cristianrosa [Mon, 13 Dec 2010 14:10:22 +0000 (14:10 +0000)]
Add parmap to raw contexts
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9152
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Mon, 13 Dec 2010 12:59:14 +0000 (12:59 +0000)]
more debugging message
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9151
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Mon, 13 Dec 2010 12:59:12 +0000 (12:59 +0000)]
Do not answer the request twice
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9150
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Mon, 13 Dec 2010 12:41:08 +0000 (12:41 +0000)]
verbose message about the type of context factory activated
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9149
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Mon, 13 Dec 2010 12:40:56 +0000 (12:40 +0000)]
function to see the name of syscall types in debugging messages (and use it on SIMIX_request_push)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9148
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Mon, 13 Dec 2010 12:40:54 +0000 (12:40 +0000)]
more verbose on problem
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9147
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
navarrop [Mon, 13 Dec 2010 12:09:44 +0000 (12:09 +0000)]
Add an hide option "enable_pcre" for enable or not pcre lib.
By default = on.
Use for personal test only.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9146
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Mon, 13 Dec 2010 10:57:18 +0000 (10:57 +0000)]
Fix tne runall functions
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9145
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Mon, 13 Dec 2010 10:57:16 +0000 (10:57 +0000)]
Remove a useless instruction
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9144
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
cristianrosa [Mon, 13 Dec 2010 08:58:22 +0000 (08:58 +0000)]
Use a dynar instead of a swag to implement the process_to_run list.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9143
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
cristianrosa [Mon, 13 Dec 2010 08:58:18 +0000 (08:58 +0000)]
Transform the thread pool into a parallel map data structure.
It better suits our needs.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9142
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
cristianrosa [Mon, 13 Dec 2010 08:58:07 +0000 (08:58 +0000)]
Remove the threadpool from the context factories.
It was way too slow.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9141
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
coldpeace [Fri, 10 Dec 2010 17:19:19 +0000 (17:19 +0000)]
fix warning msg...
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9140
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
coldpeace [Fri, 10 Dec 2010 17:12:45 +0000 (17:12 +0000)]
Fix quit_notify function
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9139
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
coldpeace [Fri, 10 Dec 2010 16:55:59 +0000 (16:55 +0000)]
add a deadline condition to finish properly the simulation when all nodes leave the chord ring (and start an implementation for the function leave(node_t)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9138
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
coldpeace [Fri, 10 Dec 2010 16:51:30 +0000 (16:51 +0000)]
Chord Deployment File :add a deadline value for each node to let them leave the ring peacefuly...
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9137
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Fri, 10 Dec 2010 14:54:31 +0000 (14:54 +0000)]
Don't do anything if the category is NULL
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9136
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Fri, 10 Dec 2010 14:45:47 +0000 (14:45 +0000)]
Remove debugging fprintf
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9135
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Fri, 10 Dec 2010 14:45:46 +0000 (14:45 +0000)]
Fix tesh files of gras-rpc-sg and gras-synchro-sg
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9134
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Fri, 10 Dec 2010 13:08:47 +0000 (13:08 +0000)]
and now, test that the file I commit actually works before doing so. Sorry for the noise
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9133
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Fri, 10 Dec 2010 13:02:48 +0000 (13:02 +0000)]
quote the arguments to --log since they are not quoted enough in tesh files to support interpretation by the shell
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9132
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Fri, 10 Dec 2010 13:02:46 +0000 (13:02 +0000)]
put pid on 2 chars to help sorting, and revalidate the output (timestamps changed, be we don't care)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9131
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Fri, 10 Dec 2010 12:57:07 +0000 (12:57 +0000)]
Update tesh files of gras-simple_token-sg
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9130
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Fri, 10 Dec 2010 12:51:52 +0000 (12:51 +0000)]
bugfix in trace replay: cannot use split_in_place if only one trace file: memory will be reclaimed while still being used
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9129
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Fri, 10 Dec 2010 12:51:50 +0000 (12:51 +0000)]
reduce verbosity when activating sexy new contextes
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9128
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
pini [Fri, 10 Dec 2010 12:38:56 +0000 (12:38 +0000)]
Do not use a gap if route has length 0.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9127
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Fri, 10 Dec 2010 12:10:55 +0000 (12:10 +0000)]
Update tesh files of gras-pmm-sg
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9126
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Fri, 10 Dec 2010 11:26:06 +0000 (11:26 +0000)]
[tesh] allow to specify the length of the sorting prefix
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9125
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Fri, 10 Dec 2010 10:55:09 +0000 (10:55 +0000)]
Initialize the process data in MSG before actually creating it,
because SIMIX may start its execution before returning from
our call to SIMIX_req_process_create().
