3 <object id="TOC" data="graphical-toc.svg" type="image/svg+xml"></object>
5 window.onload=function() { // Wait for the SVG to be loaded before changing it
6 var elem=document.querySelector("#TOC").contentDocument.getElementById("ReferenceBox")
7 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";
13 .. _platform_reference:
15 Complete XML Reference
16 **********************
18 Your platform description should follow the specification presented in the
19 `simgrid.dtd <https://simgrid.org/simgrid.dtd>`_ DTD file. The same DTD is used for both platform and deployment files.
21 -------------------------------------------------------------------------------
28 SimGrid can simulate the time it takes to read or write data on disk, even if the stored data is not made persistent in
29 any way by SimGrid. This means that your application will correctly be slowed down when doing simulated I/O, but there
30 is no way to get the data stored this way.
32 We decided to not model anything beyond raw access in SimGrid because we believe that there is not single way of doing so.
33 We provide an example model of file system as a plugin, (sparsely) documented in :ref:`plugin_filesystem`.
35 **Parent tags:** :ref:`pf_tag_host` |br|
36 **Children tags:** :ref:`pf_tag_prop` |br|
39 :``id``: A name of your choice (must be unique on this host).
40 :``read_bw``: Read bandwidth for this disk. You must specify a unit as follows.
42 **Units in bytes and powers of 2** (1 KiBps = 1,024 Bps):
43 Bps, KiBps, MiBps, GiBps, TiBps, PiBps, or EiBps. |br|
44 **Units in bits and powers of 2** (1 Bps = 8 bps):
45 bps, Kibps, Mibps, Gibps, Tibps, Pibps, or Eibps. |br|
46 **Units in bytes and powers of 10** (1 KBps = 1,000 Bps):
47 Bps, KBps, MBps, GBps, TBps, PBps, or EBps. |br|
48 **Units in bits and powers of 10:**
49 bps, Kbps, Mbps, Gbps, Tbps, Pbps, or Ebps.
51 :``write_bw``: Write bandwidth for this disk. You must specify a unit as for the read bandwidth.
55 <host id="alice" speed="1Gf">
56 <disk id="Disk1" read_bw="200MBps" write_bw="80MBps">
57 <!-- you can add properties for anything you want: they are not used by SimGrid -->
58 <prop id="content" value="storage/content/small_content.txt"/>
60 <prop id="ram" value="100B" />
63 -------------------------------------------------------------------------------
70 Adding configuration flags directly into the platform file becomes particularly
71 useful when the realism of the described platform depends on some specific
72 flags. For example, this could help you to finely tune SMPI. Almost all
73 :ref:`command-line configuration items <options_list>` can be configured this
76 Each configuration flag is described as a :ref:`pf_tag_prop` whose ``id`` is the
77 name of the flag and ``value`` is what it has to be set to.
79 **Parent tags:** :ref:`pf_tag_platform` (must appear before any other tags) |br|
80 **Children tags:** :ref:`pf_tag_prop` |br|
85 <?xml version = '1.0'?>
86 <!DOCTYPE platform SYSTEM "https://simgrid.org/simgrid.dtd">
87 <platform version = "4.1">
89 <prop id = "maxmin/precision" value = "0.000010" />
90 <prop id = "cpu/optim" value = "TI" />
91 <prop id = "network/model" value = "SMPI" />
92 <prop id = "smpi/bw-factor" value = "65472:0.940694;15424:0.697866;9376:0.58729" />
95 <!-- The rest of your platform -->
98 -------------------------------------------------------------------------------
105 A host is the computing resource on which an actor can run. See :cpp:class:`simgrid::s4u::Host`.
107 **Parent tags:** :ref:`pf_tag_zone` (only leaf zones, i.e., zones containing neither inner zones nor clusters) |br|
108 **Children tags:** :ref:`pf_tag_mount`, :ref:`pf_tag_prop`, :ref:`pf_tag_disk` |br|
112 Must be unique over the whole platform.
113 :``speed``: Computational power (per core, in flop/s).
114 If you use DVFS, provide a comma-separated list of values for each pstate (see :ref:`howto_dvfs`).
115 :``core``: Amount of cores (default: 1).
