ports/openmpi
ports/openmpi
1
1
Форкнуть 0
Код Релизы Активность
6 605 коммитов 3 Ветки 0 Теги 130 MiB
d2897b54d7
Граф коммитов

6605 Коммитов

This Branch
This Branch
Все ветки
Автор SHA1 Сообщение Дата
Jeff Squyres
d2897b54d7 Sync with 1.0 NEWS 
This commit was SVN r8652.
 2006-01-05 14:55:00 +00:00
Jeff Squyres
b2de55d72e Back out some debugging stuff from a careless r8643 commit (only 
intended to include the OMPI_DEBUG_ZERO call).

These debugging statements should not have affected correcteness
because the value of 78 will be overridden in the read() and the
assert()/abort() stuff will only be triggered on an error which should
never happen (i.e., the error should have been handled by the prior if
conditional).  But still, thise code should not be there.

This commit was SVN r8649.

The following SVN revision numbers were found above:
  r8643 --> open-mpi/ompi@a6b869ed68
 2006-01-05 14:44:10 +00:00
Jeff Squyres
268a5f5716 Ensure that seg is always initialized, even in the return_error case. 
This commit was SVN r8646.
 2006-01-04 23:01:50 +00:00
Jeff Squyres
a6b869ed68 Avoid a false positive in bcheck 
This commit was SVN r8643.
 2006-01-04 22:29:09 +00:00
Jeff Squyres
c3cc1d5bbc Avoid false positives in bcheck/valgrind -- there are holes in the 
mca_base_param_t struct, and it eventually ends up getting memcpy()'ed 

This commit was SVN r8642.
 2006-01-04 22:18:30 +00:00
Jeff Squyres
d9ec36c5fe Add OMPI_DEBUG_ZERO to a struct so that valgrind doesn't complain when 
looking for memory errors (i.e., it's ok for some fields to be
uninitialized when memcpy'ing the data)

This commit was SVN r8640.
 2006-01-04 20:57:57 +00:00
Tim Woodall
63d0438991 merge in changes from release branch 
This commit was SVN r8637.
 2006-01-04 16:34:45 +00:00
Tim Woodall
7eade5b856 added SIGFPE to default list of signals that generate a backtrace (where supported) 
This commit was SVN r8632.
 2006-01-04 16:02:45 +00:00
David Daniel
d272e02338 Need to include fcntl.h on linux -- protected for windows. 
This commit was SVN r8630.
 2006-01-04 00:54:16 +00:00
George Bosilca
29dd1a6f6b Unmap from map_addr not map_seg. 
This commit was SVN r8626.
 2005-12-31 16:11:58 +00:00
George Bosilca
68b0ba0f58 We need the copyright ... 
This commit was SVN r8625.
 2005-12-31 15:51:57 +00:00
George Bosilca
e1fbf7b0df The last commit remove the initialization of map_addr ... 
This commit was SVN r8624.
 2005-12-31 15:18:58 +00:00
George Bosilca
479d510eaf Use the common SM component to unmap the shared memory file. 
This commit was SVN r8623.
 2005-12-31 15:07:48 +00:00
George Bosilca
3baf36c6d9 Add the windows implementation or shared memory files. 
As the common part is the one that create the shred memory file it seems
logical to make it destroy the file as well. Therefore, the code for
unmapping the file is in a common place.

This commit was SVN r8622.
 2005-12-31 15:06:24 +00:00
George Bosilca
4002365f43 On windows we have include io.h in order to have access to the isatty function. 
This commit was SVN r8621.
 2005-12-31 15:03:58 +00:00
George Bosilca
228f966798 A little trick to force qsort to order the sm modules in the order we expect. 
On some systems (like windows) qsort can modify the order of modules when the
comparaison function return 0. As we expect to have the
mca_btl_sm_add_procs_same_base_addr called before mca_btl_sm_add_procs we have
force the qsort to return the SM modules in the correct order. Giving to the
same_addr module a slightly higher priority solve this problem.

This commit was SVN r8620.
 2005-12-31 14:59:53 +00:00
George Bosilca
b3c6c7b442 We need sys/types.h. 
This commit was SVN r8619.
 2005-12-31 14:55:39 +00:00
George Bosilca
4ae8c291c3 Add MPI_SIGNED_CHAR. 
This commit was SVN r8618.
 2005-12-31 14:54:57 +00:00
George Bosilca
7a88e72c1b Add more protections around the headers. 
This commit was SVN r8617.
 2005-12-31 12:35:24 +00:00
George Bosilca
d91650ea85 Do not use explicitly "ln -s" as on some systems it does not work properly ... 
(windows). Instead use the LN_S variable exported by the Makefile (set to
"ln -s" on all Unixes and to "cp -p" on windows).

When we remove an executable use the correct extension for its name
(add $(EXEEXT) to the name).

This commit was SVN r8616.
 2005-12-31 12:33:44 +00:00
George Bosilca
bb53ffda9c Include the windows compatibility header after declaring the OMPI_DECLSPEC 
define.

This commit was SVN r8615.
 2005-12-31 12:25:39 +00:00
George Bosilca
163391d71a Cleanup the win32 stuff. Add some more definition for UNIX compatibility. There 
is still an issue with the OMPI_DECLSPEC but I will investigate it a little
bit later.

This commit was SVN r8614.
 2005-12-31 12:23:28 +00:00
George Bosilca
3c95dd0801 No discrimination ! 
This commit was SVN r8613.
 2005-12-31 12:20:32 +00:00
George Bosilca
670716c489 There is no need to do the checking for socklen_t twice. 
This commit was SVN r8612.
 2005-12-31 12:19:27 +00:00
George Bosilca
208fa725c3 Add OMPI_DECLSPEC before all exported functions. 
This commit was SVN r8611.
 2005-12-31 12:18:15 +00:00
George Bosilca
9457534520 Add full path for the header files. 
This commit was SVN r8610.
 2005-12-28 19:28:32 +00:00
Brian Barrett
f23356aa02 * if we don't have mmap hook support, don't expect to get a callback from 
mmap().

