8 <object id="TOC" data="graphical-toc.svg" width="100%" 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:
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 ').
44 Another solution is to use the ``<config>`` tag in the platform file. The
45 only restriction is that this tag must occure 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 e->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 - **clean-atexit:** :ref:`cfg=clean-atexit`
89 - **contexts/factory:** :ref:`cfg=contexts/factory`
90 - **contexts/guard-size:** :ref:`cfg=contexts/guard-size`
91 - **contexts/nthreads:** :ref:`cfg=contexts/nthreads`
92 - **contexts/parallel-threshold:** :ref:`cfg=contexts/parallel-threshold`
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 - **exception/cutpath:** :ref:`cfg=exception/cutpath`
102 - **host/model:** :ref:`options_model_select`
104 - **maxmin/precision:** :ref:`cfg=maxmin/precision`
105 - **maxmin/concurrency-limit:** :ref:`cfg=maxmin/concurrency-limit`
107 - **msg/debug-multiple-use:** :ref:`cfg=msg/debug-multiple-use`
109 - **model-check:** :ref:`options_modelchecking`
110 - **model-check/checkpoint:** :ref:`cfg=model-check/checkpoint`
111 - **model-check/communications-determinism:** :ref:`cfg=model-check/communications-determinism`
112 - **model-check/dot-output:** :ref:`cfg=model-check/dot-output`
113 - **model-check/hash:** :ref:`cfg=model-checker/hash`
114 - **model-check/max-depth:** :ref:`cfg=model-check/max-depth`
115 - **model-check/property:** :ref:`cfg=model-check/property`
116 - **model-check/reduction:** :ref:`cfg=model-check/reduction`
117 - **model-check/replay:** :ref:`cfg=model-check/replay`
118 - **model-check/send-determinism:** :ref:`cfg=model-check/send-determinism`
119 - **model-check/termination:** :ref:`cfg=model-check/termination`
120 - **model-check/timeout:** :ref:`cfg=model-check/timeout`
121 - **model-check/visited:** :ref:`cfg=model-check/visited`
123 - **network/bandwidth-factor:** :ref:`cfg=network/bandwidth-factor`
124 - **network/crosstraffic:** :ref:`cfg=network/crosstraffic`
125 - **network/latency-factor:** :ref:`cfg=network/latency-factor`
126 - **network/maxmin-selective-update:** :ref:`Network Optimization Level <options_model_optim>`
127 - **network/model:** :ref:`options_model_select`
128 - **network/optim:** :ref:`Network Optimization Level <options_model_optim>`
129 - **network/TCP-gamma:** :ref:`cfg=network/TCP-gamma`
130 - **network/weight-S:** :ref:`cfg=network/weight-S`
132 - **ns3/TcpModel:** :ref:`options_pls`
133 - **path:** :ref:`cfg=path`
134 - **plugin:** :ref:`cfg=plugin`
136 - **simix/breakpoint:** :ref:`cfg=simix/breakpoint`
138 - **storage/max_file_descriptors:** :ref:`cfg=storage/max_file_descriptors`
140 - **surf/precision:** :ref:`cfg=surf/precision`
142 - **For collective operations of SMPI,** please refer to Section :ref:`cfg=smpi/coll-selector`
143 - **smpi/async-small-thresh:** :ref:`cfg=smpi/async-small-thresh`
144 - **smpi/bw-factor:** :ref:`cfg=smpi/bw-factor`
145 - **smpi/coll-selector:** :ref:`cfg=smpi/coll-selector`
146 - **smpi/comp-adjustment-file:** :ref:`cfg=smpi/comp-adjustment-file`
147 - **smpi/cpu-threshold:** :ref:`cfg=smpi/cpu-threshold`
148 - **smpi/display-timing:** :ref:`cfg=smpi/display-timing`
149 - **smpi/grow-injected-times:** :ref:`cfg=smpi/grow-injected-times`
150 - **smpi/host-speed:** :ref:`cfg=smpi/host-speed`
151 - **smpi/IB-penalty-factors:** :ref:`cfg=smpi/IB-penalty-factors`
152 - **smpi/iprobe:** :ref:`cfg=smpi/iprobe`
153 - **smpi/iprobe-cpu-usage:** :ref:`cfg=smpi/iprobe-cpu-usage`
154 - **smpi/init:** :ref:`cfg=smpi/init`
155 - **smpi/keep-temps:** :ref:`cfg=smpi/keep-temps`
156 - **smpi/lat-factor:** :ref:`cfg=smpi/lat-factor`
157 - **smpi/ois:** :ref:`cfg=smpi/ois`
158 - **smpi/or:** :ref:`cfg=smpi/or`
159 - **smpi/os:** :ref:`cfg=smpi/os`
160 - **smpi/papi-events:** :ref:`cfg=smpi/papi-events`
161 - **smpi/privatization:** :ref:`cfg=smpi/privatization`
162 - **smpi/privatize-libs:** :ref:`cfg=smpi/privatize-libs`
163 - **smpi/send-is-detached-thresh:** :ref:`cfg=smpi/send-is-detached-thresh`
164 - **smpi/shared-malloc:** :ref:`cfg=smpi/shared-malloc`
165 - **smpi/shared-malloc-hugepage:** :ref:`cfg=smpi/shared-malloc-hugepage`
166 - **smpi/simulate-computation:** :ref:`cfg=smpi/simulate-computation`
167 - **smpi/test:** :ref:`cfg=smpi/test`
168 - **smpi/wtime:** :ref:`cfg=smpi/wtime`
170 - **Tracing configuration options** can be found in Section :ref:`tracing_tracing_options`
172 - **storage/model:** :ref:`options_model_select`
173 - **verbose-exit:** :ref:`cfg=verbose-exit`
175 - **vm/model:** :ref:`options_model_select`
179 Configuring the Platform Models
180 -------------------------------
182 .. _options_model_select:
184 Choosing the Platform Models
185 ............................
187 SimGrid comes with several network, CPU and storage models built in,
188 and you can change the used model at runtime by changing the passed
189 configuration. The three main configuration items are given below.
190 For each of these items, passing the special ``help`` value gives you
191 a short description of all possible values (for example,
192 ``--cfg=network/model:help`` will present all provided network
193 models). Also, ``--help-models`` should provide information about all
194 models for all existing resources.
196 - ``network/model``: specify the used network model. Possible values:
198 - **LV08 (default one):** Realistic network analytic model
199 (slow-start modeled by multiplying latency by 13.01, bandwidth by
200 .97; bottleneck sharing uses a payload of S=20537 for evaluating
201 RTT). Described in `Accuracy Study and Improvement of Network
202 Simulation in the SimGrid Framework
203 <http://mescal.imag.fr/membres/arnaud.legrand/articles/simutools09.pdf>`_.
204 - **Constant:** Simplistic network model where all communication
205 take a constant time (one second). This model provides the lowest
206 realism, but is (marginally) faster.