This fixes the test msg-suspend.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9124
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Fri, 10 Dec 2010 10:14:22 +0000 (10:14 +0000)]
Update smpi-tracing-ptp tesh file
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9123
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Fri, 10 Dec 2010 10:14:21 +0000 (10:14 +0000)]
Reimplement SIMIX_display_process_status()
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9122
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Fri, 10 Dec 2010 10:14:20 +0000 (10:14 +0000)]
Setting process->waiting_task doesn't make sense for isend
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9121
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Fri, 10 Dec 2010 10:14:19 +0000 (10:14 +0000)]
Cosmetics
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9120
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Fri, 10 Dec 2010 10:14:16 +0000 (10:14 +0000)]
Free some memory
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9119
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
suter [Fri, 10 Dec 2010 09:32:58 +0000 (09:32 +0000)]
revalidation of the test with the line that warns users of the libc hack
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9118
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Thu, 9 Dec 2010 18:09:10 +0000 (18:09 +0000)]
Chord: the join operations work
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9117
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Thu, 9 Dec 2010 16:36:23 +0000 (16:36 +0000)]
fix raw contextes for supernovae and for parallel execution
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9116
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Thu, 9 Dec 2010 16:36:13 +0000 (16:36 +0000)]
a few more machines
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9115
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Thu, 9 Dec 2010 16:36:11 +0000 (16:36 +0000)]
Get verbose when activating the sexy new contextes (and activate them for replay)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9114
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
pini [Thu, 9 Dec 2010 15:54:19 +0000 (15:54 +0000)]
Saved one memory access per context swap.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9113
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
cristianrosa [Thu, 9 Dec 2010 15:31:10 +0000 (15:31 +0000)]
Remove the fork based checkpoint functionality.
Use the old one based on snapshots.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9112
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
cristianrosa [Thu, 9 Dec 2010 15:31:08 +0000 (15:31 +0000)]
On MC mode, return immediately after handling the wait request.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9111
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
cristianrosa [Thu, 9 Dec 2010 15:31:01 +0000 (15:31 +0000)]
Do not use XBT log on unit test.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9110
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
pini [Thu, 9 Dec 2010 15:30:17 +0000 (15:30 +0000)]
Output the assembly function in the text section.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9109
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
pini [Thu, 9 Dec 2010 15:12:40 +0000 (15:12 +0000)]
Added support for x86_64 raw contexts + some comments.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9108
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
pini [Thu, 9 Dec 2010 13:37:12 +0000 (13:37 +0000)]
Throw an error at runtime instead of breaking compilation.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9107
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
pini [Thu, 9 Dec 2010 13:24:41 +0000 (13:24 +0000)]
Added context switching support for x86 architectures.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9106
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Thu, 9 Dec 2010 13:08:23 +0000 (13:08 +0000)]
some preprocessor macros to detect which flavor of assembly we want today
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9105
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Thu, 9 Dec 2010 11:16:21 +0000 (11:16 +0000)]
draft of hardcore ultra brutal manual context switching (TBC after lunch)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9104
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Thu, 9 Dec 2010 10:10:55 +0000 (10:10 +0000)]
Working on Chord
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9103
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
cristianrosa [Wed, 8 Dec 2010 22:08:35 +0000 (22:08 +0000)]
Bugfix: save the ucontext of each running thread in a per context basis.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9102
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
cristianrosa [Wed, 8 Dec 2010 22:08:32 +0000 (22:08 +0000)]
Rename condition variable + improve debugging
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9101
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
cristianrosa [Wed, 8 Dec 2010 22:08:29 +0000 (22:08 +0000)]
Compile fix: smx_process_t->request is a s_smx_req_t now.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9100
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Wed, 8 Dec 2010 20:14:22 +0000 (20:14 +0000)]
New trace replayer, as fast as hell, and as boring as hell, too
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9099
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Wed, 8 Dec 2010 15:47:13 +0000 (15:47 +0000)]
Fix compilation warnings
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9098
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
navarrop [Wed, 8 Dec 2010 15:35:57 +0000 (15:35 +0000)]
Remove unused variables.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9097
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
navarrop [Wed, 8 Dec 2010 15:35:55 +0000 (15:35 +0000)]
Add availability_file and state_file to peer tag.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9096
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
coldpeace [Wed, 8 Dec 2010 15:07:41 +0000 (15:07 +0000)]
Fist implementation of a Chord example using MSG
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9095
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
navarrop [Wed, 8 Dec 2010 15:06:28 +0000 (15:06 +0000)]
Add availability_file and state_file in cluster.
