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 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 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 - **contexts/factory:** :ref:`cfg=contexts/factory`
88 - **contexts/guard-size:** :ref:`cfg=contexts/guard-size`
89 - **contexts/nthreads:** :ref:`cfg=contexts/nthreads`
90 - **contexts/parallel-threshold:** :ref:`cfg=contexts/parallel-threshold`
91 - **contexts/stack-size:** :ref:`cfg=contexts/stack-size`
92 - **contexts/synchro:** :ref:`cfg=contexts/synchro`
94 - **cpu/maxmin-selective-update:** :ref:`Cpu Optimization Level <options_model_optim>`
95 - **cpu/model:** :ref:`options_model_select`
96 - **cpu/optim:** :ref:`Cpu Optimization Level <options_model_optim>`
98 - **debug/breakpoint:** :ref:`cfg=debug/breakpoint`
99 - **debug/clean-atexit:** :ref:`cfg=debug/clean-atexit`
100 - **debug/verbose-exit:** :ref:`cfg=debug/verbose-exit`
102 - **exception/cutpath:** :ref:`cfg=exception/cutpath`
104 - **host/model:** :ref:`options_model_select`
106 - **maxmin/precision:** :ref:`cfg=maxmin/precision`
107 - **maxmin/concurrency-limit:** :ref:`cfg=maxmin/concurrency-limit`
109 - **msg/debug-multiple-use:** :ref:`cfg=msg/debug-multiple-use`
111 - **model-check:** :ref:`options_modelchecking`
112 - **model-check/checkpoint:** :ref:`cfg=model-check/checkpoint`
113 - **model-check/communications-determinism:** :ref:`cfg=model-check/communications-determinism`
114 - **model-check/dot-output:** :ref:`cfg=model-check/dot-output`
115 - **model-check/max-depth:** :ref:`cfg=model-check/max-depth`
116 - **model-check/property:** :ref:`cfg=model-check/property`
117 - **model-check/reduction:** :ref:`cfg=model-check/reduction`
118 - **model-check/replay:** :ref:`cfg=model-check/replay`
119 - **model-check/send-determinism:** :ref:`cfg=model-check/send-determinism`
120 - **model-check/termination:** :ref:`cfg=model-check/termination`
121 - **model-check/timeout:** :ref:`cfg=model-check/timeout`
122 - **model-check/visited:** :ref:`cfg=model-check/visited`
124 - **network/bandwidth-factor:** :ref:`cfg=network/bandwidth-factor`
125 - **network/crosstraffic:** :ref:`cfg=network/crosstraffic`
126 - **network/latency-factor:** :ref:`cfg=network/latency-factor`
127 - **network/maxmin-selective-update:** :ref:`Network Optimization Level <options_model_optim>`
128 - **network/model:** :ref:`options_model_select`
129 - **network/optim:** :ref:`Network Optimization Level <options_model_optim>`
130 - **network/TCP-gamma:** :ref:`cfg=network/TCP-gamma`
131 - **network/weight-S:** :ref:`cfg=network/weight-S`
133 - **ns3/TcpModel:** :ref:`options_pls`
134 - **path:** :ref:`cfg=path`
135 - **plugin:** :ref:`cfg=plugin`
137 - **storage/max_file_descriptors:** :ref:`cfg=storage/max_file_descriptors`
139 - **surf/precision:** :ref:`cfg=surf/precision`
141 - **For collective operations of SMPI,** please refer to Section :ref:`cfg=smpi/coll-selector`
142 - **smpi/async-small-thresh:** :ref:`cfg=smpi/async-small-thresh`
143 - **smpi/buffering:** :ref:`cfg=smpi/buffering`
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`
174 - **vm/model:** :ref:`options_model_select`
178 Configuring the Platform Models
179 -------------------------------
181 .. _options_model_select:
183 Choosing the Platform Models
184 ............................
186 SimGrid comes with several network, CPU and storage models built in,
187 and you can change the used model at runtime by changing the passed
188 configuration. The three main configuration items are given below.
189 For each of these items, passing the special ``help`` value gives you
190 a short description of all possible values (for example,
191 ``--cfg=network/model:help`` will present all provided network
192 models). Also, ``--help-models`` should provide information about all
193 models for all existing resources.
195 - ``network/model``: specify the used network model. Possible values:
197 - **LV08 (default one):** Realistic network analytic model
198 (slow-start modeled by multiplying latency by 13.01, bandwidth by
199 .97; bottleneck sharing uses a payload of S=20537 for evaluating
200 RTT). Described in `Accuracy Study and Improvement of Network
201 Simulation in the SimGrid Framework
202 <http://mescal.imag.fr/membres/arnaud.legrand/articles/simutools09.pdf>`_.
203 - **Constant:** Simplistic network model where all communication
204 take a constant time (one second). This model provides the lowest
205 realism, but is (marginally) faster.
206 - **SMPI:** Realistic network model specifically tailored for HPC
207 settings (accurate modeling of slow start with correction factors on
208 three intervals: < 1KiB, < 64 KiB, >= 64 KiB). This model can be
209 :ref:`further configured <options_model_network>`.
210 - **IB:** Realistic network model specifically tailored for HPC
211 settings with InfiniBand networks (accurate modeling contention
212 behavior, based on the model explained in `this PhD work
213 <http://mescal.imag.fr/membres/jean-marc.vincent/index.html/PhD/Vienne.pdf>`_.
214 This model can be :ref:`further configured <options_model_network>`.
215 - **CM02:** Legacy network analytic model. Very similar to LV08, but
216 without corrective factors. The timings of small messages are thus
217 poorly modeled. This model is described in `A Network Model for
218 Simulation of Grid Application
219 <https://hal.inria.fr/inria-00071989/document>`_.
220 - **ns-3** (only available if you compiled SimGrid accordingly):
221 Use the packet-level network
222 simulators as network models (see :ref:`model_ns3`).
223 This model can be :ref:`further configured <options_pls>`.
225 - ``cpu/model``: specify the used CPU model. We have only one model
228 - **Cas01:** Simplistic CPU model (time=size/power)
230 - ``host/model``: The host concept is the aggregation of a CPU with a
231 network card. Three models exists, but actually, only 2 of them are
232 interesting. The "compound" one is simply due to the way our
233 internal code is organized, and can easily be ignored. So at the
234 end, you have two host models: The default one allows aggregation of
235 an existing CPU model with an existing network model, but does not
236 allow parallel tasks because these beasts need some collaboration
237 between the network and CPU model. That is why, ptask_07 is used by
238 default when using SimDag.
240 - **default:** Default host model. Currently, CPU:Cas01 and
241 network:LV08 (with cross traffic enabled)
242 - **compound:** Host model that is automatically chosen if
243 you change the network and CPU models
244 - **ptask_L07:** Host model somehow similar to Cas01+CM02 but
245 allowing "parallel tasks", that are intended to model the moldable
246 tasks of the grid scheduling literature.
