1
1
Граф коммитов

5824 Коммитов

Автор SHA1 Сообщение Дата
George Bosilca
e501768f10 Sleep is defined in the unistd.h file, do not forget to include it on the dependencies list.
This commit was SVN r7301.
2005-09-11 20:05:17 +00:00
George Bosilca
361ff6640f Correct the progress thread function name from opal_progress_thread to opal_event_progress_thread.
This commit was SVN r7300.
2005-09-11 20:04:40 +00:00
George Bosilca
c24eb702bb Correctly compute the default sizes for the fragments.
This commit was SVN r7299.
2005-09-11 20:02:55 +00:00
George Bosilca
5b1303edef Correctly handle the convertor flags.
This commit was SVN r7298.
2005-09-11 18:51:47 +00:00
George Bosilca
6b2d154df3 Add more debug messages to the pack and unpack functions.
This commit was SVN r7297.
2005-09-11 18:51:06 +00:00
George Bosilca
95497b156e Cleanup the flags when we clone or copy a convertor. Add safety checks for the set_position function.
This commit was SVN r7296.
2005-09-11 18:49:46 +00:00
George Bosilca
df26395a83 Add some mask to cleanup the convertor flags.
This commit was SVN r7295.
2005-09-11 18:48:50 +00:00
Jeff Squyres
a107ab3897 Add missing header file
This commit was SVN r7294.
2005-09-11 10:14:29 +00:00
Brian Barrett
1fcf18c211 non-persistent signal behavior isn't quite right, so use the proper SIGNAL
macros and deregister at the appropriate time.

This commit was SVN r7293.
2005-09-10 23:22:37 +00:00
Jeff Squyres
c4731285ee protect state var with the same #if that the rest of the code is
protected with

This commit was SVN r7292.
2005-09-10 11:17:01 +00:00
Jeff Squyres
31ea80cd0b Simplify the poe configure.m4 -- if we're on AIX, always build it. If
we're not on AIX, don't build it.

This commit was SVN r7291.
2005-09-10 10:40:26 +00:00
Jeff Squyres
7e4c23d88c Add missing header file
This commit was SVN r7290.
2005-09-10 10:39:15 +00:00
Rainer Keller
3c639efa38 - Silly cleanup
This commit was SVN r7289.
2005-09-10 08:01:47 +00:00
Rainer Keller
5fed46e072 - Allow usernames to be specified in the hostfile.
The following formats are parsed:
    user@IPv4
    user@fqdn
    IPv4 or fqdn [username|user-name|user_name]=user
- Try a better error-detection when parsing (recognize wrong
  IPs, fqdns...)

This commit was SVN r7288.
2005-09-10 07:57:50 +00:00
George Bosilca
2e3764a181 Remove the unused variable p + something that SVN figure out and where I dont see any difference,
certainly just some conversion from TAB to space ...

This commit was SVN r7287.
2005-09-10 04:06:49 +00:00
George Bosilca
41c2dc6a79 Call the new unpack function from the new_unpack.c
The cleanup of the pack/unpack framework will follow shortly.

This commit was SVN r7286.
2005-09-09 23:38:31 +00:00
George Bosilca
35676f0c75 Do not forget to update the max_data before returning to the upper level.
This commit was SVN r7285.
2005-09-09 23:37:44 +00:00
George Bosilca
06d9c594ef A new unpack framework based on the pack. Quicker and simpler and hopefully less
error prone.

This commit was SVN r7284.
2005-09-09 23:37:02 +00:00
Graham Fagg
00842ad87c Fixed reduce bugs
(added some temp debug routines)

This commit was SVN r7283.
2005-09-09 23:05:17 +00:00
George Bosilca
0ad973afdd Do not modex sned and receive the proc architecture. This is now done outside the PML in the proc
init and was added there few weeks ago by Ralph.

This commit was SVN r7282.
2005-09-09 22:21:57 +00:00
Jeff Squyres
e8da20a6ae Fix VPATH build
This commit was SVN r7281.
2005-09-09 22:20:37 +00:00
George Bosilca
43e37060a8 Correct a unfortunate typo in the fortran constants ... We still have untested code inside :)
Add a macro to convert from Fortran to C MPI_IN_PLACE.

This commit was SVN r7276.
2005-09-09 20:46:47 +00:00
George Bosilca
56f8607439 Adding MPI_IN_PLACE for both languages C and Fortran. Still working on the C++ and maybe F90 stuff. It's easilly
getting ugly ...

This commit was SVN r7275.
2005-09-09 20:44:04 +00:00
George Bosilca
a4c70dd4c3 The trace macro take one argument even if we do not intent to use it. Otherwise some compilers complain and generate errors.
This commit was SVN r7271.
2005-09-09 19:59:33 +00:00
Ralph Castain
76ccec0cee Upgrade the new opal trace system to utilize verbosity. Begin building the trace command into the ORTE system.
This commit was SVN r7267.
2005-09-09 18:27:17 +00:00
Jeff Squyres
0b255830e0 Slight change to the defaults -- make the prefix "output-pid<pid>-" so
that multiple processes don't overwrite each other.  Change that
default in orte_init_stage1() to just "output-" (because the file will
be in a process-unique directory at that point; the pid is no longer
necessary).