This commit was SVN r8609.
 2005-12-24 05:34:51 +00:00
Jeff Squyres
152d5c1e99 Remove unnecessary dynamic-mca tree (was already removed in opal and 
orte).

This commit was SVN r8608.
 2005-12-23 17:36:44 +00:00
Jeff Squyres
5e9f756290 Remove Makefile and Makefile.in from svn:ignore in a bunch of 
directories where these files are no longer generated

This commit was SVN r8607.
 2005-12-23 17:18:10 +00:00
Jeff Squyres
5f96a74e33 Make user-defined MPI::Op's be thread safe (the previous 
implementation was not thread safe).  See lengthy comment in
ompi/mpi/cxx/intercepts.cc::ompi_mpi_cxx_op_intercept() for a full
explanation.

This commit was SVN r8606.
 2005-12-23 16:49:09 +00:00
Jeff Squyres
06b01b3b76 Fix uninitialized variable 
This commit was SVN r8605.
 2005-12-23 13:02:36 +00:00
Brian Barrett
45a3eccec6 re-enable the stack-trace functionality (it had been inadvertently disabled 
in a merge a long time ago).  Set default signals to print a stack trace from
none to SIGBUS,SIGSEGV.

This commit was SVN r8603.
 2005-12-22 19:39:50 +00:00
Tim Woodall
4ff5316b2d correct copy/paste error 
This commit was SVN r8601.
 2005-12-22 18:07:46 +00:00
Tim Woodall
5d91c492d6 improve diagnostic error messages 
This commit was SVN r8600.
 2005-12-22 18:01:47 +00:00
Tim Woodall
e1a49793f7 improve diagnostics when registration fails 
This commit was SVN r8599.
 2005-12-22 16:32:12 +00:00
Tim Woodall
e9498f7a75 improve error reporting when registrations fail 
This commit was SVN r8598.
 2005-12-22 16:05:28 +00:00
Jeff Squyres
d05b4d6c82 Remove a bunch of Makefile and Makefile.in entries from svn:ignore 
properties for Makefile's that are no longer generated (because
they're now included by top-level Makefile.am's)

This commit was SVN r8597.
 2005-12-22 14:39:23 +00:00
Jeff Squyres
1e93c78e2e - Rename rsh component members: argv->agent_argv, argc->agent_argc, 
and path->agent_path so that it's totally clear what these are for
- make a new rsh component param for agent_param (the value from the
  MCA param)
- delay the path check for the agent until the component init -- don't
  make it fail during open, because the MCA base will print a warning
  if a component fails open() (e.g., on clusters without rsh/ssh (!),
  this component was failing noisly even though it was
  normal/expected)

This commit was SVN r8596.
 2005-12-22 14:37:19 +00:00
Jeff Squyres
8fb5e506aa Arrgh -- should have included this in last commit: need to set a 
variable before we += it in a Makefile.am.

This commit was SVN r8595.
 2005-12-22 14:30:27 +00:00
Jeff Squyres
93b4d12d14 Add a friendly help message if no pls components are found to be 
available.

This commit was SVN r8594.
 2005-12-22 14:29:45 +00:00
Tim Woodall
0c57e2d091 correct typo 
This commit was SVN r8593.
 2005-12-22 14:28:13 +00:00
Brian Barrett
f74e24df69 * don't declare functions being checked as extern char * - that's not a 
function and causes problems with false negatives on some linux distros.
  Instead, use the test code used by AC_CHECK_FUNC.

This commit was SVN r8592.
 2005-12-22 14:26:24 +00:00
Jeff Squyres
4395d440dc Apparently we have not been following the new prefix rule in the OMPI 
layer -- public symbols in components should begin with ompi_
(vs. mca_).

This commit was SVN r8590.
 2005-12-22 13:50:15 +00:00
Jeff Squyres
54c4bd3ce2 Update to have public symbols be consistent; use new prefix rule 
(apparently we've been doing this in opal and orte, but not in ompi
yet).  All public symbols begin with "ompi_coll_tuned_" (not
mca_coll_tuned_) except the component struct.  Now this component
passes the illegal symbol report with no hits.

This commit was SVN r8589.
 2005-12-22 13:49:33 +00:00
Jeff Squyres
efe84971ce Correct copyrights and some typos 
This commit was SVN r8588.
 2005-12-22 05:37:28 +00:00
Jeff Squyres
6770facf59 Remove "static inline" from a prototype of a function that isn't 
static inline.

This commit was SVN r8587.
 2005-12-22 05:34:36 +00:00
George Bosilca
b88f2a3c32 More protection around the atomic operations on Windows. Still ahve to find 
a way to detect them correctly without popping up a Windows error message box.

This commit was SVN r8586.
 2005-12-22 00:13:09 +00:00
George Bosilca
b6036d985b Updated version of the windows compatibility file. 
This commit was SVN r8585.
 2005-12-22 00:10:56 +00:00
Jeff Squyres
4a061b7c78 Since all of the collectives (except scan and exscan) are defined on 
both intercomm and intracomm, put them up in MPI::Comm rather than
duplicating them in MPI::Intercomm and MPI::Intracomm.

This commit was SVN r8583.
 2005-12-21 21:12:52 +00:00
Tim Woodall
3c170c410c changes required by dr 
This commit was SVN r8580.
 2005-12-21 15:11:40 +00:00