207 - **SMPI:** Realistic network model specifically tailored for HPC
208 settings (accurate modeling of slow start with correction factors on
209 three intervals: < 1KiB, < 64 KiB, >= 64 KiB). This model can be
210 :ref:`further configured <options_model_network>`.
211 - **IB:** Realistic network model specifically tailored for HPC
212 settings with InfiniBand networks (accurate modeling contention
213 behavior, based on the model explained in `this PhD work
214 <http://mescal.imag.fr/membres/jean-marc.vincent/index.html/PhD/Vienne.pdf>`_.
215 This model can be :ref:`further configured <options_model_network>`.
216 - **CM02:** Legacy network analytic model. Very similar to LV08, but
217 without corrective factors. The timings of small messages are thus
218 poorly modeled. This model is described in `A Network Model for
219 Simulation of Grid Application
220 <https://hal.inria.fr/inria-00071989/document>`_.
221 - **NS3** (only available if you compiled SimGrid accordingly):
222 Use the packet-level network
223 simulators as network models (see :ref:`pls_ns3`).
224 This model can be :ref:`further configured <options_pls>`.
226 - ``cpu/model``: specify the used CPU model. We have only one model
229 - **Cas01:** Simplistic CPU model (time=size/power)
231 - ``host/model``: The host concept is the aggregation of a CPU with a
232 network card. Three models exists, but actually, only 2 of them are
233 interesting. The "compound" one is simply due to the way our
234 internal code is organized, and can easily be ignored. So at the
235 end, you have two host models: The default one allows to aggregate
236 an existing CPU model with an existing network model, but does not
237 allow parallel tasks because these beasts need some collaboration
238 between the network and CPU model. That is why, ptask_07 is used by
239 default when using SimDag.
241 - **default:** Default host model. Currently, CPU:Cas01 and
242 network:LV08 (with cross traffic enabled)
243 - **compound:** Host model that is automatically chosen if
244 you change the network and CPU models
245 - **ptask_L07:** Host model somehow similar to Cas01+CM02 but
246 allowing "parallel tasks", that are intended to model the moldable
247 tasks of the grid scheduling literature.
249 - ``storage/model``: specify the used storage model. Only one model is
251 - ``vm/model``: specify the model for virtual machines. Only one model
254 .. todo: make 'compound' the default host model.
256 .. _options_model_optim:
261 The network and CPU models that are based on lmm_solve (that
262 is, all our analytical models) accept specific optimization
265 - items ``network/optim`` and ``cpu/optim`` (both default to 'Lazy'):
267 - **Lazy:** Lazy action management (partial invalidation in lmm +
268 heap in action remaining).
269 - **TI:** Trace integration. Highly optimized mode when using
270 availability traces (only available for the Cas01 CPU model for
272 - **Full:** Full update of remaining and variables. Slow but may be
273 useful when debugging.
275 - items ``network/maxmin-selective-update`` and
276 ``cpu/maxmin-selective-update``: configure whether the underlying
277 should be lazily updated or not. It should have no impact on the
278 computed timings, but should speed up the computation. |br| It is
279 still possible to disable this feature because it can reveal
280 counter-productive in very specific scenarios where the
281 interaction level is high. In particular, if all your
282 communication share a given backbone link, you should disable it:
283 without it, a simple regular loop is used to update each
284 communication. With it, each of them is still updated (because of
285 the dependency induced by the backbone), but through a complicated
286 and slow pattern that follows the actual dependencies.
288 .. _cfg=maxmin/precision:
289 .. _cfg=surf/precision:
294 **Option** ``maxmin/precision`` **Default:** 0.00001 (in flops or bytes) |br|
295 **Option** ``surf/precision`` **Default:** 0.00001 (in seconds)
297 The analytical models handle a lot of floating point values. It is
298 possible to change the epsilon used to update and compare them through
299 this configuration item. Changing it may speedup the simulation by
300 discarding very small actions, at the price of a reduced numerical
301 precision. You can modify separately the precision used to manipulate
302 timings (in seconds) and the one used to manipulate amounts of work
305 .. _cfg=maxmin/concurrency-limit:
310 **Option** ``maxmin/concurrency-limit`` **Default:** -1 (no limit)
312 The maximum number of variables per resource can be tuned through this
313 option. You can have as many simultaneous actions per resources as you
314 want. If your simulation presents a very high level of concurrency, it
315 may help to use e.g. 100 as a value here. It means that at most 100
316 actions can consume a resource at a given time. The extraneous actions
317 are queued and wait until the amount of concurrency of the considered
318 resource lowers under the given boundary.
320 Such limitations help both to the simulation speed and simulation accuracy
321 on highly constrained scenarios, but the simulation speed suffers of this
322 setting on regular (less constrained) scenarios so it is off by default.
324 .. _options_model_network:
326 Configuring the Network Model
327 .............................
329 .. _cfg=network/TCP-gamma:
331 Maximal TCP Window Size
332 ^^^^^^^^^^^^^^^^^^^^^^^
334 **Option** ``network/TCP-gamma`` **Default:** 4194304
336 The analytical models need to know the maximal TCP window size to take
337 the TCP congestion mechanism into account. On Linux, this value can
338 be retrieved using the following commands. Both give a set of values,
339 and you should use the last one, which is the maximal size.
341 .. code-block:: shell
343 cat /proc/sys/net/ipv4/tcp_rmem # gives the sender window
344 cat /proc/sys/net/ipv4/tcp_wmem # gives the receiver window
346 .. _cfg=smpi/IB-penalty-factors:
347 .. _cfg=network/bandwidth-factor:
348 .. _cfg=network/latency-factor:
349 .. _cfg=network/weight-S:
351 Correcting Important Network Parameters
352 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
354 SimGrid can take network irregularities such as a slow startup or
355 changing behavior depending on the message size into account. You
356 should not change these values unless you really know what you're
357 doing. The corresponding values were computed through data fitting
358 one the timings of packet-level simulators, as described in `Accuracy
359 Study and Improvement of Network Simulation in the SimGrid Framework
360 <http://mescal.imag.fr/membres/arnaud.legrand/articles/simutools09.pdf>`_.
363 If you are using the SMPI model, these correction coefficients are
364 themselves corrected by constant values depending on the size of the
365 exchange. By default SMPI uses factors computed on the Stampede
366 Supercomputer at TACC, with optimal deployment of processes on
367 nodes. Again, only hardcore experts should bother about this fact.
369 InfiniBand network behavior can be modeled through 3 parameters
370 ``smpi/IB-penalty-factors:"βe;βs;γs"``, as explained in `this PhD
372 <http://mescal.imag.fr/membres/jean-marc.vincent/index.html/PhD/Vienne.pdf>`_.
374 .. todo:: This section should be rewritten, and actually explain the
375 options network/bandwidth-factor, network/latency-factor,
378 .. _cfg=network/crosstraffic:
380 Simulating Cross-Traffic
381 ^^^^^^^^^^^^^^^^^^^^^^^^
383 Since SimGrid v3.7, cross-traffic effects can be taken into account in
384 analytical simulations. It means that ongoing and incoming
385 communication flows are treated independently. In addition, the LV08
386 model adds 0.05 of usage on the opposite direction for each new
387 created flow. This can be useful to simulate some important TCP
388 phenomena such as ack compression.
