schnorr [Fri, 14 Sep 2012 08:26:00 +0000 (10:26 +0200)]
[trace] documentation for new tracing options (comment, comment_file)
schnorr [Fri, 14 Sep 2012 08:25:41 +0000 (10:25 +0200)]
[trace] missing documentation for tracing/basic option
schnorr [Fri, 14 Sep 2012 08:09:39 +0000 (10:09 +0200)]
[trace] update tesh'es to contain simgrid version that generated them
schnorr [Fri, 14 Sep 2012 08:07:17 +0000 (10:07 +0200)]
Revert "[trace] remove version, otherwise we'll have to updates tesh'es at every release"
This reverts commit
b4a1edc6568563a179fc1177a36421094d0bf6ac.
schnorr [Thu, 13 Sep 2012 22:24:19 +0000 (00:24 +0200)]
[trace] update tesh'es
schnorr [Thu, 13 Sep 2012 22:23:00 +0000 (00:23 +0200)]
[trace] remove version, otherwise we'll have to updates tesh'es at every release
and use active voice
schnorr [Thu, 13 Sep 2012 22:18:17 +0000 (00:18 +0200)]
[trace] reflect new tracing options in smpirun script
schnorr [Thu, 13 Sep 2012 22:17:19 +0000 (00:17 +0200)]
cosmetics
schnorr [Thu, 13 Sep 2012 22:10:02 +0000 (00:10 +0200)]
[trace] dumping comment and comment_file to trace files when requested
schnorr [Thu, 13 Sep 2012 22:09:41 +0000 (00:09 +0200)]
fix, removing tabs
schnorr [Thu, 13 Sep 2012 21:53:06 +0000 (23:53 +0200)]
[trace] new tracing options to add comments to trace files
schnorr [Thu, 13 Sep 2012 21:47:51 +0000 (23:47 +0200)]
fix: missing undef
Augustin Degomme [Thu, 13 Sep 2012 07:50:11 +0000 (09:50 +0200)]
fix a compilation warning/error in MSG_task_isend_with_matching instrumentation
schnorr [Wed, 12 Sep 2012 15:24:18 +0000 (17:24 +0200)]
[trace] instrumentation of MSG_task_isend_with_matching
details:
- tracing only works for asynchronous messaging on sender side for now
- let us know if you need to trace a simulation with asynchronous messaging on the receiver side
schnorr [Wed, 12 Sep 2012 15:03:13 +0000 (17:03 +0200)]
[trace] instrumenting MSG_task_dsend
suter [Tue, 11 Sep 2012 16:21:10 +0000 (18:21 +0200)]
add tracing for the trace replay tool. produces a trace, have to
visualize it now.
navarro [Tue, 11 Sep 2012 12:14:12 +0000 (14:14 +0200)]
Update file example for ls command
navarro [Tue, 11 Sep 2012 12:03:31 +0000 (14:03 +0200)]
Add a function ls to storage
suter [Mon, 10 Sep 2012 12:27:06 +0000 (14:27 +0200)]
don't need to synchronize processes anymore
suter [Mon, 10 Sep 2012 09:02:16 +0000 (11:02 +0200)]
slight change in the example that lauchs everything
suter [Mon, 10 Sep 2012 09:01:04 +0000 (11:01 +0200)]
functional implementation of trace replay on top of SMPI.
Major modifications required in the input traces:
* get rid off the "p" before the rank of the process (it was useless
BTW)
* add the message size to the recv actions (simplify a lot of thing, and
we can get it easily from the TAU traces)
suter [Mon, 10 Sep 2012 08:58:13 +0000 (10:58 +0200)]
add some test to not perform some memcpy. Mandatory for trace replay on
top of SMPI. We don't care about data, so the send and recv buffers are
always NULL. memcpy doesn't like that at all.
suter [Mon, 10 Sep 2012 08:56:33 +0000 (10:56 +0200)]
add a getter to user data attached to a smpi_process
suter [Mon, 10 Sep 2012 08:55:50 +0000 (10:55 +0200)]
add a getter on user data attached to a smpi_process
suter [Thu, 6 Sep 2012 14:16:14 +0000 (16:16 +0200)]
attach time-independent traces to processes when the executable is
called replay. It takes only one argument that is a file listing the
traces to attach. If there is only one trace in this file, all the
processes try to find their own actions in it. Otherwise, they have
their own trace to read.
suter [Thu, 23 Aug 2012 14:04:45 +0000 (16:04 +0200)]
doxygen correction: SD_READY state doesn't exist anymore, replaced by
SD_SCHEDULABLE.