116 See :ref:`howto_multicore`.
117 :``availability_file``:
118 File containing the availability profile.
119 Almost every lines of such files describe timed events as ``date ratio``.
122 .. code-block:: python
129 - At time t = 1, half of the host computational power (0.5 means 50%) is used to process some background load, hence
130 only 50% of this initial power remains available to your own simulation.
131 - At time t = 2, the available power drops at 20% of the initial value.
132 - At time t = 5, the host can compute at full speed again.
133 - At time t = 10, the profile is reset (as we are 5 seconds after the last event). Then the available speed will drop
134 again to 50% at time t = 11.
136 If your profile does not contain any LOOPAFTER line, then it will
137 be executed only once and not repeated.
139 .. warning:: Don't get fooled: Bandwidth and Latency profiles of a :ref:`pf_tag_link` contain absolute values, while
140 Availability profiles of a :ref:`pf_tag_host` contain ratios.
141 :``state_file``: File containing the state profile.
142 Almost every lines of such files describe timed events as ``date boolean``.
145 .. code-block:: python
151 - At time t = 1, the host is turned off (a zero value means OFF)
152 - At time t = 2, the host is turned back on (any other value than zero means ON)
153 - At time t = 10, the profile is reset (as we are 8 seconds after the last event). Then the host will be turned off
154 again at time t = 11.
156 If your profile does not contain any LOOPAFTER line, then it will
157 be executed only once and not repeated.
159 :``coordinates``: Vivaldi coordinates (meaningful for Vivaldi zones only).
160 See :ref:`pf_tag_peer`.
161 :``pstate``: Initial pstate (default: 0, the first one).
162 See :ref:`howto_dvfs`.
164 -------------------------------------------------------------------------------
171 SimGrid links usually represent one-hop network connections (see :cpp:class:`simgrid::s4u::Link`), i.e., a single wire.
172 They can also be used to abstract a larger network interconnect, e.g., the entire transcontinental network, into a
175 **Parent tags:** :ref:`pf_tag_zone` (both leaf zones and inner zones) |br|
176 **Children tags:** :ref:`pf_tag_prop` |br|
179 :``id``: Link name. Must be unique over the whole platform.
180 :``bandwidth``: Maximum bandwidth for this link. You must specify a unit as follows.
182 **Units in bytes and powers of 2** (1 KiBps = 1,024 Bps):
183 Bps, KiBps, MiBps, GiBps, TiBps, PiBps, or EiBps. |br|
184 **Units in bits and powers of 2** (1 Bps = 8 bps):
185 bps, Kibps, Mibps, Gibps, Tibps, Pibps, or Eibps. |br|
186 **Units in bytes and powers of 10** (1 KBps = 1,000 Bps):
187 Bps, KBps, MBps, GBps, TBps, PBps, or EBps. |br|
188 **Units in bits and powers of 10:**
189 bps, Kbps, Mbps, Gbps, Tbps, Pbps, or Ebps.
191 :``latency``: Latency for this link (default: 0.0). You must specify a unit as follows.
193 ==== =========== ======================
194 Unit Meaning Duration in seconds
195 ==== =========== ======================
196 ps picosecond 10⁻¹² = 0.000000000001
197 ns nanosecond 10⁻⁹ = 0.000000001
198 us microsecond 10⁻⁶ = 0.000001
199 ms millisecond 10⁻³ = 0.001
204 w week 60 * 60 * 24 * 7
205 ==== =========== ======================
207 :``sharing_policy``: Sharing policy for the link. Possible values are ``SHARED``, ``FATPIPE`` or ``SPLITDUPLEX``
208 (default: ``SPLITDUPLEX``).
210 If set to ``SPLITDUPLEX``, the link models the full-duplex
211 behavior, as meant in TCP or UDP. To that extend, the link is
212 actually split in two links whose names are suffixed with "_UP" and
213 "_DOWN". You should then specify the direction to use when
214 referring to that link in a :ref:`pf_tag_link_ctn`.