390 For that to work, your platform must have two links for each
391 pair of interconnected hosts. An example of usable platform is
392 available in ``examples/platforms/crosstraffic.xml``.
394 This is activated through the ``network/crosstraffic`` item, that
395 can be set to 0 (disable this feature) or 1 (enable it).
397 Note that with the default host model this option is activated by default.
399 .. _cfg=smpi/async-small-thresh:
401 Simulating Asyncronous Send
402 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
404 (this configuration item is experimental and may change or disapear)
406 It is possible to specify that messages below a certain size will be
407 sent as soon as the call to MPI_Send is issued, without waiting for
408 the correspondant receive. This threshold can be configured through
409 the ``smpi/async-small-thresh`` item. The default value is 0. This
410 behavior can also be manually set for mailboxes, by setting the
411 receiving mode of the mailbox with a call to
412 :cpp:func:`MSG_mailbox_set_async`. After this, all messages sent to
413 this mailbox will have this behavior regardless of the message size.
415 This value needs to be smaller than or equals to the threshold set at
416 @ref options_model_smpi_detached , because asynchronous messages are
417 meant to be detached as well.
424 **Option** ``ns3/TcpModel`` **Default:** "default" (NS3 default)
426 When using NS3, there is an extra item ``ns3/TcpModel``, corresponding
427 to the ``ns3::TcpL4Protocol::SocketType`` configuration item in
428 NS3. The only valid values (enforced on the SimGrid side) are
429 'default' (no change to the NS3 configuration), 'NewReno' or 'Reno' or
432 Configuring the Storage model
433 .............................
435 .. _cfg=storage/max_file_descriptors:
437 File Descriptor Cound per Host
438 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
440 **Option** ``storage/max_file_descriptors`` **Default:** 1024
442 Each host maintains a fixed-size array of its file descriptors. You
443 can change its size through this item to either enlarge it if your
444 application requires it or to reduce it to save memory space.
451 SimGrid plugins allow to extend the framework without changing its
452 source code directly. Read the source code of the existing plugins to
453 learn how to do so (in ``src/plugins``), and ask your questions to the
454 usual channels (Stack Overflow, Mailing list, IRC). The basic idea is
455 that plugins usually register callbacks to some signals of interest.
456 If they need to store some information about a given object (Link, CPU
457 or Actor), they do so through the use of a dedicated object extension.
459 Some of the existing plugins can be activated from the command line,
460 meaning that you can activate them from the command line without any
461 modification to your simulation code. For example, you can activate
462 the host energy plugin by adding ``--cfg=plugin:host_energy`` to your
465 Here is the full list of plugins that can be activated this way:
467 - **host_energy:** keeps track of the energy dissipated by
468 computations. More details in @ref plugin_energy.
469 - **link_energy:** keeps track of the energy dissipated by
470 communications. More details in @ref SURF_plugin_energy.
471 - **host_load:** keeps track of the computational load.
472 More details in @ref plugin_load.
474 .. _options_modelchecking:
476 Configuring the Model-Checking
477 ------------------------------
479 To enable the SimGrid model-checking support the program should
480 be executed using the simgrid-mc wrapper:
482 .. code-block:: shell
484 simgrid-mc ./my_program
486 Safety properties are expressed as assertions using the function
487 :cpp:func:`void MC_assert(int prop)`.
489 .. _cfg=model-check/property:
491 Specifying a liveness property
492 ..............................
494 **Option** ``model-check/property`` **Default:** unset
496 If you want to specify liveness properties, you have to pass them on
497 the command line, specifying the name of the file containing the
498 property, as formatted by the ltl2ba program.
501 .. code-block:: shell
503 simgrid-mc ./my_program --cfg=model-check/property:<filename>
505 .. _cfg=model-check/checkpoint:
507 Going for Stateful Verification
508 ...............................
510 By default, the system is backtracked to its initial state to explore
511 another path instead of backtracking to the exact step before the fork
512 that we want to explore (this is called stateless verification). This
513 is done this way because saving intermediate states can rapidly
514 exhaust the available memory. If you want, you can change the value of
515 the ``model-check/checkpoint`` item. For example,
516 ``--cfg=model-check/checkpoint:1`` asks to take a checkpoint every
517 step. Beware, this will certainly explode your memory. Larger values
518 are probably better, make sure to experiment a bit to find the right
519 setting for your specific system.
521 .. _cfg=model-check/reduction:
523 Specifying the kind of reduction
524 ................................
526 The main issue when using the model-checking is the state space
527 explosion. To counter that problem, you can chose a exploration
528 reduction techniques with
529 ``--cfg=model-check/reduction:<technique>``. For now, this
530 configuration variable can take 2 values:
532 - **none:** Do not apply any kind of reduction (mandatory for now for
534 - **dpor:** Apply Dynamic Partial Ordering Reduction. Only valid if
535 you verify local safety properties (default value for safety
538 There is unfortunately no silver bullet here, and the most efficient
539 reduction techniques cannot be applied to any properties. In
540 particular, the DPOR method cannot be applied on liveness properties
541 since our implementation of DPOR may break some cycles, while cycles
542 are very important to the soundness of the exploration for liveness
545 .. _cfg=model-check/visited:
547 Size of Cycle Detection Set
548 ...........................
550 In order to detect cycles, the model-checker needs to check if a new
551 explored state is in fact the same state than a previous one. For
552 that, the model-checker can take a snapshot of each visited state:
553 this snapshot is then used to compare it with subsequent states in the
556 The ``model-check/visited`` item is the maximum number of states which
557 are stored in memory. If the maximum number of snapshotted state is
558 reached, some states will be removed from the memory and some cycles
559 might be missed. Small values can lead to incorrect verifications, but
560 large value can exhaust your memory, so choose carefully.
562 By default, no state is snapshotted and cycles cannot be detected.
564 .. _cfg=model-check/termination:
566 Non-Termination Detection
567 .........................
569 The ``model-check/termination`` configuration item can be used to
570 report if a non-termination execution path has been found. This is a
571 path with a cycle which means that the program might never terminate.
573 This only works in safety mode, not in liveness mode.
575 This options is disabled by default.
577 .. _cfg=model-check/dot-output:
582 If set, the ``model-check/dot-output`` configuration item is the name
583 of a file in which to write a dot file of the path leading the found
584 property (safety or liveness violation) as well as the cycle for
585 liveness properties. This dot file can then fed to the graphviz dot
586 tool to generate an corresponding graphical representation.
588 .. _cfg=model-check/max-depth:
590 Exploration Depth Limit
591 .......................
593 The ``model-checker/max-depth`` can set the maximum depth of the
594 exploration graph of the model-checker. If this limit is reached, a
595 logging message is sent and the results might not be exact.