suter [Thu, 12 Jul 2012 15:17:16 +0000 (17:17 +0200)]
add a field to attach user data to a smpi process and the associate set
function
navarro [Thu, 6 Sep 2012 14:06:51 +0000 (16:06 +0200)]
Add a simcall for as and router property
navarro [Thu, 6 Sep 2012 13:43:39 +0000 (15:43 +0200)]
Add 3 msg functions to get and modify property of an AS or router
navarro [Thu, 6 Sep 2012 13:29:37 +0000 (15:29 +0200)]
Cosmetics. (Forgot from a previous renaming)
navarro [Thu, 6 Sep 2012 13:28:40 +0000 (15:28 +0200)]
Finalize properties on AS.
navarro [Mon, 3 Sep 2012 11:29:29 +0000 (13:29 +0200)]
Add properties to AS structure (Not work for now) and use SG_PLATF_AS_INITIALIZER
navarro [Wed, 29 Aug 2012 10:11:56 +0000 (12:11 +0200)]
Oups, take the information in the argument structure.
navarro [Wed, 29 Aug 2012 10:10:43 +0000 (12:10 +0200)]
Rewrite the init of bypass mecanism (tks Martin;)
navarro [Tue, 28 Aug 2012 12:51:46 +0000 (14:51 +0200)]
Completely remove route_t and replace by sg_platf_route_cbarg_t.
schnorr [Wed, 22 Aug 2012 21:39:39 +0000 (23:39 +0200)]
[trace] avoiding key clashes on paje links
schnorr [Wed, 22 Aug 2012 21:32:10 +0000 (23:32 +0200)]
[trace] only fixing some tesh'es, paje header has changed
Augustin Degomme [Tue, 21 Aug 2012 16:11:43 +0000 (18:11 +0200)]
add a little sed script to convert a paje trace (generated with basic smpi flag) into a vite-readable paje trace
Augustin Degomme [Tue, 21 Aug 2012 16:10:50 +0000 (18:10 +0200)]
forgot to add color for computing in previous commit
schnorr [Tue, 21 Aug 2012 13:37:37 +0000 (15:37 +0200)]
[trace] color fields must have type 'color'
Augustin Degomme [Tue, 21 Aug 2012 13:26:57 +0000 (15:26 +0200)]
add option to generate states for code outside smpi to allow computation timing (not completely accurate as some smpi calls are not traced yet)
jean-noel quintin [Tue, 21 Aug 2012 10:51:04 +0000 (11:51 +0100)]
Merge remote-tracking branch 'origin/master'
jean-noel quintin [Tue, 21 Aug 2012 10:46:15 +0000 (11:46 +0100)]
[smpi,simix-network] remove a useless function
schnorr [Mon, 20 Aug 2012 15:41:53 +0000 (17:41 +0200)]
[trace] for some reason, simulated time changed in smpi -> fix tesh then
schnorr [Mon, 20 Aug 2012 15:39:26 +0000 (17:39 +0200)]
[trace] updating tesh'es according to header changes
details:
- albeit recent modifications on the header, only a white
space after the events number identification was removed
schnorr [Mon, 20 Aug 2012 15:33:25 +0000 (17:33 +0200)]
[trace] header re-organization (now with a basic version)
details:
you can generate a basic version of the header by passing
--cfg=tracing/basic:1 when launching your simulator (considering
you are using other arguments to enable the tracing)
schnorr [Mon, 20 Aug 2012 13:30:14 +0000 (15:30 +0200)]
[trace] fix color for MPI operations to correctly follow color specification
details:
http://paje.sourceforge.net/download/publication/lang-paje.pdf
schnorr [Mon, 20 Aug 2012 13:25:20 +0000 (15:25 +0200)]
[trace] new option tracing/basic (used to generate impoverished trace files)
details:
- this option should be used with visualization tools that are
incapable to read the complete Paje file format specification, as
detailed in this file:
http://paje.sourceforge.net/download/publication/lang-paje.pdf
Marion Guthmuller [Sat, 18 Aug 2012 16:48:28 +0000 (18:48 +0200)]
model-checker : init result of simcalls for snapshot comparison done during simcall
Marion Guthmuller [Fri, 17 Aug 2012 07:55:59 +0000 (09:55 +0200)]
model-checker : try first to associate each block or fragment with the same positiion in the other heap
Marion Guthmuller [Thu, 16 Aug 2012 22:32:52 +0000 (00:32 +0200)]
model-checker : remove unnecessary step
Marion Guthmuller [Thu, 16 Aug 2012 22:24:49 +0000 (00:24 +0200)]
model-checker : equal_to field of blocks/fragments updated twice with match_equals
Marion Guthmuller [Thu, 16 Aug 2012 22:19:08 +0000 (00:19 +0200)]
modelc-checker : reset list of blocks/fragments pairs associated
Marion Guthmuller [Thu, 16 Aug 2012 22:13:52 +0000 (00:13 +0200)]
model-checker : init equal_to field to -1 for each allocated block/fragment at the beginning of the heap comparison
Marion Guthmuller [Wed, 8 Aug 2012 10:06:33 +0000 (12:06 +0200)]
model-checker : set size used to 0 when free fragment
Martin Quinson [Thu, 16 Aug 2012 12:14:26 +0000 (14:14 +0200)]
make it clear that this platform only documents the NREN, not the whole computational platform
Martin Quinson [Wed, 15 Aug 2012 22:54:51 +0000 (00:54 +0200)]
dafuk did I just see?
