ports/openmpi
ports/openmpi
1
1
Форкнуть 0
Код Релизы Активность
4 466 коммитов 3 Ветки 0 Теги 130 MiB
0dd11c8c32
Граф коммитов

4466 Коммитов

This Branch
This Branch
Все ветки
Автор SHA1 Сообщение Дата
Jeff Squyres
3ea253bbb2 Updates to orterun: 
- Print error messages with the basename(argv[0]) rather than
  hard-coded argv[0] so that you can see an error message beginning
  with "mpirun" when you run mpirun, etc.
- For all processes that died due to a signal:
  - If the signal was not SIGKILL, display the first N of them (where
    N defaults to 1)
  - If more than N processes died due to a non-SIGKILL signal, print
    "And X more processes aborted..." kind of message
  - Add --aborted command line parameter to change the default value
    of N
  - Also print out the total number of processes that died due to
    SIGKILL, with a disclaimer that it's impossible to know if we
    killed them or someone else killed them

This commit was SVN r5406.
 2005-04-15 21:52:58 +00:00
George Bosilca
ac987a5533 Remove some compilation warnings. 
This commit was SVN r5405.
 2005-04-15 21:36:32 +00:00
Jeff Squyres
7ad0d67a9d After consultation with Tim: when we set the ABORTED state for a proc, 
also increase the count on TERMINATED so that orterun can know when
*all* processes have died.

This commit was SVN r5404.
 2005-04-15 21:35:17 +00:00
Jeff Squyres
f9ef7d4657 Make the pls's clean up the session directory of each process that dies. 
This commit was SVN r5403.
 2005-04-15 21:34:07 +00:00
Brian Barrett
3a55cffdcd * add missing files from pty commit 
This commit was SVN r5402.
 2005-04-15 21:23:25 +00:00
George Bosilca
c6e97182cf Dont go back from the progress if there are still events. 
This commit was SVN r5401.
 2005-04-15 21:22:46 +00:00
Brian Barrett
e3587652b7 * Add support for using ptys for stdout when doing I/O forwarding. This is 
enough to make use applications be line buffered instead of block buffered,
  which makes output come much faster :)

This commit was SVN r5400.
 2005-04-15 21:18:20 +00:00
George Bosilca
fe70c2ceed Correctly compute the displacement in the stack. 
This commit was SVN r5399.
 2005-04-15 20:48:51 +00:00
George Bosilca
d7a66701a5 As this function can be used with datatype without an optimized representation compute the count by 
taking in account the size of he predefined datatype.

This commit was SVN r5398.
 2005-04-15 20:47:56 +00:00
Graham Fagg
8be3558757 missed const include 
This commit was SVN r5394.
 2005-04-15 17:40:41 +00:00
Tim Woodall
a8e13d9636 correct return value 
This commit was SVN r5393.
 2005-04-15 17:19:01 +00:00
Tim Woodall
a831729d6f split close into finalize/close so that rmgr can finalize all 
sub-components prior to entering close. moved pls logic to
wait on children from close to finalize.

This commit was SVN r5392.
 2005-04-15 17:04:57 +00:00
Jeff Squyres
1581a8feb4 If MPI_INIT hasn't been called or MPI_FINALIZE has been called, then 
just abort().

This commit was SVN r5391.
 2005-04-15 16:38:44 +00:00
Edgar Gabriel
78c00921f6 saving the current status before leaving to the airport 
This commit was SVN r5390.
 2005-04-15 16:20:30 +00:00
George Bosilca
7fd378f5d2 If everything went just fine return OMPI_SUCCESS. 
This commit was SVN r5389.
 2005-04-15 16:16:44 +00:00
George Bosilca
1ae368311a Do not allow the PML to cache the GM requests by now. I'll work on that later. 
This commit was SVN r5388.
 2005-04-15 16:13:43 +00:00
George Bosilca
478f7a226d Just indentation :) 
This commit was SVN r5387.
 2005-04-15 16:12:30 +00:00
George Bosilca
f70f4871ae If they try to create a struct with ZERO members just return an empty datatype. 
This commit was SVN r5386.
 2005-04-15 16:11:43 +00:00
George Bosilca
1f79196d78 NULL pointers are accepted if there is no associated data to retrieve. 
This commit was SVN r5385.
 2005-04-15 16:01:39 +00:00
George Bosilca
64e06bae52 Do not allow to stop the counting in the middle of a basic datatype. 
This commit was SVN r5384.
 2005-04-15 16:00:35 +00:00
George Bosilca
c3866709fa MPI_BOTTOM is a correct argument for pack/unpack at least for the part that have a datatype attached. 
This commit was SVN r5383.
 2005-04-15 15:58:59 +00:00
Edgar Gabriel
b1df79aedb these functions always have to return MPI_SUCCESS and not the error code passed to them. 
This commit was SVN r5382.
 2005-04-15 15:30:57 +00:00
George Bosilca
1010fd3690 For contifuous datatype the true_lb should be taken in account in order to receive the data in the 
correct memory location.

This commit was SVN r5381.
 2005-04-15 15:21:54 +00:00
Edgar Gabriel
c59b09a6a7 minor fix in graph_map and cart_map: in case the communicator passed to the functions 
does not have a topo-modules assigned (yet), we return as newrank their rank in the original commmunicator. This is the solution suggested by the MPI spec.