248 - ``storage/model``: specify the used storage model. Only one model is
250 - ``vm/model``: specify the model for virtual machines. Only one model
253 .. todo: make 'compound' the default host model.
255 .. _options_model_optim:
260 The network and CPU models that are based on lmm_solve (that
261 is, all our analytical models) accept specific optimization
264 - items ``network/optim`` and ``cpu/optim`` (both default to 'Lazy'):
266 - **Lazy:** Lazy action management (partial invalidation in lmm +
267 heap in action remaining).
268 - **TI:** Trace integration. Highly optimized mode when using
269 availability traces (only available for the Cas01 CPU model for
271 - **Full:** Full update of remaining and variables. Slow but may be
272 useful when debugging.
274 - items ``network/maxmin-selective-update`` and
275 ``cpu/maxmin-selective-update``: configure whether the underlying
276 should be lazily updated or not. It should have no impact on the
277 computed timings, but should speed up the computation. |br| It is
278 still possible to disable this feature because it can reveal
279 counter-productive in very specific scenarios where the
280 interaction level is high. In particular, if all your
281 communication share a given backbone link, you should disable it:
282 without it, a simple regular loop is used to update each
283 communication. With it, each of them is still updated (because of
284 the dependency induced by the backbone), but through a complicated
285 and slow pattern that follows the actual dependencies.
287 .. _cfg=maxmin/precision:
288 .. _cfg=surf/precision:
293 **Option** ``maxmin/precision`` **Default:** 0.00001 (in flops or bytes) |br|
294 **Option** ``surf/precision`` **Default:** 0.00001 (in seconds)
296 The analytical models handle a lot of floating point values. It is
297 possible to change the epsilon used to update and compare them through
298 this configuration item. Changing it may speedup the simulation by
299 discarding very small actions, at the price of a reduced numerical
300 precision. You can modify separately the precision used to manipulate
301 timings (in seconds) and the one used to manipulate amounts of work
304 .. _cfg=maxmin/concurrency-limit:
309 **Option** ``maxmin/concurrency-limit`` **Default:** -1 (no limit)
311 The maximum number of variables per resource can be tuned through this
312 option. You can have as many simultaneous actions per resources as you
313 want. If your simulation presents a very high level of concurrency, it
314 may help to use e.g. 100 as a value here. It means that at most 100
315 actions can consume a resource at a given time. The extraneous actions
316 are queued and wait until the amount of concurrency of the considered
317 resource lowers under the given boundary.
319 Such limitations help both to the simulation speed and simulation accuracy
320 on highly constrained scenarios, but the simulation speed suffers of this
321 setting on regular (less constrained) scenarios so it is off by default.
323 .. _options_model_network:
325 Configuring the Network Model
326 .............................
328 .. _cfg=network/TCP-gamma:
330 Maximal TCP Window Size
331 ^^^^^^^^^^^^^^^^^^^^^^^
333 **Option** ``network/TCP-gamma`` **Default:** 4194304
335 The analytical models need to know the maximal TCP window size to take
336 the TCP congestion mechanism into account. On Linux, this value can
337 be retrieved using the following commands. Both give a set of values,
338 and you should use the last one, which is the maximal size.
340 .. code-block:: shell
342 cat /proc/sys/net/ipv4/tcp_rmem # gives the sender window
343 cat /proc/sys/net/ipv4/tcp_wmem # gives the receiver window
345 .. _cfg=smpi/IB-penalty-factors:
346 .. _cfg=network/bandwidth-factor:
347 .. _cfg=network/latency-factor:
348 .. _cfg=network/weight-S:
350 Correcting Important Network Parameters
351 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
353 SimGrid can take network irregularities such as a slow startup or
354 changing behavior depending on the message size into account. You
355 should not change these values unless you really know what you're
356 doing. The corresponding values were computed through data fitting
357 one the timings of packet-level simulators, as described in `Accuracy
358 Study and Improvement of Network Simulation in the SimGrid Framework
359 <http://mescal.imag.fr/membres/arnaud.legrand/articles/simutools09.pdf>`_.
362 If you are using the SMPI model, these correction coefficients are
363 themselves corrected by constant values depending on the size of the
364 exchange. By default SMPI uses factors computed on the Stampede
365 Supercomputer at TACC, with optimal deployment of processes on
366 nodes. Again, only hardcore experts should bother about this fact.
368 InfiniBand network behavior can be modeled through 3 parameters
369 ``smpi/IB-penalty-factors:"βe;βs;γs"``, as explained in `this PhD
371 <http://mescal.imag.fr/membres/jean-marc.vincent/index.html/PhD/Vienne.pdf>`_.
373 .. todo:: This section should be rewritten, and actually explain the
374 options network/bandwidth-factor, network/latency-factor,
377 .. _cfg=network/crosstraffic:
379 Simulating Cross-Traffic
380 ^^^^^^^^^^^^^^^^^^^^^^^^
382 Since SimGrid v3.7, cross-traffic effects can be taken into account in
383 analytical simulations. It means that ongoing and incoming
384 communication flows are treated independently. In addition, the LV08
385 model adds 0.05 of usage on the opposite direction for each new
386 created flow. This can be useful to simulate some important TCP
387 phenomena such as ack compression.
389 For that to work, your platform must have two links for each
390 pair of interconnected hosts. An example of usable platform is
391 available in ``examples/platforms/crosstraffic.xml``.
393 This is activated through the ``network/crosstraffic`` item, that
394 can be set to 0 (disable this feature) or 1 (enable it).
396 Note that with the default host model this option is activated by default.
398 .. _cfg=smpi/async-small-thresh:
400 Simulating Asynchronous Send
401 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
403 (this configuration item is experimental and may change or disappear)
405 It is possible to specify that messages below a certain size will be
406 sent as soon as the call to MPI_Send is issued, without waiting for
407 the correspondant receive. This threshold can be configured through
408 the ``smpi/async-small-thresh`` item. The default value is 0. This
409 behavior can also be manually set for mailboxes, by setting the
410 receiving mode of the mailbox with a call to
411 :cpp:func:`MSG_mailbox_set_async`. After this, all messages sent to
412 this mailbox will have this behavior regardless of the message size.
414 This value needs to be smaller than or equals to the threshold set at
415 :ref:`cfg=smpi/send-is-detached-thresh`, because asynchronous messages
416 are meant to be detached as well.
423 **Option** ``ns3/TcpModel`` **Default:** "default" (ns-3 default)
425 When using ns-3, there is an extra item ``ns3/TcpModel``, corresponding
426 to the ``ns3::TcpL4Protocol::SocketType`` configuration item in
427 ns-3. The only valid values (enforced on the SimGrid side) are
428 'default' (no change to the ns-3 configuration), 'NewReno' or 'Reno' or
431 Configuring the Storage model
432 .............................
434 .. _cfg=storage/max_file_descriptors:
436 File Descriptor Cound per Host
437 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
439 **Option** ``storage/max_file_descriptors`` **Default:** 1024
441 Each host maintains a fixed-size array of its file descriptors. You
442 can change its size through this item to either enlarge it if your
443 application requires it or to reduce it to save memory space.