597 By default, there is not depth limit.
599 .. _cfg=model-check/timeout:
604 By default, the model-checker does not handle timeout conditions: the `wait`
605 operations never time out. With the ``model-check/timeout`` configuration item
606 set to **yes**, the model-checker will explore timeouts of `wait` operations.
608 .. _cfg=model-check/communications-determinism:
609 .. _cfg=model-check/send-determinism:
611 Communication Determinism
612 .........................
614 The ``model-check/communications-determinism`` and
615 ``model-check/send-determinism`` items can be used to select the
616 communication determinism mode of the model-checker which checks
617 determinism properties of the communications of an application.
619 Verification Performance Considerations
620 .......................................
622 The size of the stacks can have a huge impact on the memory
623 consumption when using model-checking. By default, each snapshot will
624 save a copy of the whole stacks and not only of the part which is
625 really meaningful: you should expect the contribution of the memory
626 consumption of the snapshots to be @f$ @mbox{number of processes}
627 @times @mbox{stack size} @times @mbox{number of states} @f$.
629 When compiled against the model checker, the stacks are not
630 protected with guards: if the stack size is too small for your
631 application, the stack will silently overflow on other parts of the
632 memory (see :ref:`contexts/guard-size <cfg=contexts/guard-size>`).
634 .. _cfg=model-checker/hash:
639 Usually most of the time of the model-checker is spent comparing states. This
640 process is complicated and consumes a lot of bandwidth and cache.
641 In order to speedup the state comparison, the experimental ``model-checker/hash``
642 configuration item enables the computation of a hash summarizing as much
643 information of the state as possible into a single value. This hash can be used
644 to avoid most of the comparisons: the costly comparison is then only used when
645 the hashes are identical.
647 Currently most of the state is not included in the hash because the
648 implementation was found to be buggy and this options is not as useful as
649 it could be. For this reason, it is currently disabled by default.
651 .. _cfg=model-check/replay:
653 Replaying buggy execution paths out of the model-checker
654 ........................................................
656 Debugging the problems reported by the model-checker is challenging: First, the
657 application under verification cannot be debugged with gdb because the
658 model-checker already traces it. Then, the model-checker may explore several
659 execution paths before encountering the issue, making it very difficult to
660 understand the outputs. Fortunately, SimGrid provides the execution path leading
661 to any reported issue so that you can replay this path out of the model checker,
662 enabling the usage of classical debugging tools.
664 When the model-checker finds an interesting path in the application
665 execution graph (where a safety or liveness property is violated), it
666 generates an identifier for this path. Here is an example of output:
668 .. code-block:: shell
670 [ 0.000000] (0:@) Check a safety property
671 [ 0.000000] (0:@) **************************
672 [ 0.000000] (0:@) *** PROPERTY NOT VALID ***
673 [ 0.000000] (0:@) **************************
674 [ 0.000000] (0:@) Counter-example execution trace:
675 [ 0.000000] (0:@) [(1)Tremblay (app)] MC_RANDOM(3)
676 [ 0.000000] (0:@) [(1)Tremblay (app)] MC_RANDOM(4)
677 [ 0.000000] (0:@) Path = 1/3;1/4
678 [ 0.000000] (0:@) Expanded states = 27
679 [ 0.000000] (0:@) Visited states = 68
680 [ 0.000000] (0:@) Executed transitions = 46
682 The interesting line is ``Path = 1/3;1/4``, which means that you should use
683 `--cfg=model-check/replay:1/3;1/4`` to replay your application on the buggy
684 execution path. The other options should be the same (but the model-checker
685 should be disabled). Note that format and meaning of the path may change between
688 Configuring the User Code Virtualization
689 ----------------------------------------
691 .. _cfg=contexts/factory:
693 Selecting the Virtualization Factory
694 ....................................
696 **Option** contexts/factory **Default:** "raw"
698 In SimGrid, the user code is virtualized in a specific mechanism that
699 allows the simulation kernel to control its execution: when a user
700 process requires a blocking action (such as sending a message), it is
701 interrupted, and only gets released when the simulated clock reaches
702 the point where the blocking operation is done. This is explained
703 graphically in the `relevant tutorial, available online
704 <https://simgrid.org/tutorials/simgrid-simix-101.pdf>`_.
706 In SimGrid, the containers in which user processes are virtualized are
707 called contexts. Several context factory are provided, and you can
708 select the one you want to use with the ``contexts/factory``
709 configuration item. Some of the following may not exist on your
710 machine because of portability issues. In any case, the default one
711 should be the most effcient one (please report bugs if the
712 auto-detection fails for you). They are approximately sorted here from
713 the slowest to the most efficient:
715 - **thread:** very slow factory using full featured threads (either
716 pthreads or windows native threads). They are slow but very
717 standard. Some debuggers or profilers only work with this factory.
718 - **java:** Java applications are virtualized onto java threads (that
719 are regular pthreads registered to the JVM)
720 - **ucontext:** fast factory using System V contexts (Linux and FreeBSD only)
721 - **boost:** This uses the `context
722 implementation <http://www.boost.org/doc/libs/1_59_0/libs/context/doc/html/index.html>`_
723 of the boost library for a performance that is comparable to our
725 |br| Install the relevant library (e.g. with the
726 libboost-contexts-dev package on Debian/Ubuntu) and recompile
728 - **raw:** amazingly fast factory using a context switching mechanism
729 of our own, directly implemented in assembly (only available for x86
730 and amd64 platforms for now) and without any unneeded system call.
732 The main reason to change this setting is when the debugging tools get
733 fooled by the optimized context factories. Threads are the most
734 debugging-friendly contextes, as they allow to set breakpoints
735 anywhere with gdb and visualize backtraces for all processes, in order
736 to debug concurrency issues. Valgrind is also more comfortable with
737 threads, but it should be usable with all factories (Exception: the
738 callgrind tool really dislikes raw and ucontext factories).
740 .. _cfg=contexts/stack-size:
742 Adapting the Stack Size
743 .......................
745 **Option** ``contexts/stack-size`` **Default:** 8192 KiB
747 Each virtualized used process is executed using a specific system
748 stack. The size of this stack has a huge impact on the simulation
749 scalability, but its default value is rather large. This is because
750 the error messages that you get when the stack size is too small are
751 rather disturbing: this leads to stack overflow (overwriting other
752 stacks), leading to segfaults with corrupted stack traces.
754 If you want to push the scalability limits of your code, you might
755 want to reduce the ``contexts/stack-size`` item. Its default value is
756 8192 (in KiB), while our Chord simulation works with stacks as small
757 as 16 KiB, for example. This *setting is ignored* when using the
758 thread factory. Instead, you should compile SimGrid and your
759 application with ``-fsplit-stack``. Note that this compilation flag is
760 not compatible with the model-checker right now.
762 The operating system should only allocate memory for the pages of the
763 stack which are actually used and you might not need to use this in
764 most cases. However, this setting is very important when using the
765 model checker (see :ref:`options_mc_perf`).
767 .. _cfg=contexts/guard-size:
769 Disabling Stack Guard Pages
770 ...........................