Martin Quinson [Wed, 15 Aug 2012 22:37:53 +0000 (00:37 +0200)]
if this line induce a bug (see commit
05cd623), then kill it
Jean-Baptiste Hervé [Fri, 10 Aug 2012 15:52:53 +0000 (17:52 +0200)]
Add a master/slave example using a generated platform, and using generated event traces
This example fails at the end for the moment, apparently because of autorestart.
Jean-Baptiste Hervé [Fri, 10 Aug 2012 15:48:42 +0000 (17:48 +0200)]
Event trace generation: Change interface of constructors
The previous interface was not intuitive, even for me. I hope this one is better.
Replace 1 function by 3, and add the choice of the first value for state trace.
Jean-Baptiste Hervé [Fri, 10 Aug 2012 13:38:46 +0000 (15:38 +0200)]
Platform generation: add a bittorent example using a generated platform
Jean-Baptiste Hervé [Fri, 10 Aug 2012 12:31:48 +0000 (14:31 +0200)]
Platform generation: fex modification of the master/slave example
Change the topology generation parameter, so the program will have to retry the topology generation, as the first produced graph is not connected.
Correct a typo mistake.
Jean-Baptiste Hervé [Fri, 10 Aug 2012 12:29:24 +0000 (14:29 +0200)]
Platform generation: fix the remove links function
Jean-Baptiste Hervé [Fri, 10 Aug 2012 12:12:42 +0000 (14:12 +0200)]
Platform generation: record the router id
Augustin Degomme [Fri, 10 Aug 2012 08:55:18 +0000 (10:55 +0200)]
deactivate change induced by re7d52b71aae7 in logging for smpi which caused problems, waiting for a better idea
Augustin Degomme [Fri, 10 Aug 2012 08:53:24 +0000 (10:53 +0200)]
now that the simcall has been created, use it
Jean-Baptiste Hervé [Thu, 9 Aug 2012 15:36:23 +0000 (17:36 +0200)]
Platform generation: add a master/slave example using a generated platform
Cristian Rosa [Thu, 9 Aug 2012 13:17:14 +0000 (10:17 -0300)]
Add a new simcall to get/set the receiver of a rendez-vous point.
Jean-Baptiste Hervé [Thu, 9 Aug 2012 11:56:44 +0000 (13:56 +0200)]
Platform generation: little modifications to the connectedness check function
"Speed improvement" is not strong enough to describe the difference...
Jean-Baptiste Hervé [Thu, 9 Aug 2012 09:56:56 +0000 (11:56 +0200)]
Platform generation: "clique" topology does not need to duplicate links
Jean-Baptiste Hervé [Thu, 9 Aug 2012 09:15:16 +0000 (11:15 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Jean-Baptiste Hervé [Thu, 9 Aug 2012 09:14:42 +0000 (11:14 +0200)]
Platform generation: generate routes
Jean-Baptiste Hervé [Thu, 9 Aug 2012 09:13:56 +0000 (11:13 +0200)]
Platform generation: record the router id given while generation
Martin Quinson [Thu, 9 Aug 2012 07:59:35 +0000 (09:59 +0200)]
Various hacks to unbench the logging to disk when run in SMPI
- First use of weak symbols aliases in our source tree :-/
I hope it won't induce portability issues
- Only very partially tested, sorry.
Read: it compiles, that's all I know
Augustin Degomme [Wed, 8 Aug 2012 15:58:02 +0000 (17:58 +0200)]
add getter/setter functions in simix to access more cleanly the process associated to the mailbox for permanent receive mode
Jean-Baptiste Hervé [Wed, 8 Aug 2012 14:49:00 +0000 (16:49 +0200)]
Platform generation: Make link labeling work
Also check if the link is labeled when the callbacks have been called.