This commit was SVN r5380.
 2005-04-15 15:15:29 +00:00
Edgar Gabriel
b28ca88025 remove a stupid free() statement 
This commit was SVN r5378.
 2005-04-15 14:34:40 +00:00
Tim Woodall
70210a961e dont try to free the send request if it completes while in the 
scheduler - defer this until the scheduler completes

This commit was SVN r5377.
 2005-04-15 13:57:12 +00:00
Josh Hursey
e264707c64 final commit for environ changes for windows build... 
This commit was SVN r5376.
 2005-04-15 13:46:08 +00:00
Tim Woodall
020ecd2360 allow max_slots == slots 
This commit was SVN r5375.
 2005-04-15 13:42:14 +00:00
Tim Woodall
744e924d76 fix typo 
This commit was SVN r5374.
 2005-04-15 13:36:41 +00:00
Jeff Squyres
61f55f1011 Fix the problem with the nightly unit tests -- do *not* override 
OMPI_ENABLE_DEBUG because that changes the size of struct's (e.g.,
ompi_object) in the unit tests as compared to what may have been
compiled in the library.

This commit was SVN r5373.
 2005-04-15 13:32:18 +00:00
Tim Woodall
f4454823ff vastly simplify allocator 
This commit was SVN r5372.
 2005-04-15 13:28:17 +00:00
Josh Hursey
2b25046781 Make some environ naming changes. This was motivated by a Windows 
naming conflict.

This commit was SVN r5371.
 2005-04-15 13:16:02 +00:00
Tim Woodall
280122e4a6 check for no ptls availables - force an error to be returned on send 
This commit was SVN r5370.
 2005-04-15 13:13:39 +00:00
Jeff Squyres
a6937e4b91 Update svn:ignore 
This commit was SVN r5369.
 2005-04-15 09:47:14 +00:00
Brian Barrett
b31baccd74 * need to strdup as the destructor will free() 
This commit was SVN r5368.
 2005-04-14 22:13:12 +00:00
Jeff Squyres
81043fa139 - Rename environ.c -> ompi_environ.c to avoid filename clashes on 
windows
- If the real environ is passed into ompi_setenv(), turn around and do
  a putenv() (which is a potential memory leak -- noted in comments in
  ompi_environ.c and doxygen comments in environ.h)

This commit was SVN r5367.
 2005-04-14 21:31:45 +00:00
Edgar Gabriel
52d3e1f876 checkpoint of the hierarch collective module. 
This commit was SVN r5366.
 2005-04-14 21:27:24 +00:00
Jeff Squyres
cba83190af Remove extra environ declaration 
This commit was SVN r5365.
 2005-04-14 21:20:51 +00:00
Jeff Squyres
3f8e29136a Remove useless extern for environ 
This commit was SVN r5364.
 2005-04-14 21:19:15 +00:00
Tim Woodall
e157491589 back out this change from the am 
This commit was SVN r5363.
 2005-04-14 21:10:47 +00:00
Tim Woodall
ef20da3d12 cleanup environment 
This commit was SVN r5362.
 2005-04-14 21:08:36 +00:00
Tim Woodall
eb7bf1ca58 post non-blocking recv before processing callbacks - as callback 
could block - waiting for a gpr subscription

This commit was SVN r5361.
 2005-04-14 21:08:03 +00:00
Jeff Squyres
0ac08d5f49 A bunch of f90 build process fixes: 
- ensure to dist mpi_kinds.f90
- move a bunch of things inside the AM_CONDITIONAL for building f90
  stuff
  - make mpi-f90-interfaces.h depend on fortran_kinds.sh (the output
    from configure), so that if configure is run again with a different
    --with-90-max-array-dim, the file will be re-generated
  - only generate the .f90 files if we're building the f90 library
- remove some old kruft
- fully remove the use of BUILT_SOURCES

This commit was SVN r5353.
 2005-04-14 20:54:47 +00:00
Jeff Squyres
2a9c10260c Ensure that if we don't find what we're unregistering that we don't 
try to shrink the array.

This commit was SVN r5350.
 2005-04-14 20:10:27 +00:00
Jeff Squyres
7153cc6dbd Updates 
This commit was SVN r5346.
 2005-04-14 19:37:22 +00:00
Josh Hursey
78ae61c535 minor define for windows build 
This commit was SVN r5344.
 2005-04-14 19:04:16 +00:00
Brian Barrett
815964f491 * start of code to dynamically decide whether to have the sched_yield or not 
in the ompi_progress() function.  Not yet fully operational, as there 
  is something going on with the subscription not getting called.  E-mail
  coming to those interested.

This commit was SVN r5343.
 2005-04-14 18:55:53 +00:00
Tim Woodall
9b57669593 not an error if the endpoint already exists 
This commit was SVN r5342.
 2005-04-14 18:54:55 +00:00
Brian Barrett
d22594ec48 * set default value for the want MPI param checking test to be the value of 
the configure-time param checking variable, rather than "true".  Removes
  the warning about not being able to enable MPI param checking during
  MPI_INIT.

This commit was SVN r5341.
 2005-04-14 16:35:56 +00:00
Edgar Gabriel
6b6aa6431a remove the warnings which George pointed out 
This commit was SVN r5340.
 2005-04-14 16:28:15 +00:00