450 SimGrid plugins allow one to extend the framework without changing its
451 source code directly. Read the source code of the existing plugins to
452 learn how to do so (in ``src/plugins``), and ask your questions to the
453 usual channels (Stack Overflow, Mailing list, IRC). The basic idea is
454 that plugins usually register callbacks to some signals of interest.
455 If they need to store some information about a given object (Link, CPU
456 or Actor), they do so through the use of a dedicated object extension.
458 Some of the existing plugins can be activated from the command line,
459 meaning that you can activate them from the command line without any
460 modification to your simulation code. For example, you can activate
461 the host energy plugin by adding ``--cfg=plugin:host_energy`` to your
464 Here is the full list of plugins that can be activated this way:
466 - **host_energy:** keeps track of the energy dissipated by
467 computations. More details in @ref plugin_energy.
468 - **link_energy:** keeps track of the energy dissipated by
469 communications. More details in @ref SURF_plugin_energy.
470 - **host_load:** keeps track of the computational load.
471 More details in @ref plugin_load.
473 .. _options_modelchecking:
475 Configuring the Model-Checking
476 ------------------------------
478 To enable the SimGrid model-checking support the program should
479 be executed using the simgrid-mc wrapper:
481 .. code-block:: shell
483 simgrid-mc ./my_program
485 Safety properties are expressed as assertions using the function
486 :cpp:func:`void MC_assert(int prop)`.
488 .. _cfg=smpi/buffering:
490 Specifying the MPI buffering behavior
491 .....................................
493 **Option** ``smpi/buffering`` **Default:** infty
495 Buffering in MPI has a huge impact on the communication semantic. For example,
496 standard blocking sends are synchronous calls when the system buffers are full
497 while these calls can complete immediately without even requiring a matching
498 receive call for small messages sent when the system buffers are empty.
500 In SMPI, this depends on the message size, that is compared against two thresholds:
502 - if (size < :ref:`smpi/async-small-thresh <cfg=smpi/async-small-thresh>`) then
503 MPI_Send returns immediately, even if the corresponding receive has not be issued yet.
504 - 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
505 MPI_Send returns as soon as the corresponding receive has been issued. This is known as the eager mode.
506 - if (:ref:`smpi/send-is-detached-thresh <cfg=smpi/send-is-detached-thresh>` < size) then
507 MPI_Send returns only when the message has actually been sent over the network. This is known as the rendez-vous mode.
509 The ``smpi/buffering`` option gives an easier interface to choose between these semantics. It can take two values:
511 - **zero:** means that buffering should be disabled. Blocking communications are actually blocking.
512 - **infty:** means that buffering should be made infinite. Blocking communications are non-blocking.
514 .. _cfg=model-check/property:
516 Specifying a liveness property
517 ..............................
519 **Option** ``model-check/property`` **Default:** unset
521 If you want to specify liveness properties, you have to pass them on
522 the command line, specifying the name of the file containing the
523 property, as formatted by the ltl2ba program. Note that ltl2ba is not
524 part of SimGrid and must be installed separatly.
526 .. code-block:: shell
528 simgrid-mc ./my_program --cfg=model-check/property:<filename>
530 .. _cfg=model-check/checkpoint:
532 Going for Stateful Verification
533 ...............................
535 By default, the system is backtracked to its initial state to explore
536 another path instead of backtracking to the exact step before the fork
537 that we want to explore (this is called stateless verification). This
538 is done this way because saving intermediate states can rapidly
539 exhaust the available memory. If you want, you can change the value of
540 the ``model-check/checkpoint`` item. For example,
541 ``--cfg=model-check/checkpoint:1`` asks to take a checkpoint every
542 step. Beware, this will certainly explode your memory. Larger values
543 are probably better, make sure to experiment a bit to find the right
544 setting for your specific system.
546 .. _cfg=model-check/reduction:
548 Specifying the kind of reduction
549 ................................
551 The main issue when using the model-checking is the state space
552 explosion. You can activate some reduction technique with
553 ``--cfg=model-check/reduction:<technique>``. For now, this
554 configuration variable can take 2 values:
556 - **none:** Do not apply any kind of reduction (mandatory for
557 liveness properties, as our current DPOR algorithm breaks cycles)
558 - **dpor:** Apply Dynamic Partial Ordering Reduction. Only valid if
559 you verify local safety properties (default value for safety
562 Another way to mitigate the state space explosion is to search for
563 cycles in the exploration with the :ref:`cfg=model-check/visited`
564 configuration. Note that DPOR and state-equality reduction may not
565 play well together. You should choose between them.
567 Our current DPOR implementation could be improved in may ways. We are
568 currently improving its efficiency (both in term of reduction ability
569 and computational speed), and future work could make it compatible
570 with liveness properties.
572 .. _cfg=model-check/visited:
574 Size of Cycle Detection Set (state equality reduction)
575 ......................................................
577 Mc SimGrid can be asked to search for cycles during the exploration,
578 i.e. situations where a new explored state is in fact the same state
579 than a previous one.. This can prove useful to mitigate the state
580 space explosion with safety properties, and this is the crux when
581 searching for counter-examples to the liveness properties.
583 Note that this feature may break the current implementation of the
584 DPOR reduction technique.
586 The ``model-check/visited`` item is the maximum number of states which
587 are stored in memory. If the maximum number of snapshotted state is
588 reached, some states will be removed from the memory and some cycles
589 might be missed. Small values can lead to incorrect verifications, but
590 large values can exhaust your memory and be CPU intensive as each new
591 state must be compared to that amount of older saved states.
593 The default settings depend on the kind of exploration. With safety
594 checking, no state is snapshotted and cycles cannot be detected. With
595 liveness checking, all states are snapshotted because missing a cycle
596 could hinder the exploration soundness.
598 .. _cfg=model-check/termination:
600 Non-Termination Detection
601 .........................
603 The ``model-check/termination`` configuration item can be used to
604 report if a non-termination execution path has been found. This is a
605 path with a cycle which means that the program might never terminate.
607 This only works in safety mode, not in liveness mode.
609 This options is disabled by default.
611 .. _cfg=model-check/dot-output:
616 If set, the ``model-check/dot-output`` configuration item is the name
617 of a file in which to write a dot file of the path leading the found
618 property (safety or liveness violation) as well as the cycle for
619 liveness properties. This dot file can then fed to the graphviz dot
620 tool to generate an corresponding graphical representation.
622 .. _cfg=model-check/max-depth:
624 Exploration Depth Limit
625 .......................
627 The ``model-checker/max-depth`` can set the maximum depth of the
628 exploration graph of the model checker. If this limit is reached, a
629 logging message is sent and the results might not be exact.
631 By default, there is not depth limit.
633 .. _cfg=model-check/timeout:
638 By default, the model checker does not handle timeout conditions: the `wait`
639 operations never time out. With the ``model-check/timeout`` configuration item
640 set to **yes**, the model checker will explore timeouts of `wait` operations.