772 **Option** ``contexts/guard-size`` **Default** 1 page in most case (0 pages on Windows or with MC)
774 Unless you use the threads context factory (see
775 :ref:`cfg=contexts/factory`), a stack guard page is usually used
776 which prevents the stack of a given actor from overflowing on another
777 stack. But the performance impact may become prohibitive when the
778 amount of actors increases. The option ``contexts/guard-size`` is the
779 number of stack guard pages used. By setting it to 0, no guard pages
780 will be used: in this case, you should avoid using small stacks (with
781 :ref:`contexts/stack-size <cfg=contexts/stack-size>`) as the stack
782 will silently overflow on other parts of the memory.
784 When no stack guard page is created, stacks may then silently overflow
785 on other parts of the memory if their size is too small for the
788 .. _cfg=contexts/nthreads:
789 .. _cfg=contexts/parallel-threshold:
790 .. _cfg=contexts/synchro:
792 Running User Code in Parallel
793 .............................
795 Parallel execution of the user code is only considered stable in
796 SimGrid v3.7 and higher, and mostly for MSG simulations. SMPI
797 simulations may well fail in parallel mode. It is described in
798 `INRIA RR-7653 <http://hal.inria.fr/inria-00602216/>`_.
800 If you are using the **ucontext** or **raw** context factories, you can
801 request to execute the user code in parallel. Several threads are
802 launched, each of them handling as much user contexts at each run. To
803 actiave this, set the ``contexts/nthreads`` item to the amount of
804 cores that you have in your computer (or lower than 1 to have
805 the amount of cores auto-detected).
807 Even if you asked several worker threads using the previous option,
808 you can request to start the parallel execution (and pay the
809 associated synchronization costs) only if the potential parallelism is
810 large enough. For that, set the ``contexts/parallel-threshold``
811 item to the minimal amount of user contexts needed to start the
812 parallel execution. In any given simulation round, if that amount is
813 not reached, the contexts will be run sequentially directly by the
814 main thread (thus saving the synchronization costs). Note that this
815 option is mainly useful when the grain of the user code is very fine,
816 because our synchronization is now very efficient.
818 When parallel execution is activated, you can choose the
819 synchronization schema used with the ``contexts/synchro`` item,
820 which value is either:
822 - **futex:** ultra optimized synchronisation schema, based on futexes
823 (fast user-mode mutexes), and thus only available on Linux systems.
824 This is the default mode when available.
825 - **posix:** slow but portable synchronisation using only POSIX
827 - **busy_wait:** not really a synchronisation: the worker threads
828 constantly request new contexts to execute. It should be the most
829 efficient synchronisation schema, but it loads all the cores of
830 your machine for no good reason. You probably prefer the other less
833 Configuring the Tracing
834 -----------------------
836 The :ref:`tracing subsystem <outcomes_vizu>` can be configured in
837 several different ways depending on the nature of the simulator (MSG,
838 SimDag, SMPI) and the kind of traces that need to be obtained. See the
839 :ref:`Tracing Configuration Options subsection
840 <tracing_tracing_options>` to get a detailed description of each
841 configuration option.
843 We detail here a simple way to get the traces working for you, even if
844 you never used the tracing API.
847 - Any SimGrid-based simulator (MSG, SimDag, SMPI, ...) and raw traces:
849 .. code-block:: shell
851 --cfg=tracing:yes --cfg=tracing/uncategorized:yes --cfg=triva/uncategorized:uncat.plist
853 The first parameter activates the tracing subsystem, the second
854 tells it to trace host and link utilization (without any
855 categorization) and the third creates a graph configuration file to
856 configure Triva when analysing the resulting trace file.
858 - MSG or SimDag-based simulator and categorized traces (you need to
859 declare categories and classify your tasks according to them)
861 .. code-block:: shell
863 --cfg=tracing:yes --cfg=tracing/categorized:yes --cfg=triva/categorized:cat.plist
865 The first parameter activates the tracing subsystem, the second
866 tells it to trace host and link categorized utilization and the
867 third creates a graph configuration file to configure Triva when
868 analysing the resulting trace file.
870 - SMPI simulator and traces for a space/time view:
872 .. code-block:: shell
876 The `-trace` parameter for the smpirun script runs the simulation
877 with ``--cfg=tracing:yes --cfg=tracing/smpi:yes``. Check the
878 smpirun's `-help` parameter for additional tracing options.
880 Sometimes you might want to put additional information on the trace to
881 correctly identify them later, or to provide data that can be used to
882 reproduce an experiment. You have two ways to do that:
884 - Add a string on top of the trace file as comment:
886 .. code-block:: shell
888 --cfg=tracing/comment:my_simulation_identifier
890 - Add the contents of a textual file on top of the trace file as comment:
892 .. code-block:: shell
894 --cfg=tracing/comment-file:my_file_with_additional_information.txt
896 Please, use these two parameters (for comments) to make reproducible
897 simulations. For additional details about this and all tracing
898 options, check See the :ref:`tracing_tracing_options`.
903 .. _cfg=msg/debug-multiple-use:
908 **Option** ``msg/debug-multiple-use`` **Default:** off
910 Sometimes your application may try to send a task that is still being
911 executed somewhere else, making it impossible to send this task. However,
912 for debugging purposes, one may want to know what the other host is/was
913 doing. This option shows a backtrace of the other process.
918 The SMPI interface provides several specific configuration items.
919 These are uneasy to see since the code is usually launched through the
920 ``smiprun`` script directly.
922 .. _cfg=smpi/host-speed:
923 .. _cfg=smpi/cpu-threshold:
924 .. _cfg=smpi/simulate-computation:
926 Automatic Benchmarking of SMPI Code
927 ...................................
929 In SMPI, the sequential code is automatically benchmarked, and these
930 computations are automatically reported to the simulator. That is to
931 say that if you have a large computation between a ``MPI_Recv()`` and
932 a ``MPI_Send()``, SMPI will automatically benchmark the duration of
933 this code, and create an execution task within the simulator to take
934 this into account. For that, the actual duration is measured on the
935 host machine and then scaled to the power of the corresponding
936 simulated machine. The variable ``smpi/host-speed`` allows to specify
937 the computational speed of the host machine (in flop/s) to use when
938 scaling the execution times. It defaults to 20000, but you really want
939 to update it to get accurate simulation results.
941 When the code is constituted of numerous consecutive MPI calls, the
942 previous mechanism feeds the simulation kernel with numerous tiny
943 computations. The ``smpi/cpu-threshold`` item becomes handy when this
944 impacts badly the simulation performance. It specifies a threshold (in
945 seconds) below which the execution chunks are not reported to the
946 simulation kernel (default value: 1e-6).