Augustin Degomme [Wed, 8 Aug 2012 13:37:17 +0000 (15:37 +0200)]
add missing includes
Jean-Baptiste Hervé [Wed, 8 Aug 2012 12:57:22 +0000 (14:57 +0200)]
Platform generation: add some documentation
Jean-Baptiste Hervé [Wed, 8 Aug 2012 12:56:42 +0000 (14:56 +0200)]
Trace generation: add some documentation
navarro [Wed, 8 Aug 2012 12:41:22 +0000 (14:41 +0200)]
Factorize route ASroute bypassRoute and bypassASroute from sg_platf.
Replace char* gw_src/dst for edge.
navarro [Wed, 8 Aug 2012 12:08:58 +0000 (14:08 +0200)]
First example for bypass platform wit sg_platf_new* functions.
Augustin Degomme [Wed, 8 Aug 2012 09:18:40 +0000 (11:18 +0200)]
Add a MSG_mailbox_set_async function to allow to set a mailbox to eagerly receive messages (the message is sent without waiting for the receive)
navarro [Wed, 8 Aug 2012 08:33:54 +0000 (10:33 +0200)]
Remove link_list from surf_routing.c and add it into the structure of *route and *bypassRoute
Marion Guthmuller [Wed, 8 Aug 2012 07:50:52 +0000 (09:50 +0200)]
model-checker : MC_ignore_init in MSG for ucontext
Augustin Degomme [Tue, 7 Aug 2012 16:07:59 +0000 (18:07 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Augustin Degomme [Tue, 7 Aug 2012 15:56:58 +0000 (17:56 +0200)]
add an "smpi/async_small_thres" configure option to smpi, to allow to toy with the threshold used to determine whether a message has to be sent eagerly (without waiting for the receiver) or not. The default value is 0, to deactivate this behavior for now. (note: this is independent from the "detached" state that is for now used for messages < 64*1024 bytes, which should become a new config option also)
Augustin Degomme [Tue, 7 Aug 2012 15:53:48 +0000 (17:53 +0200)]
add an immediate asynchronous send possibility for messages. This can be set by adding a permanent receiver to a mailbox. This will be used by default in smpi (commit to come), and proposed for msg
Conflicts:
src/simix/smx_network.c
navarro [Tue, 7 Aug 2012 15:45:13 +0000 (17:45 +0200)]
Remove unused structure and function
Augustin Degomme [Tue, 7 Aug 2012 12:08:21 +0000 (14:08 +0200)]
fix a bug causing surf to sometimes loop forever
navarro [Tue, 7 Aug 2012 14:32:00 +0000 (16:32 +0200)]
Completely remove surfxml_callback.
Clean the way to create arg for the process.
Jean-Baptiste Hervé [Tue, 7 Aug 2012 09:44:39 +0000 (11:44 +0200)]
Platform generation : REALLY check if the graph is connected
navarro [Mon, 6 Aug 2012 16:28:01 +0000 (18:28 +0200)]
More cleaning for surfxml callback (property and trace*)
navarro [Mon, 6 Aug 2012 14:35:49 +0000 (16:35 +0200)]
CleanUp unused xbt_dynar_t from callback
Jean-Baptiste Hervé [Mon, 6 Aug 2012 16:00:39 +0000 (18:00 +0200)]
More condensed debug...
Jean-Baptiste Hervé [Mon, 6 Aug 2012 15:32:08 +0000 (17:32 +0200)]
Event generator: Correct a mistake and add tiny debug
navarro [Mon, 6 Aug 2012 09:38:35 +0000 (11:38 +0200)]
Add function sg_platf_new_* for *route and bypass*route.
Jean-Baptiste Hervé [Mon, 6 Aug 2012 09:23:00 +0000 (11:23 +0200)]
Platform generation : Add a function to remove links in the graph
Jean-Baptiste Hervé [Mon, 6 Aug 2012 07:33:08 +0000 (09:33 +0200)]
Platform generation : add a function to check if the graph is connected
Martin Quinson [Sat, 4 Aug 2012 15:41:13 +0000 (17:41 +0200)]
and now, get the execution sampling macros work also when the user gives a maximal stderr AND a minimal execution count (hopefully)
Martin Quinson [Sat, 4 Aug 2012 15:15:42 +0000 (17:15 +0200)]
dafuk? don't mix seconds and flops!
SMPI_SAMPLE_DELAY parameter were an amount of flops. Consider it as a
delay instead.
Introduce SMPI_SAMPLE_FLOPS to report an amount of flops into the
simulo (and use it in the replay part of SMPI)