642 .. _cfg=model-check/communications-determinism:
643 .. _cfg=model-check/send-determinism:
645 Communication Determinism
646 .........................
648 The ``model-check/communications-determinism`` and
649 ``model-check/send-determinism`` items can be used to select the
650 communication determinism mode of the model checker which checks
651 determinism properties of the communications of an application.
653 Verification Performance Considerations
654 .......................................
656 The size of the stacks can have a huge impact on the memory
657 consumption when using model-checking. By default, each snapshot will
658 save a copy of the whole stacks and not only of the part which is
659 really meaningful: you should expect the contribution of the memory
660 consumption of the snapshots to be @f$ @mbox{number of processes}
661 @times @mbox{stack size} @times @mbox{number of states} @f$.
663 When compiled against the model checker, the stacks are not
664 protected with guards: if the stack size is too small for your
665 application, the stack will silently overflow on other parts of the
666 memory (see :ref:`contexts/guard-size <cfg=contexts/guard-size>`).
668 .. _cfg=model-check/replay:
670 Replaying buggy execution paths from the model checker
671 ......................................................
673 Debugging the problems reported by the model checker is challenging: First, the
674 application under verification cannot be debugged with gdb because the
675 model checker already traces it. Then, the model checker may explore several
676 execution paths before encountering the issue, making it very difficult to
677 understand the outputs. Fortunately, SimGrid provides the execution path leading
678 to any reported issue so that you can replay this path out of the model checker,
679 enabling the usage of classical debugging tools.
681 When the model checker finds an interesting path in the application
682 execution graph (where a safety or liveness property is violated), it
683 generates an identifier for this path. Here is an example of output:
685 .. code-block:: shell
687 [ 0.000000] (0:@) Check a safety property
688 [ 0.000000] (0:@) **************************
689 [ 0.000000] (0:@) *** PROPERTY NOT VALID ***
690 [ 0.000000] (0:@) **************************
691 [ 0.000000] (0:@) Counter-example execution trace:
692 [ 0.000000] (0:@) [(1)Tremblay (app)] MC_RANDOM(3)
693 [ 0.000000] (0:@) [(1)Tremblay (app)] MC_RANDOM(4)
694 [ 0.000000] (0:@) Path = 1/3;1/4
695 [ 0.000000] (0:@) Expanded states = 27
696 [ 0.000000] (0:@) Visited states = 68
697 [ 0.000000] (0:@) Executed transitions = 46
699 The interesting line is ``Path = 1/3;1/4``, which means that you should use
700 ``--cfg=model-check/replay:1/3;1/4`` to replay your application on the buggy
701 execution path. All options (but the model checker related ones) must
702 remain the same. In particular, if you ran your application with
703 ``smpirun -wrapper simgrid-mc``, then do it again. Remove all
704 MC-related options, keep the other ones and add
705 ``--cfg=model-check/replay``.
707 Currently, if the path is of the form ``X;Y;Z``, each number denotes
708 the actor's pid that is selected at each indecision point. If it's of
709 the form ``X/a;Y/b``, the X and Y are the selected pids while the a
710 and b are the return values of their simcalls. In the previouse
711 example, ``1/3;1/4``, you can see from the full output that the actor
712 1 is doing MC_RANDOM simcalls, so the 3 and 4 simply denote the values
713 that these simcall return.
715 Configuring the User Code Virtualization
716 ----------------------------------------
718 .. _cfg=contexts/factory:
720 Selecting the Virtualization Factory
721 ....................................
723 **Option** contexts/factory **Default:** "raw"
725 In SimGrid, the user code is virtualized in a specific mechanism that
726 allows the simulation kernel to control its execution: when a user
727 process requires a blocking action (such as sending a message), it is
728 interrupted, and only gets released when the simulated clock reaches
729 the point where the blocking operation is done. This is explained
730 graphically in the `relevant tutorial, available online
731 <https://simgrid.org/tutorials/simgrid-simix-101.pdf>`_.
733 In SimGrid, the containers in which user processes are virtualized are
734 called contexts. Several context factory are provided, and you can
735 select the one you want to use with the ``contexts/factory``
736 configuration item. Some of the following may not exist on your
737 machine because of portability issues. In any case, the default one
738 should be the most effcient one (please report bugs if the
739 auto-detection fails for you). They are approximately sorted here from
740 the slowest to the most efficient:
742 - **thread:** very slow factory using full featured threads (either
743 pthreads or windows native threads). They are slow but very
744 standard. Some debuggers or profilers only work with this factory.
745 - **java:** Java applications are virtualized onto java threads (that
746 are regular pthreads registered to the JVM)
747 - **ucontext:** fast factory using System V contexts (Linux and FreeBSD only)
748 - **boost:** This uses the `context
749 implementation <http://www.boost.org/doc/libs/1_59_0/libs/context/doc/html/index.html>`_
750 of the boost library for a performance that is comparable to our
752 |br| Install the relevant library (e.g. with the
753 libboost-contexts-dev package on Debian/Ubuntu) and recompile
755 - **raw:** amazingly fast factory using a context switching mechanism
756 of our own, directly implemented in assembly (only available for x86
757 and amd64 platforms for now) and without any unneeded system call.
759 The main reason to change this setting is when the debugging tools get
760 fooled by the optimized context factories. Threads are the most
761 debugging-friendly contexts, as they allow one to set breakpoints
762 anywhere with gdb and visualize backtraces for all processes, in order
763 to debug concurrency issues. Valgrind is also more comfortable with
764 threads, but it should be usable with all factories (Exception: the
765 callgrind tool really dislikes raw and ucontext factories).
767 .. _cfg=contexts/stack-size:
769 Adapting the Stack Size
770 .......................
772 **Option** ``contexts/stack-size`` **Default:** 8192 KiB
774 Each virtualized used process is executed using a specific system
775 stack. The size of this stack has a huge impact on the simulation
776 scalability, but its default value is rather large. This is because
777 the error messages that you get when the stack size is too small are
778 rather disturbing: this leads to stack overflow (overwriting other
779 stacks), leading to segfaults with corrupted stack traces.
781 If you want to push the scalability limits of your code, you might
782 want to reduce the ``contexts/stack-size`` item. Its default value is
783 8192 (in KiB), while our Chord simulation works with stacks as small
784 as 16 KiB, for example. This *setting is ignored* when using the
785 thread factory. Instead, you should compile SimGrid and your
786 application with ``-fsplit-stack``. Note that this compilation flag is
787 not compatible with the model checker right now.
789 The operating system should only allocate memory for the pages of the
790 stack which are actually used and you might not need to use this in
791 most cases. However, this setting is very important when using the
792 model checker (see :ref:`options_mc_perf`).
794 .. _cfg=contexts/guard-size:
796 Disabling Stack Guard Pages
797 ...........................