948 .. note:: The option ``smpi/cpu-threshold`` ignores any computation
949 time spent below this threshold. SMPI does not consider the
950 `amount` of these computations; there is no offset for this. Hence,
951 a value that is too small, may lead to unreliable simulation
954 In some cases, however, one may wish to disable simulation of
955 application computation. This is the case when SMPI is used not to
956 simulate an MPI applications, but instead an MPI code that performs
957 "live replay" of another MPI app (e.g., ScalaTrace's replay tool,
958 various on-line simulators that run an app at scale). In this case the
959 computation of the replay/simulation logic should not be simulated by
960 SMPI. Instead, the replay tool or on-line simulator will issue
961 "computation events", which correspond to the actual MPI simulation
962 being replayed/simulated. At the moment, these computation events can
963 be simulated using SMPI by calling internal smpi_execute*() functions.
965 To disable the benchmarking/simulation of computation in the simulated
966 application, the variable ``smpi/simulate-computation`` should be set
967 to no. This option just ignores the timings in your simulation; it
968 still executes the computations itself. If you want to stop SMPI from
969 doing that, you should check the SMPI_SAMPLE macros, documented in
970 Section :ref:`SMPI_adapting_speed`.
972 +------------------------------------+-------------------------+-----------------------------+
973 | Solution | Computations executed? | Computations simulated? |
974 +====================================+=========================+=============================+
975 | --cfg=smpi/simulate-computation:no | Yes | Never |
976 +------------------------------------+-------------------------+-----------------------------+
977 | --cfg=smpi/cpu-threshold:42 | Yes, in all cases | If it lasts over 42 seconds |
978 +------------------------------------+-------------------------+-----------------------------+
979 | SMPI_SAMPLE() macro | Only once per loop nest | Always |
980 +------------------------------------+-------------------------+-----------------------------+
982 .. _cfg=smpi/comp-adjustment-file:
984 Slow-down or speed-up parts of your code
985 ........................................
987 **Option** ``smpi/comp-adjustment-file:`` **Default:** unset
989 This option allows you to pass a file that contains two columns: The
990 first column defines the section that will be subject to a speedup;
991 the second column is the speedup. For instance:
993 .. code-block:: shell
996 "exchange_1.f:30:exchange_1.f:130",1.18244559422142
998 The first line is the header - you must include it. The following
999 line means that the code between two consecutive MPI calls on line 30
1000 in exchange_1.f and line 130 in exchange_1.f should receive a speedup
1001 of 1.18244559422142. The value for the second column is therefore a
1002 speedup, if it is larger than 1 and a slow-down if it is smaller
1003 than 1. Nothing will be changed if it is equal to 1.
1005 Of course, you can set any arbitrary filenames you want (so the start
1006 and end don't have to be in the same file), but be aware that this
1007 mechanism only supports `consecutive calls!`
1009 Please note that you must pass the ``-trace-call-location`` flag to
1010 smpicc or smpiff, respectively. This flag activates some internal
1011 macro definitions that help with obtaining the call location.
1013 .. _cfg=smpi/bw-factor:
1018 **Option** ``smpi/bw-factor``
1019 |br| **Default:** 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
1021 The possible throughput of network links is often dependent on the
1022 message sizes, as protocols may adapt to different message sizes. With
1023 this option, a series of message sizes and factors are given, helping
1024 the simulation to be more realistic. For instance, the current default
1025 value means that messages with size 65472 and more will get a total of
1026 MAX_BANDWIDTH*0.940694, messages of size 15424 to 65471 will get
1027 MAX_BANDWIDTH*0.697866 and so on (where MAX_BANDWIDTH denotes the
1028 bandwidth of the link).
1030 An experimental script to compute these factors is available online. See
1031 https://framagit.org/simgrid/platform-calibration/
1032 https://simgrid.org/contrib/smpi-saturation-doc.html
1034 .. _cfg=smpi/display-timing:
1036 Reporting Simulation Time
1037 .........................
1039 **Option** ``smpi/display-timing`` **Default:** 0 (false)
1041 Most of the time, you run MPI code with SMPI to compute the time it
1042 would take to run it on a platform. But since the code is run through
1043 the ``smpirun`` script, you don't have any control on the launcher
1044 code, making it difficult to report the simulated time when the
1045 simulation ends. If you enable the ``smpi/display-timing`` item,
1046 ``smpirun`` will display this information when the simulation
1049 .. _cfg=smpi/keep-temps:
1051 Keeping temporary files after simulation
1052 ........................................
1054 **Option** ``smpi/keep-temps`` **default:** 0 (false)
1056 SMPI usually generates a lot of temporary files that are cleaned after
1057 use. This option request to preserve them, for example to debug or
1058 profile your code. Indeed, the binary files are removed very early
1059 under the dlopen privatization schema, which tend to fool the
1062 .. _cfg=smpi/lat-factor:
1067 **Option** ``smpi/lat-factor`` |br|
1068 **default:** 65472:11.6436;15424:3.48845;9376:2.59299;5776:2.18796;3484:1.88101;1426:1.61075;732:1.9503;257:1.95341;0:2.01467
1070 The motivation and syntax for this option is identical to the motivation/syntax
1071 of :ref:`cfg=smpi/bw-factor`.
1073 There is an important difference, though: While smpi/bw-factor `reduces` the
1074 actual bandwidth (i.e., values between 0 and 1 are valid), latency factors
1075 increase the latency, i.e., values larger than or equal to 1 are valid here.
1077 .. _cfg=smpi/papi-events:
1079 Trace hardware counters with PAPI
1080 .................................
1082 **Option** ``smpi/papi-events`` **default:** unset
1084 When the PAPI support was compiled in SimGrid, this option takes the
1085 names of PAPI counters and adds their respective values to the trace
1086 files (See Section :ref:`tracing_tracing_options`).
1090 This feature currently requires superuser privileges, as registers
1091 are queried. Only use this feature with code you trust! Call
1092 smpirun for instance via ``smpirun -wrapper "sudo "
1093 <your-parameters>`` or run ``sudo sh -c "echo 0 >
1094 /proc/sys/kernel/perf_event_paranoid"`` In the later case, sudo
1095 will not be required.
1097 It is planned to make this feature available on a per-process (or per-thread?) basis.
1098 The first draft, however, just implements a "global" (i.e., for all processes) set
1099 of counters, the "default" set.
1101 .. code-block:: shell
1103 --cfg=smpi/papi-events:"default:PAPI_L3_LDM:PAPI_L2_LDM"
1105 .. _cfg=smpi/privatization:
1107 Automatic Privatization of Global Variables
1108 ...........................................
1110 **Option** ``smpi/privatization`` **default:** "dlopen" (when using smpirun)
1112 MPI executables are usually meant to be executed in separated
1113 processes, but SMPI is executed in only one process. Global variables
1114 from executables will be placed in the same memory zone and shared
1115 between processes, causing intricate bugs. Several options are
1116 possible to avoid this, as described in the main `SMPI publication
1117 <https://hal.inria.fr/hal-01415484>`_ and in the :ref:`SMPI
1118 documentation <SMPI_what_globals>`. SimGrid provides two ways of
1119 automatically privatizing the globals, and this option allows to
1120 choose between them.