216 If set to ``FATPIPE``, flows have no impact on each other, hence
217 each flow can exploit the full bandwidth. This models Internet
218 backbones that cannot get saturated by your application. From your
219 application point of view, there is no congestion on these
222 If set to ``SHARED``, the available bandwidth is fairly shared
223 among ALL flows traversing this link. The resulting link is not
224 full-duplex (as UDP or TCP would be): communications in both
225 directions share the same link. Prefer ``SPLITDUPLEX`` for TCP flows.
227 :``bandwidth_file``: File containing the bandwidth profile.
228 Almost every lines of such files describe timed events as ``date
229 bandwidth`` (in bytes per second).
232 .. code-block:: python
238 - At time t = 4, the bandwidth is of 40 MBps.
239 - At time t = 8, it raises to 60MBps.
240 - At time t = 24, it drops at 40 MBps again.
242 If your profile does not contain any LOOPAFTER line, then it will
243 be executed only once and not repeated.
245 .. warning:: Don't get fooled: Bandwidth and Latency profiles of a :ref:`pf_tag_link` contain absolute values, while
246 Availability profiles of a :ref:`pf_tag_host` contain ratios.
248 :``latency_file``: File containing the latency profile.
249 Almost every lines of such files describe timed events as ``date
250 latency`` (in seconds).
253 .. code-block:: python
259 - At time t = 1, the latency is of 1ms (0.001 second)
260 - At time t = 3, the latency is of 100ms (0.1 second)
261 - At time t = 8 (5 seconds after the last event), the profile loops.
262 - At time t = 9 (1 second after the loop reset), the latency is back at 1ms.
264 If your profile does not contain any LOOPAFTER line, then it will
265 be executed only once and not repeated.
267 .. warning:: Don't get fooled: Bandwidth and Latency profiles of a :ref:`pf_tag_link` contain absolute values, while
268 Availability profiles of a :ref:`pf_tag_host` contain ratios.
270 :``state_file``: File containing the state profile. See :ref:`pf_tag_host`.
272 -------------------------------------------------------------------------------
279 An element in a route, representing a previously defined link.
281 **Parent tags:** :ref:`pf_tag_route` |br|
282 **Children tags:** none |br|
285 :``id``: Link that is to be included in this route.
286 :``direction``: either ``UP`` (by default) or ``DOWN``, specifying whether to
287 use the uplink or downlink component of the link (that must
288 follow the ``SPLITDUPLEX`` sharing policy). |br|
289 Please refer to the ``sharing_policy`` attribute in
292 -------------------------------------------------------------------------------
299 This tag represents a peer, as in Peer-to-Peer (P2P) networks. It is
300 handy to model situations where hosts have an asymmetric
301 connectivity. Computers connected through set-top-boxes usually have a
302 much better download rate than their upload rate. To model this,
303 <peer> creates and connects several elements: a host, an upload link
306 **Parent tags:** :ref:`pf_tag_zone` (only with Vivaldi routing) |br|
307 **Children tags:** none |br|
310 :``id``: Name of the host. Must be unique on the whole platform.
311 :``speed``: Computational power (in flop/s).
313 If you use DVFS, provide a comma-separated list of values for each pstate (see :ref:`howto_dvfs`).
314 :``bw_in``: Bandwidth of the private downstream link, along with its
315 unit. See :ref:`pf_tag_link`.
316 :``bw_out``: Bandwidth of the private upstream link, along with its
317 unit. See :ref:`pf_tag_link`.
318 :``lat``: Latency of both private links. See :ref:`pf_tag_link`.
319 :``coordinates``: Coordinates of the gateway for this peer.
321 The communication latency between a host A = (xA,yA,zA) and a host B = (xB,yB,zB) is computed as follows:
323 latency = sqrt( (xA-xB)² + (yA-yB)² ) + zA + zB
325 See the documentation of
326 :cpp:class:`simgrid::kernel::routing::VivaldiZone` for details on
327 how the latency is computed from the coordinates, and on how the up
328 and down bandwidth are used.
329 :``availability_file``: File containing the availability profile.
330 See the full description in :ref:`pf_tag_host`
331 :``state_file``: File containing the state profile.