799 **Option** ``contexts/guard-size`` **Default** 1 page in most case (0 pages on Windows or with MC)
801 Unless you use the threads context factory (see
802 :ref:`cfg=contexts/factory`), a stack guard page is usually used
803 which prevents the stack of a given actor from overflowing on another
804 stack. But the performance impact may become prohibitive when the
805 amount of actors increases. The option ``contexts/guard-size`` is the
806 number of stack guard pages used. By setting it to 0, no guard pages
807 will be used: in this case, you should avoid using small stacks (with
808 :ref:`contexts/stack-size <cfg=contexts/stack-size>`) as the stack
809 will silently overflow on other parts of the memory.
811 When no stack guard page is created, stacks may then silently overflow
812 on other parts of the memory if their size is too small for the
815 .. _cfg=contexts/nthreads:
816 .. _cfg=contexts/parallel-threshold:
817 .. _cfg=contexts/synchro:
819 Running User Code in Parallel
820 .............................
822 Parallel execution of the user code is only considered stable in
823 SimGrid v3.7 and higher, and mostly for MSG simulations. SMPI
824 simulations may well fail in parallel mode. It is described in
825 `INRIA RR-7653 <http://hal.inria.fr/inria-00602216/>`_.
827 If you are using the **ucontext** or **raw** context factories, you can
828 request to execute the user code in parallel. Several threads are
829 launched, each of them handling the same number of user contexts at each
830 run. To activate this, set the ``contexts/nthreads`` item to the amount
831 of cores that you have in your computer (or lower than 1 to have the
832 amount of cores auto-detected).
834 Even if you asked several worker threads using the previous option,
835 you can request to start the parallel execution (and pay the
836 associated synchronization costs) only if the potential parallelism is
837 large enough. For that, set the ``contexts/parallel-threshold``
838 item to the minimal amount of user contexts needed to start the
839 parallel execution. In any given simulation round, if that amount is
840 not reached, the contexts will be run sequentially directly by the
841 main thread (thus saving the synchronization costs). Note that this
842 option is mainly useful when the grain of the user code is very fine,
843 because our synchronization is now very efficient.
845 When parallel execution is activated, you can choose the
846 synchronization schema used with the ``contexts/synchro`` item,
847 which value is either:
849 - **futex:** ultra optimized synchronisation schema, based on futexes
850 (fast user-mode mutexes), and thus only available on Linux systems.
851 This is the default mode when available.
852 - **posix:** slow but portable synchronisation using only POSIX
854 - **busy_wait:** not really a synchronisation: the worker threads
855 constantly request new contexts to execute. It should be the most
856 efficient synchronisation schema, but it loads all the cores of
857 your machine for no good reason. You probably prefer the other less
860 Configuring the Tracing
861 -----------------------
863 The :ref:`tracing subsystem <outcomes_vizu>` can be configured in
864 several different ways depending on the nature of the simulator (MSG,
865 SimDag, SMPI) and the kind of traces that need to be obtained. See the
866 :ref:`Tracing Configuration Options subsection
867 <tracing_tracing_options>` to get a detailed description of each
868 configuration option.
870 We detail here a simple way to get the traces working for you, even if
871 you never used the tracing API.
874 - Any SimGrid-based simulator (MSG, SimDag, SMPI, ...) and raw traces:
876 .. code-block:: shell
878 --cfg=tracing:yes --cfg=tracing/uncategorized:yes --cfg=triva/uncategorized:uncat.plist
880 The first parameter activates the tracing subsystem, the second
881 tells it to trace host and link utilization (without any
882 categorization) and the third creates a graph configuration file to
883 configure Triva when analysing the resulting trace file.
885 - MSG or SimDag-based simulator and categorized traces (you need to
886 declare categories and classify your tasks according to them)
888 .. code-block:: shell
890 --cfg=tracing:yes --cfg=tracing/categorized:yes --cfg=triva/categorized:cat.plist
892 The first parameter activates the tracing subsystem, the second
893 tells it to trace host and link categorized utilization and the
894 third creates a graph configuration file to configure Triva when
895 analysing the resulting trace file.
897 - SMPI simulator and traces for a space/time view:
899 .. code-block:: shell
903 The `-trace` parameter for the smpirun script runs the simulation
904 with ``--cfg=tracing:yes --cfg=tracing/smpi:yes``. Check the
905 smpirun's `-help` parameter for additional tracing options.
907 Sometimes you might want to put additional information on the trace to
908 correctly identify them later, or to provide data that can be used to
909 reproduce an experiment. You have two ways to do that:
911 - Add a string on top of the trace file as comment:
913 .. code-block:: shell
915 --cfg=tracing/comment:my_simulation_identifier
917 - Add the contents of a textual file on top of the trace file as comment:
919 .. code-block:: shell
921 --cfg=tracing/comment-file:my_file_with_additional_information.txt
923 Please, use these two parameters (for comments) to make reproducible
924 simulations. For additional details about this and all tracing
925 options, check See the :ref:`tracing_tracing_options`.
930 .. _cfg=msg/debug-multiple-use:
935 **Option** ``msg/debug-multiple-use`` **Default:** off
937 Sometimes your application may try to send a task that is still being
938 executed somewhere else, making it impossible to send this task. However,
939 for debugging purposes, one may want to know what the other host is/was
940 doing. This option shows a backtrace of the other process.
945 The SMPI interface provides several specific configuration items.
946 These are uneasy to see since the code is usually launched through the
947 ``smiprun`` script directly.
949 .. _cfg=smpi/host-speed:
950 .. _cfg=smpi/cpu-threshold:
951 .. _cfg=smpi/simulate-computation:
953 Automatic Benchmarking of SMPI Code
954 ...................................
956 In SMPI, the sequential code is automatically benchmarked, and these
957 computations are automatically reported to the simulator. That is to
958 say that if you have a large computation between a ``MPI_Recv()`` and
959 a ``MPI_Send()``, SMPI will automatically benchmark the duration of
960 this code, and create an execution task within the simulator to take
961 this into account. For that, the actual duration is measured on the
962 host machine and then scaled to the power of the corresponding
963 simulated machine. The variable ``smpi/host-speed`` allows one to specify
964 the computational speed of the host machine (in flop/s) to use when
965 scaling the execution times. It defaults to 20000, but you really want
966 to update it to get accurate simulation results.
968 When the code is constituted of numerous consecutive MPI calls, the
969 previous mechanism feeds the simulation kernel with numerous tiny
970 computations. The ``smpi/cpu-threshold`` item becomes handy when this
971 impacts badly the simulation performance. It specifies a threshold (in
972 seconds) below which the execution chunks are not reported to the
973 simulation kernel (default value: 1e-6).
