ports/openmpi
ports/openmpi
1
1
Форкнуть 0
Код Релизы Активность
1 344 коммитов 3 Ветки 0 Теги 130 MiB
39a341be52
Граф коммитов

1344 Коммитов

This Branch
This Branch
Все ветки
Автор SHA1 Сообщение Дата
Edgar Gabriel
5f570a8b62 modified the value of MPI_COMM_NULL to two, since we need to have a valid 
entry in the f_to_c translation table for this object. We still wanted
to be MPI_COMM_WORLD zero and MPI_COMM_SELF 1.

Reviewed and agreed by Jeff.

This commit was SVN r1146.
 2004-05-21 19:30:39 +00:00
Ralph Castain
2e866b9d46 Provide an updated version of sys_info that does a little more error checking. It returns LAM_ERROR if there is a problem, and LAM_SUCCESS if everything goes okay. Also, the system information has been extended to include the path separator character, usually either a '\' or '/', depending upon operating system. 
This utility is planned to be called early (like perhaps in mpi_init) to fill the global variable with the relevant information. Since this info might change if the process migrates, we will update it with another call to ompi_sys_info as part of the migration process. If you want to call it yourself, however, there is no harm done - it will just update the global variable and merrily sail along.

This commit was SVN r1145.
 2004-05-20 21:28:59 +00:00
Edgar Gabriel
48df884e40 added a function to check for the availability of an element in the 
array-list and reserve it, if available. Needed for the communicator
cid allocation.
 
Reviewed by Tim.

This commit was SVN r1144.
 2004-05-20 21:13:11 +00:00
Ralph Castain
8cc3ecb8e5 Remove the runtime-environment directory RTE/universe from the makefile list since it only exists on my machine! Update the sys_info header file 
This commit was SVN r1143.
 2004-05-20 17:38:54 +00:00
Ralph Castain
2d709f768a This adds into the system a utility for detecting what kind of system we are on. The utility provides a global variable ompi_system_info that contains information on the operating system, release and version, machine architecture, and network node name. I'll provide additional documentation via Doxygen once I figure out how to put it into the .h file. 
Code was reviewed by Jeff - thanks Jeff!

This commit was SVN r1141.
 2004-05-20 13:54:14 +00:00
Tim Woodall
9fa9cd9b0a changed to PML/PTL I/F to support both put/get semantics 
and receiver side scheduling

This commit was SVN r1140.
 2004-05-20 13:45:34 +00:00
Jeff Squyres
63fb9ea1ee Move the docs from the .c to the .h file (I think the docs were 
originally written before we decided to put all docs in .h files)

This commit was SVN r1139.
 2004-05-20 12:39:27 +00:00
Jeff Squyres
e8e3646c9d Allow lds==NULL on lam_output_open(), giving a default set of output 
destinations (i.e., stderr)

This commit was SVN r1137.
 2004-05-19 19:52:24 +00:00
Tim Woodall
e68af0dde8 - inlining multiple functions - changed to macros functions the compiler wouldn't inline 
- removed unused event handling code from critical path
- fixed threading issue in tcp code

This commit was SVN r1136.
 2004-05-18 21:06:11 +00:00
Jeff Squyres
c39ba9ea6e Add -finline-functions if we're gcc and it works 
This commit was SVN r1135.
 2004-05-18 14:14:12 +00:00
Edgar Gabriel
aab81a9261 Added a function to lookup a lam_proc_t struct based on the vpid. 
Used currently in intercomm_create functions.

Reviewed by Tim.

This commit was SVN r1134.
 2004-05-17 21:28:32 +00:00
Jeff Squyres
c951247181 Fix a few typos 
This commit was SVN r1133.
 2004-05-17 17:56:21 +00:00
Ralph Castain
c04a7efd96 Although I understand Jeff's explanations about why/when autogen needs to be run, I prefer the lazy approach and offer this script as a way to avoid having to worry about the silly thing. 
This commit was SVN r1132.
 2004-05-14 22:11:08 +00:00
Ralph Castain
905a84901d Update the INSTALL file with instructions on how to compile and run this software. 
This commit was SVN r1131.
 2004-05-12 05:00:53 +00:00
Rich Graham
d08f246990 initialize grp_ok_to_free for group_emty, and group_null. 
Reviewed by Edgar G.

This commit was SVN r1130.
 2004-05-11 22:04:39 +00:00
Rich Graham
784693826a add the grp_ok_to_free flag indicating if it is ok to actually 
free the group.  By default this is set to true, with special
code setting this to false for the groups associated with
COMM_NULL, COMM_SELF, and COMM_WORLD.  Also, moved the reference
count decrement on the procs to the group destructor.

Reviewed by Ralph C.