332 See the full description in :ref:`pf_tag_host`
334 -------------------------------------------------------------------------------
341 **Parent tags:** none (this is the root tag of every file) |br|
342 **Children tags:** :ref:`pf_tag_config` (must come first),
343 :ref:`pf_tag_cluster`, :ref:`pf_tag_cabinet`, :ref:`pf_tag_peer`,
344 :ref:`pf_tag_zone`, :ref:`pf_tag_trace`, :ref:`pf_tag_trace_connect`, or
345 :ref:`pf_tag_actor` in :ref:`deployment <deploy>` files. |br|
348 :``version``: Version of the DTD, describing the whole XML format.
349 This versioning allow future evolutions, even if we
350 avoid backward-incompatible changes. The current version
351 is **4.1**. The ``simgrid_update_xml`` program can
352 upgrade most of the past platform files to the most recent
355 -------------------------------------------------------------------------------
362 This tag can be used to attach user-defined properties to some
363 platform elements. Both the name and the value can be any string of
364 your wish. You can use this to pass extra parameters to your code and
367 From your code, you can interact with these properties using the
370 - Actor: :cpp:func:`simgrid::s4u::Actor::get_property` or :cpp:func:`MSG_process_get_property_value`
371 - Cluster: this is a zone, see below.
372 - Host: :cpp:func:`simgrid::s4u::Host::get_property` or :cpp:func:`MSG_host_get_property_value`
373 - Link: :cpp:func:`simgrid::s4u::Link::get_property`
374 - Disk: :cpp:func:`simgrid::s4u::Disk::get_property`
375 - Storage :cpp:func:`MSG_storage_get_property_value` (deprecated)
376 - Zone: :cpp:func:`simgrid::s4u::Zone::get_property` of :cpp:func:`MSG_zone_get_property_value`
378 **Parent tags:** :ref:`pf_tag_actor`, :ref:`pf_tag_config`, :ref:`pf_tag_cluster`, :ref:`pf_tag_host`,
379 :ref:`pf_tag_link`, :ref:`pf_tag_disk`,:ref:`pf_tag_zone` |br|
380 **Children tags:** none |br|
383 :``id``: Name of the defined property.
384 :``value``: Value of the defined property.
386 -------------------------------------------------------------------------------
393 A path between two network locations, composed of several occurrences
394 of :ref:`pf_tag_link` .
396 **Parent tags:** :ref:`pf_tag_zone` |br|
397 **Children tags:** :ref:`pf_tag_link_ctn` |br|
400 :``src``: Host from which this route starts. Must be an existing host.
401 :``dst``: Host to which this route leads. Must be an existing host.
402 :``symmetrical``: Whether this route is symmetrical, ie, whether we
403 are defining the route ``dst -> src`` at the same
404 time. Valid values: ``yes``, ``no``, ``YES``, ``NO``.
406 -------------------------------------------------------------------------------
411 ------------------------------------------------------------------
413 A router is similar to a :ref:`pf_tag_host`, but it cannot contain
414 any actor. It is only useful to some routing algorithms. In
415 particular, they are useful when you want to use the NS3 bindings to
416 break the routes that are longer than 1 hop.
418 **Parent tags:** :ref:`pf_tag_zone` (only leaf zones, i.e., zones containing neither inner zones nor clusters) |br|
421 :``id``: Router name.
422 No other host or router may have the same name over the whole platform.
423 :``coordinates``: Vivaldi coordinates. See :ref:`pf_tag_peer`.
425 -------------------------------------------------------------------------------
432 A networking zone is an area in which elements are located. See :cpp:class:`simgrid::s4u::Zone`.
434 **Parent tags:** :ref:`pf_tag_platform`, :ref:`pf_tag_zone` (only internal nodes, i.e., zones
435 containing only inner zones or clusters but no basic
436 elements such as host or peer) |br|
437 **Children tags (if internal zone):** :ref:`pf_tag_cluster`, :ref:`pf_tag_link`, :ref:`pf_tag_zone` |br|
438 **Children tags (if leaf zone):** :ref:`pf_tag_host`, :ref:`pf_tag_link`, :ref:`pf_tag_peer` |br|
442 No other zone may have the same name over the whole platform.
443 :``routing``: Routing algorithm to use.