975 .. note:: The option ``smpi/cpu-threshold`` ignores any computation
976 time spent below this threshold. SMPI does not consider the
977 `amount` of these computations; there is no offset for this. Hence,
978 a value that is too small, may lead to unreliable simulation
981 In some cases, however, one may wish to disable simulation of
982 application computation. This is the case when SMPI is used not to
983 simulate an MPI applications, but instead an MPI code that performs
984 "live replay" of another MPI app (e.g., ScalaTrace's replay tool,
985 various on-line simulators that run an app at scale). In this case the
986 computation of the replay/simulation logic should not be simulated by
987 SMPI. Instead, the replay tool or on-line simulator will issue
988 "computation events", which correspond to the actual MPI simulation
989 being replayed/simulated. At the moment, these computation events can
990 be simulated using SMPI by calling internal smpi_execute*() functions.
992 To disable the benchmarking/simulation of computation in the simulated
993 application, the variable ``smpi/simulate-computation`` should be set
994 to no. This option just ignores the timings in your simulation; it
995 still executes the computations itself. If you want to stop SMPI from
996 doing that, you should check the SMPI_SAMPLE macros, documented in
997 Section :ref:`SMPI_adapting_speed`.
999 +------------------------------------+-------------------------+-----------------------------+
1000 | Solution | Computations executed? | Computations simulated? |
1001 +====================================+=========================+=============================+
1002 | --cfg=smpi/simulate-computation:no | Yes | Never |
1003 +------------------------------------+-------------------------+-----------------------------+
1004 | --cfg=smpi/cpu-threshold:42 | Yes, in all cases | If it lasts over 42 seconds |
1005 +------------------------------------+-------------------------+-----------------------------+
1006 | SMPI_SAMPLE() macro | Only once per loop nest | Always |
1007 +------------------------------------+-------------------------+-----------------------------+
1009 .. _cfg=smpi/comp-adjustment-file:
1011 Slow-down or speed-up parts of your code
1012 ........................................
1014 **Option** ``smpi/comp-adjustment-file:`` **Default:** unset
1016 This option allows you to pass a file that contains two columns: The
1017 first column defines the section that will be subject to a speedup;
1018 the second column is the speedup. For instance:
1020 .. code-block:: shell
1022 "start:stop","ratio"
1023 "exchange_1.f:30:exchange_1.f:130",1.18244559422142
1025 The first line is the header - you must include it. The following
1026 line means that the code between two consecutive MPI calls on line 30
1027 in exchange_1.f and line 130 in exchange_1.f should receive a speedup
1028 of 1.18244559422142. The value for the second column is therefore a
1029 speedup, if it is larger than 1 and a slow-down if it is smaller
1030 than 1. Nothing will be changed if it is equal to 1.
1032 Of course, you can set any arbitrary filenames you want (so the start
1033 and end don't have to be in the same file), but be aware that this
1034 mechanism only supports `consecutive calls!`
1036 Please note that you must pass the ``-trace-call-location`` flag to
1037 smpicc or smpiff, respectively. This flag activates some internal
1038 macro definitions that help with obtaining the call location.
1040 .. _cfg=smpi/bw-factor:
1045 **Option** ``smpi/bw-factor``
1046 |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
1048 The possible throughput of network links is often dependent on the
1049 message sizes, as protocols may adapt to different message sizes. With
1050 this option, a series of message sizes and factors are given, helping
1051 the simulation to be more realistic. For instance, the current default
1052 value means that messages with size 65472 and more will get a total of
1053 MAX_BANDWIDTH*0.940694, messages of size 15424 to 65471 will get
1054 MAX_BANDWIDTH*0.697866 and so on (where MAX_BANDWIDTH denotes the
1055 bandwidth of the link).
1057 An experimental script to compute these factors is available online. See
1058 https://framagit.org/simgrid/platform-calibration/
1059 https://simgrid.org/contrib/smpi-saturation-doc.html
1061 .. _cfg=smpi/display-timing:
1063 Reporting Simulation Time
1064 .........................
1066 **Option** ``smpi/display-timing`` **Default:** 0 (false)
1068 Most of the time, you run MPI code with SMPI to compute the time it
1069 would take to run it on a platform. But since the code is run through
1070 the ``smpirun`` script, you don't have any control on the launcher
1071 code, making it difficult to report the simulated time when the
1072 simulation ends. If you enable the ``smpi/display-timing`` item,
1073 ``smpirun`` will display this information when the simulation
1076 .. _cfg=smpi/keep-temps:
1078 Keeping temporary files after simulation
1079 ........................................
1081 **Option** ``smpi/keep-temps`` **default:** 0 (false)
1083 SMPI usually generates a lot of temporary files that are cleaned after
1084 use. This option request to preserve them, for example to debug or
1085 profile your code. Indeed, the binary files are removed very early
1086 under the dlopen privatization schema, which tend to fool the
1089 .. _cfg=smpi/lat-factor:
1094 **Option** ``smpi/lat-factor`` |br|
1095 **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
1097 The motivation and syntax for this option is identical to the motivation/syntax
1098 of :ref:`cfg=smpi/bw-factor`.
1100 There is an important difference, though: While smpi/bw-factor `reduces` the
1101 actual bandwidth (i.e., values between 0 and 1 are valid), latency factors
1102 increase the latency, i.e., values larger than or equal to 1 are valid here.
1104 .. _cfg=smpi/papi-events:
1106 Trace hardware counters with PAPI
1107 .................................
1109 **Option** ``smpi/papi-events`` **default:** unset
1111 When the PAPI support was compiled in SimGrid, this option takes the
1112 names of PAPI counters and adds their respective values to the trace
1113 files (See Section :ref:`tracing_tracing_options`).
1117 This feature currently requires superuser privileges, as registers
1118 are queried. Only use this feature with code you trust! Call
1119 smpirun for instance via ``smpirun -wrapper "sudo "
1120 <your-parameters>`` or run ``sudo sh -c "echo 0 >
1121 /proc/sys/kernel/perf_event_paranoid"`` In the later case, sudo
1122 will not be required.
1124 It is planned to make this feature available on a per-process (or per-thread?) basis.
1125 The first draft, however, just implements a "global" (i.e., for all processes) set
1126 of counters, the "default" set.
1128 .. code-block:: shell
1130 --cfg=smpi/papi-events:"default:PAPI_L3_LDM:PAPI_L2_LDM"
1132 .. _cfg=smpi/privatization:
1134 Automatic Privatization of Global Variables
1135 ...........................................
1137 **Option** ``smpi/privatization`` **default:** "dlopen" (when using smpirun)
1139 MPI executables are usually meant to be executed in separated
1140 processes, but SMPI is executed in only one process. Global variables
1141 from executables will be placed in the same memory zone and shared
1142 between processes, causing intricate bugs. Several options are
1143 possible to avoid this, as described in the main `SMPI publication
1144 <https://hal.inria.fr/hal-01415484>`_ and in the :ref:`SMPI
1145 documentation <SMPI_what_globals>`. SimGrid provides two ways of
1146 automatically privatizing the globals, and this option allows one to
1147 choose between them.
1149 - **no** (default when not using smpirun): Do not automatically
1150 privatize variables. Pass ``-no-privatize`` to smpirun to disable
1152 - **dlopen** or **yes** (default when using smpirun): Link multiple
1153 times against the binary.