This commit was SVN r7256.
2005-09-09 12:37:39 +00:00
Jeff Squyres
86cca77046 Clarification on docs
This commit was SVN r7255.
2005-09-09 12:25:55 +00:00
Jeff Squyres
7eadfc4bdc Add a new hook function into the opal_output system:
opal_output_set_output_file_info().  This allows getting and setting
the default directory where output stream files will be opened (for
all *new* streams).  Before this function is not invoked, the default
location is $TMPDIR or $HOME (if $TMPDIR is not defined).  

Added a call into orte_init_stage1() to call this function
immediately after the session directory is created and set the default
location of stream files to be the process' session directory.

This commit was SVN r7254.
2005-09-09 12:18:39 +00:00
Jeff Squyres
4aa75fa739 - Make opal_output_stream_t be a real opal_object_t so that it can use
a constructor, like the rest of the code base
- Convert usage in the tree to use the constructor to zero out an
  instance of opal_output_stream_t
- Still need to re-enable output files

This commit was SVN r7253.
2005-09-09 10:46:54 +00:00
Jeff Squyres
8ec6721f24 Prior fix for the ROMIO re-#define problem didn't really fix the
problem because Autoconf replaced the "#undef ..." with "#define
...".  Fix this by not putting the "#undef ..." statement directly in
romioconf.h[.in] -- but rather having romioconf.h[.in] #include
romioconf-undefs.h, which has the #undef statements.

This commit was SVN r7252.
2005-09-09 10:33:04 +00:00
Jeff Squyres
841b199167 Update svn:ignore
This commit was SVN r7251.
2005-09-09 10:14:38 +00:00
Jeff Squyres
d19e5b4af8 Remove unused variable
This commit was SVN r7250.
2005-09-09 10:11:46 +00:00
Jeff Squyres
eeee36aa73 Style changes only (this was originally written before many of the
style guidelines existed) so that the next patch concentrates on
fixes / functionality changes

This commit was SVN r7249.
2005-09-09 09:38:43 +00:00
George Bosilca
778468a5b6 Add comment about how to compile these 2 tests.
This commit was SVN r7248.
2005-09-09 06:17:19 +00:00
George Bosilca
b2bb2393fc Remove all useless debug stuff.
This commit was SVN r7247.
2005-09-09 06:16:44 +00:00
George Bosilca
63fb0f3341 Correct the eager problem for some devices. The bug was because we stop packing too early, not on the
data where I have to start the next time, but on the place where I detect there is no more space in the
iovecs. That lead to a wrong computation of the displacement on the next call to pack.

Additionally, I change some macros on statis inline functions. They allow me to go inside step by step
with a debugger. They will go back to macros one I'm 100% sure they cannot generate any more errors.

This commit was SVN r7246.
2005-09-09 06:07:37 +00:00
George Bosilca
8280ce8f6d A new flag has been added. To detect when the convertor has finish its work a complex condition
was computed. This informations can be updated on the convertor flag by the pack/unpack function,
making the test a lot simpler. The pack/unpack functions will be modified soon.

This commit was SVN r7245.
2005-09-09 06:04:29 +00:00
George Bosilca
3693816d08 Update the old datatype test and add a new one based on the test Tim wrote.
This commit was SVN r7244.
2005-09-09 06:01:59 +00:00
Graham Fagg
9596d6e2f3 use correct names or cannot call functions...
use correct ddt calls (and params ext vs lb etc)

This commit was SVN r7243.
2005-09-09 02:39:57 +00:00
Galen Shipman
e2186a1cc2 Updates to the rb rcache..
This commit was SVN r7238.
2005-09-08 21:34:26 +00:00
Galen Shipman
ec2b99986f remove rcache_base_init.c no longer needed..
This commit was SVN r7237.
2005-09-08 16:31:17 +00:00
Ralph Castain
7fbe575edd Make sure rc is initialized.
This commit was SVN r7233.
2005-09-08 13:20:38 +00:00
Jeff Squyres
eb25be25be Add missing "\"
This commit was SVN r7232.
2005-09-08 10:02:21 +00:00
Thara Angskun
e3feddfdd4 Move from stone aged configure.stub to configure.m4
This commit was SVN r7231.
2005-09-08 09:49:21 +00:00
Jeff Squyres
1d20f800ba Add another flag to MPI_Op's -- whether it is associative with
floating point numbers or not (e.g., MAX is, but SUM is not).

This commit was SVN r7230.
2005-09-08 09:47:27 +00:00
Jeff Squyres
67cde6c212 - Minor cleanups
- Add --enable-trace which turns on some internal tracing and dumps a
  file per process in the session directory tree.  Meant for internal
  developer tracing, *not* for tracing MPI applications in the
  traditional sense.

This commit was SVN r7229.
2005-09-08 09:44:50 +00:00
Jeff Squyres
e8b120fbf2 - Add some #undef's so that we don't conflict with the
AC-automatically-generated macros
- Remove a useless #include

This commit was SVN r7228.
2005-09-08 09:42:19 +00:00
Galen Shipman
6c8e9ab661 more rcache changes
This commit was SVN r7227.
2005-09-07 20:37:17 +00:00
Galen Shipman
a0b82acece Add fleshed out open,create and close
This commit was SVN r7226.
2005-09-07 20:21:15 +00:00
George Bosilca
d27f17f406 Solve the reference count problem. As the MPI functions always use the ompi_ddt_set_args
and ompi_ddt_release_args functions, increase the reference count for the used datatypes here
not in the dt_add.c. Tus, we have a correct reference count for all data-type independing if
the datatypes are used in MPI or directly in others modules.

This commit was SVN r7225.
2005-09-07 20:10:24 +00:00