1122 - **no** (default when not using smpirun): Do not automatically
1123 privatize variables. Pass ``-no-privatize`` to smpirun to disable
1125 - **dlopen** or **yes** (default when using smpirun): Link multiple
1126 times against the binary.
1127 - **mmap** (slower, but maybe somewhat more stable):
1128 Runtime automatic switching of the data segments.
1131 This configuration option cannot be set in your platform file. You can only
1132 pass it as an argument to smpirun.
1134 .. _cfg=smpi/privatize-libs:
1136 Automatic privatization of global variables inside external libraries
1137 .....................................................................
1139 **Option** ``smpi/privatize-libs`` **default:** unset
1141 **Linux/BSD only:** When using dlopen (default) privatization,
1142 privatize specific shared libraries with internal global variables, if
1143 they can't be linked statically. For example libgfortran is usually
1144 used for Fortran I/O and indexes in files can be mixed up.
1146 Multiple libraries can be given, semicolon separated.
1148 This configuration option can only use either full paths to libraries,
1149 or full names. Check with ldd the name of the library you want to
1152 .. code-block:: shell
1156 libgfortran.so.3 => /usr/lib/x86_64-linux-gnu/libgfortran.so.3 (0x00007fbb4d91b000)
1159 Then you can use ``--cfg=smpi/privatize-libs:libgfortran.so.3``
1160 or ``--cfg=smpi/privatize-libs:/usr/lib/x86_64-linux-gnu/libgfortran.so.3``,
1161 but not ``libgfortran`` nor ``libgfortran.so``.
1163 .. _cfg=smpi/send-is-detached-thresh:
1165 Simulating MPI detached send
1166 ............................
1168 **Option** ``smpi/send-is-detached-thresh`` **default:** 65536
1170 This threshold specifies the size in bytes under which the send will
1171 return immediately. This is different from the threshold detailed in
1172 :ref:`options_model_network_asyncsend` because the message is not
1173 effectively sent when the send is posted. SMPI still waits for the
1174 correspondant receive to be posted to perform the communication
1177 .. _cfg=smpi/coll-selector:
1179 Simulating MPI collective algorithms
1180 ....................................
1182 **Option** ``smpi/coll-selector`` **Possible values:** naive (default), ompi, mpich
1184 SMPI implements more than 100 different algorithms for MPI collective
1185 communication, to accurately simulate the behavior of most of the
1186 existing MPI libraries. The ``smpi/coll-selector`` item can be used to
1187 use the decision logic of either OpenMPI or MPICH libraries (by
1188 default SMPI uses naive version of collective operations).
1190 Each collective operation can be manually selected with a
1191 ``smpi/collective_name:algo_name``. Available algorithms are listed in
1192 :ref:`SMPI_use_colls`.
1194 .. TODO:: All available collective algorithms will be made available
1195 via the ``smpirun --help-coll`` command.
1197 .. _cfg=smpi/iprobe:
1199 Inject constant times for MPI_Iprobe
1200 ....................................
1202 **Option** ``smpi/iprobe`` **default:** 0.0001
1204 The behavior and motivation for this configuration option is identical
1205 with :ref:`smpi/test <cfg=smpi/test>`, but for the function
1208 .. _cfg=smpi/iprobe-cpu-usage:
1210 Reduce speed for iprobe calls
1211 .............................
1213 **Option** ``smpi/iprobe-cpu-usage`` **default:** 1 (no change)
1215 MPI_Iprobe calls can be heavily used in applications. To account
1216 correctly for the energy cores spend probing, it is necessary to
1217 reduce the load that these calls cause inside SimGrid.
1219 For instance, we measured a max power consumption of 220 W for a
1220 particular application but only 180 W while this application was
1221 probing. Hence, the correct factor that should be passed to this
1222 option would be 180/220 = 0.81.
1226 Inject constant times for MPI_Init
1227 ..................................
1229 **Option** ``smpi/init`` **default:** 0
1231 The behavior and motivation for this configuration option is identical
1232 with :ref:`smpi/test <cfg=smpi/test>`, but for the function ``MPI_Init()``.
1236 Inject constant times for MPI_Isend()
1237 .....................................
1239 **Option** ``smpi/ois``
1241 The behavior and motivation for this configuration option is identical
1242 with :ref:`smpi/os <cfg=smpi/os>`, but for the function ``MPI_Isend()``.
1246 Inject constant times for MPI_send()
1247 ....................................
1249 **Option** ``smpi/os``
1251 In several network models such as LogP, send (MPI_Send, MPI_Isend) and
1252 receive (MPI_Recv) operations incur costs (i.e., they consume CPU
1253 time). SMPI can factor these costs in as well, but the user has to
1254 configure SMPI accordingly as these values may vary by machine. This
1255 can be done by using ``smpi/os`` for MPI_Send operations; for MPI_Isend
1256 and MPI_Recv, use ``smpi/ois`` and ``smpi/or``, respectively. These work
1257 exactly as ``smpi/ois``.
1259 This item can consist of multiple sections; each section takes three
1260 values, for example ``1:3:2;10:5:1``. The sections are divided by ";"
1261 so this example contains two sections. Furthermore, each section
1262 consists of three values.
1264 1. The first value denotes the minimum size for this section to take effect;
1265 read it as "if message size is greater than this value (and other section has a larger
1266 first value that is also smaller than the message size), use this".
1267 In the first section above, this value is "1".
1269 2. The second value is the startup time; this is a constant value that will always
1270 be charged, no matter what the size of the message. In the first section above,
1273 3. The third value is the `per-byte` cost. That is, it is charged for every
1274 byte of the message (incurring cost messageSize*cost_per_byte)
1275 and hence accounts also for larger messages. In the first
1276 section of the example above, this value is "2".
1278 Now, SMPI always checks which section it should take for a given
1279 message; that is, if a message of size 11 is sent with the
1280 configuration of the example above, only the second section will be
1281 used, not the first, as the first value of the second section is
1282 closer to the message size. Hence, when ``smpi/os=1:3:2;10:5:1``, a
1283 message of size 11 incurs the following cost inside MPI_Send:
1284 ``5+11*1`` because 5 is the startup cost and 1 is the cost per byte.
1286 Note that the order of sections can be arbitrary; they will be ordered internally.
1290 Inject constant times for MPI_Recv()
1291 ....................................
1293 **Option** ``smpi/or``
1295 The behavior and motivation for this configuration option is identical
1296 with :ref:`smpi/os <cfg=smpi/os>`, but for the function ``MPI_Recv()``.
1299 .. _cfg=smpi/grow-injected-times:
1301 Inject constant times for MPI_Test
1302 ..................................