1154 - **mmap** (slower, but maybe somewhat more stable):
1155 Runtime automatic switching of the data segments.
1158 This configuration option cannot be set in your platform file. You can only
1159 pass it as an argument to smpirun.
1161 .. _cfg=smpi/privatize-libs:
1163 Automatic privatization of global variables inside external libraries
1164 .....................................................................
1166 **Option** ``smpi/privatize-libs`` **default:** unset
1168 **Linux/BSD only:** When using dlopen (default) privatization,
1169 privatize specific shared libraries with internal global variables, if
1170 they can't be linked statically. For example libgfortran is usually
1171 used for Fortran I/O and indexes in files can be mixed up.
1173 Multiple libraries can be given, semicolon separated.
1175 This configuration option can only use either full paths to libraries,
1176 or full names. Check with ldd the name of the library you want to
1179 .. code-block:: shell
1183 libgfortran.so.3 => /usr/lib/x86_64-linux-gnu/libgfortran.so.3 (0x00007fbb4d91b000)
1186 Then you can use ``--cfg=smpi/privatize-libs:libgfortran.so.3``
1187 or ``--cfg=smpi/privatize-libs:/usr/lib/x86_64-linux-gnu/libgfortran.so.3``,
1188 but not ``libgfortran`` nor ``libgfortran.so``.
1190 .. _cfg=smpi/send-is-detached-thresh:
1192 Simulating MPI detached send
1193 ............................
1195 **Option** ``smpi/send-is-detached-thresh`` **default:** 65536
1197 This threshold specifies the size in bytes under which the send will
1198 return immediately. This is different from the threshold detailed in
1199 :ref:`cfg=smpi/async-small-thresh` because the message is not
1200 effectively sent when the send is posted. SMPI still waits for the
1201 correspondant receive to be posted to perform the communication
1204 .. _cfg=smpi/coll-selector:
1206 Simulating MPI collective algorithms
1207 ....................................
1209 **Option** ``smpi/coll-selector`` **Possible values:** naive (default), ompi, mpich
1211 SMPI implements more than 100 different algorithms for MPI collective
1212 communication, to accurately simulate the behavior of most of the
1213 existing MPI libraries. The ``smpi/coll-selector`` item can be used to
1214 use the decision logic of either OpenMPI or MPICH libraries (by
1215 default SMPI uses naive version of collective operations).
1217 Each collective operation can be manually selected with a
1218 ``smpi/collective_name:algo_name``. Available algorithms are listed in
1219 :ref:`SMPI_use_colls`.
1221 .. TODO:: All available collective algorithms will be made available
1222 via the ``smpirun --help-coll`` command.
1224 .. _cfg=smpi/iprobe:
1226 Inject constant times for MPI_Iprobe
1227 ....................................
1229 **Option** ``smpi/iprobe`` **default:** 0.0001
1231 The behavior and motivation for this configuration option is identical
1232 with :ref:`smpi/test <cfg=smpi/test>`, but for the function
1235 .. _cfg=smpi/iprobe-cpu-usage:
1237 Reduce speed for iprobe calls
1238 .............................
1240 **Option** ``smpi/iprobe-cpu-usage`` **default:** 1 (no change)
1242 MPI_Iprobe calls can be heavily used in applications. To account
1243 correctly for the energy cores spend probing, it is necessary to
1244 reduce the load that these calls cause inside SimGrid.
1246 For instance, we measured a max power consumption of 220 W for a
1247 particular application but only 180 W while this application was
1248 probing. Hence, the correct factor that should be passed to this
1249 option would be 180/220 = 0.81.
1253 Inject constant times for MPI_Init
1254 ..................................
1256 **Option** ``smpi/init`` **default:** 0
1258 The behavior and motivation for this configuration option is identical
1259 with :ref:`smpi/test <cfg=smpi/test>`, but for the function ``MPI_Init()``.
1263 Inject constant times for MPI_Isend()
1264 .....................................
1266 **Option** ``smpi/ois``
1268 The behavior and motivation for this configuration option is identical
1269 with :ref:`smpi/os <cfg=smpi/os>`, but for the function ``MPI_Isend()``.
1273 Inject constant times for MPI_send()
1274 ....................................
1276 **Option** ``smpi/os``
1278 In several network models such as LogP, send (MPI_Send, MPI_Isend) and
1279 receive (MPI_Recv) operations incur costs (i.e., they consume CPU
1280 time). SMPI can factor these costs in as well, but the user has to
1281 configure SMPI accordingly as these values may vary by machine. This
1282 can be done by using ``smpi/os`` for MPI_Send operations; for MPI_Isend
1283 and MPI_Recv, use ``smpi/ois`` and ``smpi/or``, respectively. These work
1284 exactly as ``smpi/ois``.
1286 This item can consist of multiple sections; each section takes three
1287 values, for example ``1:3:2;10:5:1``. The sections are divided by ";"
1288 so this example contains two sections. Furthermore, each section
1289 consists of three values.
1291 1. The first value denotes the minimum size for this section to take effect;
1292 read it as "if message size is greater than this value (and other section has a larger
1293 first value that is also smaller than the message size), use this".
1294 In the first section above, this value is "1".
1296 2. The second value is the startup time; this is a constant value that will always
1297 be charged, no matter what the size of the message. In the first section above,
1300 3. The third value is the `per-byte` cost. That is, it is charged for every
1301 byte of the message (incurring cost messageSize*cost_per_byte)
1302 and hence accounts also for larger messages. In the first
1303 section of the example above, this value is "2".
1305 Now, SMPI always checks which section it should take for a given
1306 message; that is, if a message of size 11 is sent with the
1307 configuration of the example above, only the second section will be
1308 used, not the first, as the first value of the second section is
1309 closer to the message size. Hence, when ``smpi/os=1:3:2;10:5:1``, a
1310 message of size 11 incurs the following cost inside MPI_Send:
1311 ``5+11*1`` because 5 is the startup cost and 1 is the cost per byte.
1313 Note that the order of sections can be arbitrary; they will be ordered internally.
1317 Inject constant times for MPI_Recv()
1318 ....................................
1320 **Option** ``smpi/or``
1322 The behavior and motivation for this configuration option is identical
1323 with :ref:`smpi/os <cfg=smpi/os>`, but for the function ``MPI_Recv()``.
1326 .. _cfg=smpi/grow-injected-times:
1328 Inject constant times for MPI_Test
1329 ..................................
1331 **Option** ``smpi/test`` **default:** 0.0001
1333 By setting this option, you can control the amount of time a process
1334 sleeps when MPI_Test() is called; this is important, because SimGrid
1335 normally only advances the time while communication is happening and
1336 thus, MPI_Test will not add to the time, resulting in a deadlock if
1337 used as a break-condition as in the following example:
1342 MPI_Test(request, flag, status);
1346 To speed up execution, we use a counter to keep track on how often we
1347 already checked if the handle is now valid or not. Hence, we actually
1348 use counter*SLEEP_TIME, that is, the time MPI_Test() causes the
1349 process to sleep increases linearly with the number of previously
1350 failed tests. This behavior can be disabled by setting
1351 ``smpi/grow-injected-times`` to **no**. This will also disable this
1352 behavior for MPI_Iprobe.