Can use variable :
${id}
${prefix}
${suffix}
${radical}
example :
<cluster id="bob_cluster" prefix="bob" suffix=".hamburger.edu"
radical="0,2-4,6" power="
1000000000.00" bw="
125000000.00" lat="5e-5"
bb_bw="
2250000000.00" bb_lat="5e-4"
state_file="foo/${prefix}${radical}_state_file.trace"
availability_file="foo/${prefix}${radical}_availability_file.trace"/>
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9094
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
navarrop [Wed, 8 Dec 2010 15:06:27 +0000 (15:06 +0000)]
Remove old buildname and options.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9093
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Wed, 8 Dec 2010 15:00:56 +0000 (15:00 +0000)]
Starting to implement a Chord example with MSG
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9092
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Wed, 8 Dec 2010 15:00:54 +0000 (15:00 +0000)]
David's patch: replace != by strcmp
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9091
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
cristianrosa [Wed, 8 Dec 2010 13:54:59 +0000 (13:54 +0000)]
Bugfixes + test suitx
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9090
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
cristianrosa [Wed, 8 Dec 2010 13:54:49 +0000 (13:54 +0000)]
Include threadpool.h
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9089
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
cristianrosa [Wed, 8 Dec 2010 13:54:47 +0000 (13:54 +0000)]
Wait for all the context to finish it's execution
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9088
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
suter [Tue, 7 Dec 2010 22:03:23 +0000 (22:03 +0000)]
one & too much
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9087
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Tue, 7 Dec 2010 21:58:00 +0000 (21:58 +0000)]
last try for tonight to speed the trace replayer
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9086
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Tue, 7 Dec 2010 21:41:43 +0000 (21:41 +0000)]
new function: xbt_str_split_quoted_in_place (w/o mallocs)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9085
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Tue, 7 Dec 2010 21:41:20 +0000 (21:41 +0000)]
xbt_dynar_to_array now destroys the dynar. The caller is now only responsible of freeing the received array
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9084
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
suter [Tue, 7 Dec 2010 19:38:12 +0000 (19:38 +0000)]
revalidating the tesh outputs for ctest
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9083
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Tue, 7 Dec 2010 17:43:34 +0000 (17:43 +0000)]
do not store the requests on the process's stack nor in the big vector
Instead, they are directly in the s_smx_process_t structure.
There is no stack when going in replay contextes (still to be done)
And the vector may be realloced when creating processes, so saving
pointers to its content is not stable.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9082
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Tue, 7 Dec 2010 17:43:13 +0000 (17:43 +0000)]
rename a badly named variable
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9081
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Tue, 7 Dec 2010 17:42:51 +0000 (17:42 +0000)]
Do not traverse the whole array of request to find non-null ones, use a heap instead
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9080
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Tue, 7 Dec 2010 16:38:33 +0000 (16:38 +0000)]
Fix SIMIX calls
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9079
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Tue, 7 Dec 2010 16:38:32 +0000 (16:38 +0000)]
Make SIMIX_process_create() private
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9078
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Tue, 7 Dec 2010 15:56:51 +0000 (15:56 +0000)]
Use a raw array instead of a dynar to store the requests
It can be made big enough at every point, so that we don't have to
expend it on need automatically from the dynar stuff.
