrid::kernel::cfg_breakpoint = 3.1416
1739 .. _cfg=debug/verbose-exit:
1744 **Option** ``debug/verbose-exit`` **default:** on
1746 By default, when Ctrl-C is pressed, the status of all existing actors
1747 is displayed before exiting the simulation. This is very useful to
1748 debug your code, but it can become troublesome if you have many
1749 actors. Set this configuration item to **off** to disable this
1752 .. _cfg=exception/cutpath:
1754 Truncate local path from exception backtrace
1755 ............................................
1757 **Option** ``exception/cutpath`` **default:** off
1759 This configuration option is used to remove the path from the
1760 backtrace shown when an exception is thrown. This is mainly useful for
1761 the tests: the full file path would makes the tests non-reproducible because
1762 the paths of source files depend of the build settings. That would
1763 break most of the tests since their output is continually compared.
1767 Logging configuration
1768 ---------------------
1770 As introduced in :ref:`outcome_logs`, the SimGrid logging mechanism allows to configure at runtime the messages that should be displayed and those that should be omitted. Each
1771 message produced in the code is given a category (denoting its topic) and a priority. Then at runtime, each category is given a threshold (only messages of priority higher than
1772 that threshold are displayed), a layout (deciding how the messages in this category are formatted), and an appender (deciding what to do with the message: either print on stderr or
1775 This section explains how to configure this logging features. You can also refer to the documentation of the :ref:`programmer's interface <logging_prog>`, that allows to produce
1776 messages from your code.
1778 Most of the time, the logging mechanism is configured at runtime using the ``--log`` command-line argument, even if you can also use :c:func:`xbt_log_control_set()` to control it from
1779 your program. To pass configure more than one setting, you can either pass several ``--log`` arguments, or separate your settings with spaces, that must be quoted accordingly. In
1780 practice, the following is equivalent to the above settings: ``--log=root.thresh:error --log=s4u_host.thresh:debug``.
1782 If you want to specify more than one setting, you can either pass several ``--log`` argument to your program as above, or separate them with spaces. In this case, you want to quote
1783 your settings, as in ``--log="root.thresh:error s4u_host.thresh:debug"``. The parameters are interpreted in order, from left to right.
1786 Threshold configuration
1787 .......................
1789 The keyword ``threshold`` controls which logging event will get displayed in a given category. For example, ``--log=root.threshold:debug`` displays *every* message produced in the
1790 ``root`` category and its subcategories (i.e., every message produced -- this is *extremely* verbose), while ``--log=root.thres:critical`` turns almost everything off. As you can
1791 see, ``threshold`` can be abbreviated here.
1793 Existing thresholds:
1795 - ``trace`` some functions display a message at this level when entering or returning
1796 - ``debug`` output that is mostly useful when debugging the corresponding module.
1797 - ``verbose`` verbose output that is only mildly interesting and can easily be ignored
1798 - ``info`` usual output (this is the default threshold of all categories)
1799 - ``warning`` minor issue encountered
1800 - ``error`` issue encountered
1801 - ``critical`` major issue encountered, such as assertions failures
1805 Format configuration
1806 ....................
1808 The keyword ``fmt`` controls the layout (the format) of a logging category. For example, ``--log=root.fmt:%m`` reduces the output to the user-message only, removing any decoration such
1809 as the date, or the actor ID, everything. Existing format directives:
1812 - %n: line separator (LOG4J compatible)
1813 - %e: plain old space (SimGrid extension)
1815 - %m: user-provided message
1817 - %c: Category name (LOG4J compatible)
1818 - %p: Priority name (LOG4J compatible)
1820 - %h: Hostname (SimGrid extension)
1821 - %a: Actor name (SimGrid extension -- note that with SMPI this is the integer value of the process rank)
1822 - %i: Actor PID (SimGrid extension -- this is a 'i' as in 'i'dea)
1823 - %t: Thread "name" (LOG4J compatible -- actually the address of the thread in memory)
1825 - %F: file name where the log event was raised (LOG4J compatible)
1826 - %l: location where the log event was raised (LOG4J compatible, like '%%F:%%L' -- this is a l as in 'l'etter)
1827 - %L: line number where the log event was raised (LOG4J compatible)
1828 - %M: function name (LOG4J compatible -- called method name here of course).
1830 - %d: date (UNIX-like epoch)
1831 - %r: application age (time elapsed since the beginning of the application)
1834 ``--log=root.fmt:'[%h:%a:(%i) %r] %l: %m%n'`` gives you the default layout used for info messages while ``--log=root.fmt:'[%h:%a:(%i) %r] %l: [%c/%p] %m%n'`` gives you the default
1835 layout for the other priorities (it adds the source code location). Also, the actor identification is omitted by the default layout for the messages coming directly from the
1836 SimGrid kernel, so info messages are formatted with ``[%r] [%c/%p] %m%n`` in this case. When specifying the layout manually, such distinctions are currently impossible, and the
1837 provided layout is used for every messages.
1839 As with printf, you can specify the precision and width of the fields. For example, ``%.4r`` limits the date precision to four digits while ``%15h`` limits the host name to at most
1843 If you want to have spaces in your log format, you should protect it. Otherwise, SimGrid will consider that this is a space-separated list of several parameters. But you should
1844 also protect it from the shell that also splits command line arguments on spaces. At the end, you should use something such as ``--log="'root.fmt:%l: [%p/%c]: %m%n'"``.
1845 Another option is to use the ``%e`` directive for spaces, as in ``--log=root.fmt:%l:%e[%p/%c]:%e%m%n``.
1850 The keyword ``app`` controls the appended of a logging category. For example ``--log=root.app:file:mylogfile`` redirects every output to the file ``mylogfile``.
1852 With the ``splitfile`` appender, a new file is created when the size of the output reaches the specified size. The format is ``--log=root.app:splitfile:<size>:<file name>``. For
1853 example, ``--log=root.app:splitfile:500:mylog_%`` creates log files of at most 500 bytes, using the names ``mylog_0``, ``mylog_1``, ``mylog_2``, etc.
1855 The ``rollfile`` appender uses one file only, but the file is emptied and recreated when its size reaches the specified maximum. For example, ``--log=root.app:rollfile:500:mylog``
1856 ensures that the log file ``mylog`` will never overpass 500 bytes in size.
1858 Any appender setup this way have its own layout format, that you may change afterward. When specifying a new appender, its additivity is set to false to prevent log event displayed
1859 by this appender to "leak" to any other appender higher in the hierarchy. You can naturally change that if you want your messages to be displayed twice.
1864 The keyword ``add`` controls the additivity of a logging category. By default, the messages are only passed one appender only: the more specific, i.e. the first one found when
1865 climbing the tree from the category in which they were produced. In Log4J parlance, it is said that the default additivity of appenders is false. If you change this setting to
1866 ``on`` (or ``yes`` or ``1``), the produced messages will also be passed to the upper appender.
1868 Let's consider a more complex example: ``--log="root.app:file:all.log s4u.app:file:iface.log xbt.app:file:xbt.log xbt.add:yes``. Here, the logging of s4u will be sent to the
1869 ``iface.log`` file; the logging of the xbt toolbox will be sent to both the ``xbt.log`` file and the ``all.log`` file (because xbt additivity was enabled); and every other loggings
1870 will only be sent to ``all.log``.
1875 ``--help-logs`` displays a complete help message about logging in SimGrid.
1877 ``--help-log-categories`` displays the actual hierarchy of log categories for this binary.
1879 ``--log=no_loc`` hides the source locations (file names and line numbers) from the messages. This is useful to make tests reproducible.