445 -------------------------------------------------------------------------------
447 .. _pf_tag_zoneRoute:
452 The purpose of this entity is to define a route between two zones.
453 Recall that all zones form a tree, so to connect two sibling zones,
454 you must give such a zoneRoute specifying the source and destination zones,
455 along with the gateway in each zone (i.e., the point to reach within that zone to reach the zone),
456 and the list of links to go from one zone to another.
458 **Parent tags:** :ref:`pf_tag_zone` |br|
459 **Children tags:** :ref:`pf_tag_link_ctn` |br|
462 :``src``: Zone from which this route starts. Must be an existing zone.
463 :``dst``: Zone to which this route leads. Must be an existing zone.
464 :``gw_src``: Netpoint (within src zone) from which this route starts. Must be an existing host/router.
465 :``gw_dst``: Netpoint (within dst zone) to which this route leads. Must be an existing host/router.
466 :``symmetrical``: Whether this route is symmetrical, ie, whether we
467 are defining the route ``dst -> src`` at the same
468 time. Valid values: ``yes``, ``no``, ``YES``, ``NO``.
470 -------------------------------------------------------------------------------
477 This tag is commonly used when one wants to define many hosts and a network quickly.
478 Technically, cluster is a meta-tag: **from the inner SimGrid point of
479 view, a cluster is a network zone where some optimized routing is defined**.
480 The default inner organization of the cluster is as follow:
487 ____________|__________|_____________ backbone
489 l0| l1| l2| l97| l96 | | l99
495 Here, a set of **hosts** is defined. Each of them has a **link**
496 to a central backbone (backbone is a link itself, as a link can
497 be used to represent a switch).
498 A **router** allows to connect a **cluster** to the outside world. Internally,
499 SimGrid treats a cluster as a network zone containing all hosts: the router is the
500 gateway for the cluster.
502 There is an alternative organization, which is as follows:
517 The principle is the same, except that there is no backbone. This representation
518 can be obtained easily: just do not set the bb_* attributes.
520 **Parent tags:** :ref:`pf_tag_platform`, :ref:`pf_tag_zone` |br|
521 **Children tags:** none |br|
524 :``id``: The identifier of the cluster. Facilitates referring to this cluster.
525 :``prefix``: Each node of the cluster has to have a name. This name will be prefixed with this prefix.
526 :``suffix``: Each node of the cluster will be suffixed with this suffix
527 :``radical``: Regexp used to generate cluster nodes name.
529 Syntax: "10-20" will give you 11 machines numbered from 10 to 20, "10-20;2" will give you 12 machines, one with the number 2, others numbered as before.
531 The produced number is concatenated between prefix and suffix to form machine names.
532 :``speed``: Same as the ``speed`` attribute of the :ref:`pf_tag_host` tag.
533 :``core``: Same as the ``core`` attribute of the :ref:`pf_tag_host` tag.
534 :``bw``: Bandwidth for the links between nodes and backbone (if any). See :ref:`pf_tag_link` for syntax/details.
535 :``lat``: Latency for the links between nodes and backbone (if any). See :ref:`pf_tag_link` for syntax/details.
536 :``sharing_policy``: Sharing policy for the links between nodes and backbone (if any). See :ref:`pf_tag_link` for syntax/details.
537 :``bb_bw``: Bandwidth for backbone (if any). See :ref:`pf_tag_link` for syntax/details.
539 If bb_bw and bb_lat attributes are omitted, no backbone is created (alternative cluster architecture described earlier).
540 :``bb_lat``: Latency for backbone (if any). See :ref:`pf_tag_link` section for syntax/details.
542 If bb_lat and bb_bw attributes are omitted, no backbone is created (alternative cluster architecture described earlier).
543 :``bb_sharing_policy``: Sharing policy for the backbone (if any). See :ref:`pf_tag_link` section for syntax/details.
544 :``limiter_link``: Bandwidth for limiter link (if any).
546 This adds a specific link for each node, to set the maximum bandwidth reached when communicating in both directions at the same time.