This was the only remaining reason why we had to lock that structure,
so kill the locking stuff too.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9077
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Tue, 7 Dec 2010 15:56:40 +0000 (15:56 +0000)]
new function (private to simix): SIMIX_process_get_maxpid
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9076
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Tue, 7 Dec 2010 15:56:38 +0000 (15:56 +0000)]
stop using private functions from simix; use requests instead
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9075
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
pini [Tue, 7 Dec 2010 15:47:21 +0000 (15:47 +0000)]
Build SMPI on top of smx_action_t.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9074
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
pini [Tue, 7 Dec 2010 15:47:19 +0000 (15:47 +0000)]
Distinguish access to sender-side and receiver-side user data in smx_action_t.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9073
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
cristianrosa [Tue, 7 Dec 2010 15:23:50 +0000 (15:23 +0000)]
No need to lock on request pop.
Only maestro does it and running alone.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9072
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
coldpeace [Tue, 7 Dec 2010 15:22:40 +0000 (15:22 +0000)]
no need to check, since the default value of core is set to 1 by the lua user code
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9071
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
coldpeace [Tue, 7 Dec 2010 15:21:10 +0000 (15:21 +0000)]
Fix lua console by adding new core element when bypass the parser
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9070
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Tue, 7 Dec 2010 15:18:43 +0000 (15:18 +0000)]
new function: MSG_action_paranoid_mode_set (allowing to tell the parser that our input is clean and that he don't need to be paranoid about it
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9069
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Tue, 7 Dec 2010 15:18:41 +0000 (15:18 +0000)]
do not trim in split_quoted, that's expensive, and the caller can do it if his input is not clean
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9068
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Tue, 7 Dec 2010 15:05:20 +0000 (15:05 +0000)]
Call elements_father() only if necessary
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9067
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Tue, 7 Dec 2010 14:24:27 +0000 (14:24 +0000)]
Fix more memory leaks
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9066
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Tue, 7 Dec 2010 14:15:09 +0000 (14:15 +0000)]
no need to lock the structure if we run in sequential
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9065
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Tue, 7 Dec 2010 14:15:07 +0000 (14:15 +0000)]
given the following line, this FIXME seems now done ;)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9064
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
cristianrosa [Tue, 7 Dec 2010 14:02:54 +0000 (14:02 +0000)]
Add support for custom communication matching to SIMIX network interface
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9063
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
cristianrosa [Tue, 7 Dec 2010 14:02:49 +0000 (14:02 +0000)]
Rework context factory's initialization logic.
Now if both ucontext and threads are detected, allow the parallel execution of ucontexts using a thread pool.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9062
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Tue, 7 Dec 2010 13:41:29 +0000 (13:41 +0000)]
MSG does not auto-destroy the MSG_comm_t objects anymore
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9061
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Tue, 7 Dec 2010 13:23:36 +0000 (13:23 +0000)]
Another useless and leaky xbt_strdup
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9060
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Tue, 7 Dec 2010 13:20:55 +0000 (13:20 +0000)]
implement actions irecv and wait w/o spawning; kill spawning functions
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9059
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
suter [Tue, 7 Dec 2010 12:26:51 +0000 (12:26 +0000)]
reimplement reduce and bcast (non tested) w/o spawns
uncomment the comm_destroy
modify the action files to add init and finalize
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9058
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Tue, 7 Dec 2010 12:23:40 +0000 (12:23 +0000)]
Don't auto-destroy the MSG_comm_t in MSG_comm_wait()
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9057
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Tue, 7 Dec 2010 12:23:39 +0000 (12:23 +0000)]
Don't auto-destroy the MSG_comm_t in MSG_comm_wait(), it's the
responsability of the user code
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9056
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Tue, 7 Dec 2010 12:23:38 +0000 (12:23 +0000)]
Fix memory leaks
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9055
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
thiery [Tue, 7 Dec 2010 12:23:37 +0000 (12:23 +0000)]
Remove leaky xbt_strdup: MSG_mailbox_new already does it (and takes a const char*)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9054
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
mquinson [Tue, 7 Dec 2010 12:16:38 +0000 (12:16 +0000)]
implement simix_comm_testany properly
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/simgrid/simgrid/trunk@9053
48e7efb5-ca39-0410-a469-
dd3cf9ba447f