1304 **Option** ``smpi/test`` **default:** 0.0001
1306 By setting this option, you can control the amount of time a process
1307 sleeps when MPI_Test() is called; this is important, because SimGrid
1308 normally only advances the time while communication is happening and
1309 thus, MPI_Test will not add to the time, resulting in a deadlock if
1310 used as a break-condition as in the following example:
1315 MPI_Test(request, flag, status);
1319 To speed up execution, we use a counter to keep track on how often we
1320 already checked if the handle is now valid or not. Hence, we actually
1321 use counter*SLEEP_TIME, that is, the time MPI_Test() causes the
1322 process to sleep increases linearly with the number of previously
1323 failed tests. This behavior can be disabled by setting
1324 ``smpi/grow-injected-times`` to **no**. This will also disable this
1325 behavior for MPI_Iprobe.
1327 .. _cfg=smpi/shared-malloc:
1328 .. _cfg=smpi/shared-malloc-hugepage:
1333 **Option** ``smpi/shared-malloc`` **Possible values:** global (default), local
1335 If your simulation consumes too much memory, you may want to modify
1336 your code so that the working areas are shared by all MPI ranks. For
1337 example, in a bloc-cyclic matrix multiplication, you will only
1338 allocate one set of blocs, and every processes will share them.
1339 Naturally, this will lead to very wrong results, but this will save a
1340 lot of memory so this is still desirable for some studies. For more on
1341 the motivation for that feature, please refer to the `relevant section
1342 <https://simgrid.github.io/SMPI_CourseWare/topic_understanding_performance/matrixmultiplication>`_
1343 of the SMPI CourseWare (see Activity #2.2 of the pointed
1344 assignment). In practice, change the call to malloc() and free() into
1345 SMPI_SHARED_MALLOC() and SMPI_SHARED_FREE().
1347 SMPI provides two algorithms for this feature. The first one, called
1348 ``local``, allocates one bloc per call to SMPI_SHARED_MALLOC() in your
1349 code (each call location gets its own bloc) and this bloc is shared
1350 amongst all MPI ranks. This is implemented with the shm_* functions
1351 to create a new POSIX shared memory object (kept in RAM, in /dev/shm)
1352 for each shared bloc.
1354 With the ``global`` algorithm, each call to SMPI_SHARED_MALLOC()
1355 returns a new adress, but it only points to a shadow bloc: its memory
1356 area is mapped on a 1MiB file on disk. If the returned bloc is of size
1357 N MiB, then the same file is mapped N times to cover the whole bloc.
1358 At the end, no matter how many SMPI_SHARED_MALLOC you do, this will
1359 only consume 1 MiB in memory.
1361 You can disable this behavior and come back to regular mallocs (for
1362 example for debugging purposes) using @c "no" as a value.
1364 If you want to keep private some parts of the buffer, for instance if these
1365 parts are used by the application logic and should not be corrupted, you
1366 can use SMPI_PARTIAL_SHARED_MALLOC(size, offsets, offsets_count). Example:
1370 mem = SMPI_PARTIAL_SHARED_MALLOC(500, {27,42 , 100,200}, 2);
1372 This will allocate 500 bytes to mem, such that mem[27..41] and
1373 mem[100..199] are shared while other area remain private.
1375 Then, it can be deallocated by calling SMPI_SHARED_FREE(mem).
1377 When smpi/shared-malloc:global is used, the memory consumption problem
1378 is solved, but it may induce too much load on the kernel's pages table.
1379 In this case, you should use huge pages so that we create only one
1380 entry per Mb of malloced data instead of one entry per 4k.
1381 To activate this, you must mount a hugetlbfs on your system and allocate
1382 at least one huge page:
1384 .. code-block:: shell
1387 sudo mount none /home/huge -t hugetlbfs -o rw,mode=0777
1388 sudo sh -c 'echo 1 > /proc/sys/vm/nr_hugepages' # echo more if you need more
1390 Then, you can pass the option
1391 ``--cfg=smpi/shared-malloc-hugepage:/home/huge`` to smpirun to
1392 actually activate the huge page support in shared mallocs.
1396 Inject constant times for MPI_Wtime, gettimeofday and clock_gettime
1397 ...................................................................
1399 **Option** ``smpi/wtime`` **default:** 10 ns
1401 This option controls the amount of (simulated) time spent in calls to
1402 MPI_Wtime(), gettimeofday() and clock_gettime(). If you set this value
1403 to 0, the simulated clock is not advanced in these calls, which leads
1404 to issue if your application contains such a loop:
1408 while(MPI_Wtime() < some_time_bound) {
1409 /* some tests, with no communication nor computation */
1412 When the option smpi/wtime is set to 0, the time advances only on
1413 communications and computations, so the previous code results in an
1414 infinite loop: the current [simulated] time will never reach
1415 ``some_time_bound``. This infinite loop is avoided when that option
1416 is set to a small amount, as it is by default since SimGrid v3.21.
1418 Note that if your application does not contain any loop depending on
1419 the current time only, then setting this option to a non-zero value
1420 will slow down your simulations by a tiny bit: the simulation loop has
1421 to be broken and reset each time your code ask for the current time.
1422 If the simulation speed really matters to you, you can avoid this
1423 extra delay by setting smpi/wtime to 0.
1425 Other Configurations
1426 --------------------
1428 .. _cfg=clean-atexit:
1430 Cleanup at Termination
1431 ......................
1433 **Option** ``clean-atexit`` **default:** on
1435 If your code is segfaulting during its finalization, it may help to
1436 disable this option to request SimGrid to not attempt any cleanups at
1437 the end of the simulation. Since the Unix process is ending anyway,
1438 the operating system will wipe it all.
1445 **Option** ``path`` **default:** . (current dir)
1447 It is possible to specify a list of directories to search into for the
1448 trace files (see :ref:`pf_trace`) by using this configuration
1449 item. To add several directory to the path, set the configuration
1450 item several times, as in ``--cfg=path:toto --cfg=path:tutu``
1452 .. _cfg=simix/breakpoint:
1457 **Option** ``simix/breakpoint`` **default:** unset
1459 This configuration option sets a breakpoint: when the simulated clock
1460 reaches the given time, a SIGTRAP is raised. This can be used to stop
1461 the execution and get a backtrace with a debugger.
1463 It is also possible to set the breakpoint from inside the debugger, by
1464 writing in global variable simgrid::simix::breakpoint. For example,
1467 .. code-block:: shell
1469 set variable simgrid::simix::breakpoint = 3.1416
1471 .. _cfg=verbose-exit:
1476 **Option** ``verbose-exit`` **default:** on
1478 By default, when Ctrl-C is pressed, the status of all existing actors
1479 is displayed before exiting the simulation. This is very useful to
1480 debug your code, but it can reveal troublesome if you have many
1481 actors. Set this configuration item to **off** to disable this
1484 .. _cfg=exception/cutpath:
1486 Truncate local path from exception backtrace
1487 ............................................
1489 **Option** ``exception/cutpath`` **default:** off
1491 This configuration option is used to remove the path from the
1492 backtrace shown when an exception is thrown. This is mainly useful for
1493 the tests: the full file path makes the tests not reproducible because
1494 the path of source files depend of the build settings. That would
1495 break most of our tests as we keep comparing output.
1497 Logging Configuration
1498 ---------------------
1500 It can be done by using XBT. Go to :ref:`XBT_log` for more details.