548 In theory this value should be 2*bw for splitduplex links, but in reality this might be less. This value will depend heavily on the communication model, and on the cluster's hardware, so no default value can be set, this has to be measured.
550 More details can be obtained in `Toward Better Simulation of MPI Applications on Ethernet/TCP Networks <https://hal.inria.fr/hal-00919507/>`_
551 :``loopback_bw``: Bandwidth for loopback (if any). See :ref:`pf_tag_link` section for syntax/details.
553 If loopback_bw and loopback_lat attributes are omitted, no loopback link is created and all intra-node communication will use the main network link of the node.
555 The sharing policy of a loopback link is **FATPIPE** :ref:`pf_tag_link`.
556 :``loopback_lat``: Latency for loopback (if any). See loopback_bw for more info.
557 :``topology``: Network topology to use.
559 SimGrid currently supports FLAT (with or without backbone, as described before), TORUS, FAT_TREE and DRAGONFLY attributes for this tag.
561 See :ref:`platform_examples` for more details.
563 :``topo_parameters``: Specific parameters to pass for the topology defined in the topology tag.
565 For torus networks, comma-separated list of the number of nodes in each dimension of the torus.
567 Please refer to :ref:`platform_examples`.
573 Consider the following two (and independent) uses of the ``cluster`` tag:
577 <cluster id="my_cluster_1" prefix="" suffix="" radical="0-262144"
578 speed="1e9" bw="125e6" lat="5E-5"/>
580 <cluster id="my_cluster_2" prefix="c-" suffix=".me" radical="0-99"
581 speed="1e9" bw="125e6" lat="5E-5"
582 bb_bw="2.25e9" bb_lat="5E-4"/>
585 The second example creates one router and 100 machines with the following names:
589 c-my_cluster_2_router.me
598 The router name is defined as the resulting string: prefix + clusterId + "_router" + suffix.
599 In this case: *my_cluster_1_router* and *my_cluster_2_router.me*.
601 -------------------------------------------------------------------------------
609 This tag is only available when the routing mode of the network zone
610 is set to ``Cluster``.
612 The cabinet tag is, like the :ref:`pf_tag_cluster` tag,
613 a meta-tag. This means that it is simply a shortcut for creating a set of (homogenous) hosts and links quickly;
614 unsurprisingly, this tag was introduced to setup cabinets in data centers. Unlike
615 :ref:`pf_tag_cluster`, however, the :ref:`pf_tag_cabinet` assumes that you create the backbone
616 and routers yourself; see our examples below.
618 **Parent tags:** :ref:`pf_tag_zone` |br|
619 **Children tags:** none |br|
622 :``id``: The identifier of the cabinet. Facilitates referring to it.
623 :``prefix``: Each node of the cabinet has to have a name. This name will be prefixed with this prefix.
624 :``suffix``: Each node of the cabinet will be suffixed with this suffix.
625 :``radical``: Regexp used to generate cabinet nodes name.
626 Syntax: "10-20" will give you 11 machines numbered from 10 to 20, "10-20;2" will give you 12 machines, one with the number 2, others numbered as before.
628 The produced number is concatenated between prefix and suffix to form machine names.
629 :``speed``: Same as the ``speed`` attribute of the :ref:`pf_tag_host` tag.
630 :``bw``: Bandwidth for the links between nodes and backbone (if any). See the :ref:`pf_tag_link` for syntax/details.
631 :``lat``: Latency for the links between nodes and backbone (if any). See the :ref:`pf_tag_link` for syntax/details.
634 Please note that as of now, it is impossible to change attributes such as,
635 amount of cores (always set to 1), the initial state of hosts/links
636 (always set to ON), the sharing policy of the links (always set to **SPLITDUPLEX** :ref:`pf_tag_link` ).
641 The following example shows how to use the cabinet tag.
643 .. literalinclude:: ../../examples/platforms/meta_cluster.xml
646 Please note that you must specify the :ref:`pf_backbone` tag by yourself;
647 this is not done automatically and there are no checks
648 that ensure this backbone was defined.
650 The hosts generated in the above example are named host-1.cluster, host-2.cluster1