1354 .. _cfg=smpi/shared-malloc:
1355 .. _cfg=smpi/shared-malloc-hugepage:
1360 **Option** ``smpi/shared-malloc`` **Possible values:** global (default), local
1362 If your simulation consumes too much memory, you may want to modify
1363 your code so that the working areas are shared by all MPI ranks. For
1364 example, in a bloc-cyclic matrix multiplication, you will only
1365 allocate one set of blocs, and every processes will share them.
1366 Naturally, this will lead to very wrong results, but this will save a
1367 lot of memory so this is still desirable for some studies. For more on
1368 the motivation for that feature, please refer to the `relevant section
1369 <https://simgrid.github.io/SMPI_CourseWare/topic_understanding_performance/matrixmultiplication>`_
1370 of the SMPI CourseWare (see Activity #2.2 of the pointed
1371 assignment). In practice, change the call to malloc() and free() into
1372 SMPI_SHARED_MALLOC() and SMPI_SHARED_FREE().
1374 SMPI provides two algorithms for this feature. The first one, called
1375 ``local``, allocates one bloc per call to SMPI_SHARED_MALLOC() in your
1376 code (each call location gets its own bloc) and this bloc is shared
1377 amongst all MPI ranks. This is implemented with the shm_* functions
1378 to create a new POSIX shared memory object (kept in RAM, in /dev/shm)
1379 for each shared bloc.
1381 With the ``global`` algorithm, each call to SMPI_SHARED_MALLOC()
1382 returns a new address, but it only points to a shadow bloc: its memory
1383 area is mapped on a 1MiB file on disk. If the returned bloc is of size
1384 N MiB, then the same file is mapped N times to cover the whole bloc.
1385 At the end, no matter how many SMPI_SHARED_MALLOC you do, this will
1386 only consume 1 MiB in memory.
1388 You can disable this behavior and come back to regular mallocs (for
1389 example for debugging purposes) using @c "no" as a value.
1391 If you want to keep private some parts of the buffer, for instance if these
1392 parts are used by the application logic and should not be corrupted, you
1393 can use SMPI_PARTIAL_SHARED_MALLOC(size, offsets, offsets_count). Example:
1397 mem = SMPI_PARTIAL_SHARED_MALLOC(500, {27,42 , 100,200}, 2);
1399 This will allocate 500 bytes to mem, such that mem[27..41] and
1400 mem[100..199] are shared while other area remain private.
1402 Then, it can be deallocated by calling SMPI_SHARED_FREE(mem).
1404 When smpi/shared-malloc:global is used, the memory consumption problem
1405 is solved, but it may induce too much load on the kernel's pages table.
1406 In this case, you should use huge pages so that we create only one
1407 entry per Mb of malloced data instead of one entry per 4k.
1408 To activate this, you must mount a hugetlbfs on your system and allocate
1409 at least one huge page:
1411 .. code-block:: shell
1414 sudo mount none /home/huge -t hugetlbfs -o rw,mode=0777
1415 sudo sh -c 'echo 1 > /proc/sys/vm/nr_hugepages' # echo more if you need more
1417 Then, you can pass the option
1418 ``--cfg=smpi/shared-malloc-hugepage:/home/huge`` to smpirun to
1419 actually activate the huge page support in shared mallocs.
1423 Inject constant times for MPI_Wtime, gettimeofday and clock_gettime
1424 ...................................................................
1426 **Option** ``smpi/wtime`` **default:** 10 ns
1428 This option controls the amount of (simulated) time spent in calls to
1429 MPI_Wtime(), gettimeofday() and clock_gettime(). If you set this value
1430 to 0, the simulated clock is not advanced in these calls, which leads
1431 to issue if your application contains such a loop:
1435 while(MPI_Wtime() < some_time_bound) {
1436 /* some tests, with no communication nor computation */
1439 When the option smpi/wtime is set to 0, the time advances only on
1440 communications and computations, so the previous code results in an
1441 infinite loop: the current [simulated] time will never reach
1442 ``some_time_bound``. This infinite loop is avoided when that option
1443 is set to a small amount, as it is by default since SimGrid v3.21.
1445 Note that if your application does not contain any loop depending on
1446 the current time only, then setting this option to a non-zero value
1447 will slow down your simulations by a tiny bit: the simulation loop has
1448 to be broken and reset each time your code ask for the current time.
1449 If the simulation speed really matters to you, you can avoid this
1450 extra delay by setting smpi/wtime to 0.
1452 Other Configurations
1453 --------------------
1455 .. _cfg=debug/clean-atexit:
1457 Cleanup at Termination
1458 ......................
1460 **Option** ``debug/clean-atexit`` **default:** on
1462 If your code is segfaulting during its finalization, it may help to
1463 disable this option to request SimGrid to not attempt any cleanups at
1464 the end of the simulation. Since the Unix process is ending anyway,
1465 the operating system will wipe it all.
1472 **Option** ``path`` **default:** . (current dir)
1474 It is possible to specify a list of directories to search into for the
1475 trace files (see :ref:`pf_trace`) by using this configuration
1476 item. To add several directory to the path, set the configuration
1477 item several times, as in ``--cfg=path:toto --cfg=path:tutu``
1479 .. _cfg=debug/breakpoint:
1484 **Option** ``debug/breakpoint`` **default:** unset
1486 This configuration option sets a breakpoint: when the simulated clock
1487 reaches the given time, a SIGTRAP is raised. This can be used to stop
1488 the execution and get a backtrace with a debugger.
1490 It is also possible to set the breakpoint from inside the debugger, by
1491 writing in global variable simgrid::simix::breakpoint. For example,
1494 .. code-block:: shell
1496 set variable simgrid::simix::breakpoint = 3.1416
1498 .. _cfg=debug/verbose-exit:
1503 **Option** ``debug/verbose-exit`` **default:** on
1505 By default, when Ctrl-C is pressed, the status of all existing actors
1506 is displayed before exiting the simulation. This is very useful to
1507 debug your code, but it can reveal troublesome if you have many
1508 actors. Set this configuration item to **off** to disable this
1511 .. _cfg=exception/cutpath:
1513 Truncate local path from exception backtrace
1514 ............................................
1516 **Option** ``exception/cutpath`` **default:** off
1518 This configuration option is used to remove the path from the
1519 backtrace shown when an exception is thrown. This is mainly useful for
1520 the tests: the full file path makes the tests not reproducible because
1521 the path of source files depend of the build settings. That would
1522 break most of our tests as we keep comparing output.
1524 Logging Configuration
1525 ---------------------
1527 It can be done by using XBT. Go to :ref:`XBT_log` for more details.