This commit was SVN r1129.
 2004-05-11 21:28:24 +00:00
Jeff Squyres
05f8a6c999 Update the checklist 
This commit was SVN r1128.
 2004-05-08 11:16:24 +00:00
Vishal Sahay
62e5cbea60 Remove typos and add one more defn 
This commit was SVN r1127.
 2004-05-07 23:45:40 +00:00
Vishal Sahay
148689e3d1 front end funcitions for coll and some dt functions 
This commit was SVN r1126.
 2004-05-07 23:44:49 +00:00
Vishal Sahay
037d7e39fc Correcct a typo 
This commit was SVN r1125.
 2004-05-07 23:43:49 +00:00
Vishal Sahay
a9eb043310 Correctly take in the coll pointers 
This commit was SVN r1124.
 2004-05-07 23:42:54 +00:00
Vishal Sahay
dd430f905f Datatype changes 
This commit was SVN r1123.
 2004-05-07 23:41:30 +00:00
Vishal Sahay
21408ee4d3 mpi init changes for coll 
This commit was SVN r1122.
 2004-05-07 23:23:03 +00:00
Vishal Sahay
ec7b437428 Add more functionality in comm 
This commit was SVN r1121.
 2004-05-07 23:15:10 +00:00
Vishal Sahay
c01488ab28 Collective MPI function's back end implementation 
This commit was SVN r1120.
 2004-05-07 23:09:55 +00:00
Vishal Sahay
13b2738768 Collective base functions 
This commit was SVN r1119.
 2004-05-07 23:04:59 +00:00
Jeff Squyres
67ec10c46c Update svn:ignore for these directories 
This commit was SVN r1118.
 2004-05-06 10:51:08 +00:00
David Daniel
588d08a999 Adding caveat 
This commit was SVN r1117.
 2004-05-05 23:22:30 +00:00
David Daniel
c174b2ca7c Adding more atomic stuff before I vanish... 
This commit was SVN r1116.
 2004-05-05 23:19:32 +00:00
David Daniel
aa36122c66 Adding current status of atomic operations. 
Not complete, tested or reviewed.

Currently ifdef-ed out so should not impact build.

This commit was SVN r1115.
 2004-05-05 19:19:47 +00:00
Tim Woodall
208a2e3ec0 allow the socket buffers to default to kernel values 
This commit was SVN r1114.
 2004-05-04 22:51:04 +00:00
Tim Woodall
9e4c239720 fixed a bug for the threaded case that required a select/poll 
timeout (10secs) before new connections are noticed

This commit was SVN r1113.
 2004-05-04 22:45:09 +00:00
Tim Woodall
0326edc437 fixed bug - need to decrement list length when item is removed 
This commit was SVN r1112.
 2004-05-04 22:06:49 +00:00
Jeff Squyres
7f10fd0d50 Minor corrections to documentation 
This commit was SVN r1111.
 2004-05-03 14:58:38 +00:00
George Bosilca
1f1bfee77f Quick implementation for MPI_Get_address. 
This commit was SVN r1110.
 2004-04-30 20:36:40 +00:00
George Bosilca
85b2f9ba5f Add default error handlers to the predefined communicators. 
This commit was SVN r1109.
 2004-04-30 20:31:29 +00:00
George Bosilca
94a84cbaff I correct the creation of the MPI_ERROR_ARE_FATAL, MPI_ERRHANDLER_NULL and MPI_ERRORS_RETURN variables. Less code but same effect as the initializer for the lam_errhandler_t insert the errhandler in the array as the previsious code. The only big difference is that now the error handler class is correctly derived from the parent. 
This commit was SVN r1108.
 2004-04-30 20:11:07 +00:00
George Bosilca
d86e6858ff Introduce mpi_error_check parameter to detect at runtime if the user want to check the consistency of MPI functions arguments. 
This commit was SVN r1107.
 2004-04-30 20:09:11 +00:00
George Bosilca
a8047b9229 Use LAM_ENABLE_MEM_PROFILING correctly. Now we can handle the configure options --disable-mem-profile and --disable-meme-debug. 
This commit was SVN r1106.
 2004-04-30 20:07:47 +00:00
George Bosilca
152ac14685 Improuve the computation of the required space for MPI_Type_struct. 
Change size_t to unsigned int, as I really need just a 32 bits integer here.

This commit was SVN r1105.
 2004-04-30 20:06:11 +00:00
George Bosilca
daac7a1788 Correct 2 typos. 
This commit was SVN r1104.
 2004-04-30 20:05:02 +00:00
George Bosilca
8502661385 Update atomic operations for x86 64 bits architectures. 
This commit was SVN r1103.
 2004-04-30 20:04:11 +00:00
Jeff Squyres
e4bbefd57b Fix for bug 845: elminiated an #if/#endif block that is only necessary 
if we're going to suport non-ANSI compilers (in which case we would
need a corresponding #else block with a set of non-ANSI typedefs),
which we're not going to do.  :-)

This commit was SVN r1102.
 2004-04-28 01:52:56 +00:00
Jeff Squyres
7a81bb118c Add missing "\" that causes problems with some "make"'s, including 
Solaris make

This commit was SVN r1101.
 2004-04-28 01:28:24 +00:00
George Bosilca
d45d74b36c Add the type for 2 more MPI predefined datatypes: lam_mpi_2cplex and lam_mpi_2dblcplex. 
This commit was SVN r1100.
 2004-04-27 18:26:05 +00:00
George Bosilca
3084aeeb59 Use the fake element to detect the end of data description (remove some tests about the position in the data description. 
Correctly define the boundaries in the data description loops.
Remove useless local variables ( like end_loop as it already present on the stack).
Increase the position in the datatype description BEFORE looking for the next data

This commit was SVN r1099.
 2004-04-27 18:19:03 +00:00
George Bosilca
e49b59e29c Remove useless local variables 
Add the fake last element in the data description.

This commit was SVN r1098.
 2004-04-27 18:14:16 +00:00
George Bosilca
2a6e129269 The fake element will be added in lam_ddt_commit function. 
This commit was SVN r1097.
 2004-04-27 18:13:16 +00:00
George Bosilca
a82a47a33e Remove some memory leaks in the test program. (Thanks valgrind). 
This commit was SVN r1096.
 2004-04-27 18:12:33 +00:00
George Bosilca
258677c453 Add the lam_complex_long_double_t typedef. 
Add a test to abort the compilation if the DT_MAX_PREFEDINED does not match the expected value.
Add some additional MPI2 predefined datatypes.

This commit was SVN r1095.
 2004-04-27 18:11:49 +00:00