ports/openmpi
ports/openmpi
1
1
Форкнуть 0
Код Релизы Активность
3 623 коммитов 3 Ветки 0 Теги
Граф коммитов

3623 Коммитов

This Branch
This Branch
Все ветки
Автор SHA1 Сообщение Дата
Rainer Keller
cd8143dab9 Fix for the fix ,-] 
The remote_leader is guarantueed to be valid only at the local_leader rank.

This commit was SVN r4238.
 2005-01-29 03:32:15 +00:00
Rainer Keller
b8a35fe71f Fix a few off-by-one errors: the (remote) leader/process/root 
can not be of rank == comm_size.

This commit was SVN r4237.
 2005-01-29 03:01:28 +00:00
Jeff Squyres
97515c126d Fix bug potentially preventing the generation of some Makefiles 
This commit was SVN r4235.
 2005-01-29 02:09:15 +00:00
Rainer Keller
813a5b2062 Really(tm) fix the llm_hostfile_hosfile -> -hostfile issue. 
Thanks Jeff.

This commit was SVN r4234.
 2005-01-29 02:09:13 +00:00
Rainer Keller
f21536202f Reset the component parameter for hostfile selection from 
"hostfile" to the more conforming name "llm_hostfile_hostfile".

The shortcut for hostfile-selection with -hostfile, e.g.:
mpirun -np 98234 -hostfile big_machine ./test
still works.

This commit was SVN r4233.
 2005-01-29 01:58:29 +00:00
Rainer Keller
c8a13e27db Allow out of srcdir compilation. 
Here: add -I$(top_ompi_builddir)/include directory.

This commit was SVN r4232.
 2005-01-29 01:51:02 +00:00
Rainer Keller
77c54ddf69 Just a few spelling mistakes... 
This commit was SVN r4229.
 2005-01-29 00:57:40 +00:00
Brian Barrett
113fba12db * Fix bug 1127. Update powerpc 32 code to do 64bit compare and set when 
running on hardware with such support.  Since the calling convention is
  to use the registers as 32 bit values (even if they are 64 bit registers),
  we need to pack things up properly before doing the compare and set.

This commit was SVN r4226.
 2005-01-28 23:32:44 +00:00
Rainer Keller
f822ffc5ff Dough. Get rid of warnings 
This commit was SVN r4221.
 2005-01-28 20:58:38 +00:00
Rich Graham
418d184c45 starting to correct the logic for IB device selection. 
This commit was SVN r4217.
 2005-01-28 20:09:39 +00:00
Brian Barrett
79d4e8b60c * add some tests that explicitly disable any compiler inline of the assembly, 
to make sure the generated files also work

This commit was SVN r4215.
 2005-01-28 19:31:24 +00:00
Brian Barrett
891a0286d2 * add data line for PPC Linux 
This commit was SVN r4214.
 2005-01-28 19:16:04 +00:00
Jeff Squyres
6362a63459 Distinguish between success/failure and the number of requests 
progressed

This commit was SVN r4212.
 2005-01-28 18:54:34 +00:00
Jeff Squyres
ba3016bc1e Set ROMIO threads flags properly 
This commit was SVN r4211.
 2005-01-28 18:51:04 +00:00
Jeff Squyres
3b3279b267 Add clarification comment 
This commit was SVN r4210.
 2005-01-28 18:47:52 +00:00
Jeff Squyres
90fb658177 Fix minor type mismatch 
This commit was SVN r4209.
 2005-01-28 18:47:29 +00:00
Brian Barrett
75f38c5bd3 * fix nightly build - turns out we do need cache.h 
This commit was SVN r4208.
 2005-01-28 17:06:01 +00:00
George Bosilca
601836f2f5 Correctly detect if the datatype is contiguous. 
This commit was SVN r4203.
 2005-01-28 04:50:52 +00:00
Rich Graham
3fbe029333 Some versions of Mellanox (v3.1) require linking to the thread 
libraries.  So if we fail to find a symbol in the vapi library, try
looking for -lpthread, and if that succeeds, then try looking for a
second symbol in the vapi library.

This especially happens if you use Mellanox v3.1 and --without-threads
in the top-level configure.

This commit was SVN r4201.
 2005-01-28 01:49:33 +00:00
Brian Barrett
bf94b9df70 * Don't make a helper library, since it seems to play nice with some 
compilers

This commit was SVN r4200.
 2005-01-28 01:25:10 +00:00
Brian Barrett
b9ccc45008 * Move comment in powerpc atomic to where it should actually be 
* Remove old version of ifinit since no one has complained recently

This commit was SVN r4199.
 2005-01-28 01:10:50 +00:00
Brian Barrett
5955006dde Want noinst not lib libararies, as we want this in libmpi. 
This commit was SVN r4198.
 2005-01-28 01:06:47 +00:00
Rainer Keller
5c5cd76df3 Perl tries to create the generated assembler-ouput 
in directory src/asm/generated/, which for VPATH-builds
is at configure-time not yet created.
mkdir -p it...

This commit was SVN r4197.
 2005-01-28 00:43:49 +00:00
Brian Barrett
e724142970 use right macro for threads 
This commit was SVN r4196.
 2005-01-27 23:11:01 +00:00
Brian Barrett
f4b135a829 * make the 64/32 size warninsg go away 
This commit was SVN r4195.
 2005-01-27 22:59:31 +00:00
Brian Barrett
1e0e279de0 * add $(libs) to pick up -pthread if we need it 
This commit was SVN r4194.
 2005-01-27 22:53:36 +00:00
Brian Barrett
abee275d36 * update to run make check along the way of testing the nightly 
This commit was SVN r4193.
 2005-01-27 21:32:56 +00:00
Brian Barrett
85eda952b8 * add missing file 
This commit was SVN r4192.
 2005-01-27 21:16:45 +00:00
Brian Barrett
9f61d1d8a3 * Clean up some problems with the atomic operation interface discovered 
during testing.  In particular, we can't have a cmpset macro that can
  deal with pointers without some really evil voodoo.  So have a
  different macro for pointers
* Add more detailed testing of the atomics using AM's test framework.
  More to come...

This commit was SVN r4191.
 2005-01-27 21:08:35 +00:00
Jeff Squyres
b80698eff4 Fix a minor bug with the right rotate. 
This commit was SVN r4190.
 2005-01-27 18:11:36 +00:00
Jeff Squyres
4b2d3fcd2b Add svn:ignore 
This commit was SVN r4189.
 2005-01-27 17:19:36 +00:00
Brian Barrett
80967bec8d * fix make dist for the atomic code 
This commit was SVN r4188.
 2005-01-27 16:49:47 +00:00
Jeff Squyres
9dee6d8c03 Make the sym links a little more robust -- delete them before creating 
them (e.g., if the source tree was moved and they now point to bad
source files).

This commit was SVN r4187.
 2005-01-27 16:13:52 +00:00
Brian Barrett
9c1a277804 * Update build system to support non-GCC inline assembly, including detection 
of assembler format
* Fix minor bugs in AMD64, PPC, and IA32 assembly for atomic operations
* Clean up the #defines to look for when examining level of atomic operation
  support

This commit was SVN r4183.
 2005-01-27 01:39:55 +00:00
Jeff Squyres
b240395d9a Fix use of --with-ptl-*-libdir 
This commit was SVN r4182.
 2005-01-27 00:14:33 +00:00
Prabhanjan Kambadur
6075354290 Removing everything in CFLAGS which adds debug version of windows libraries. There are some issues with the way we use memory which results in perpetual assertion failures on windows 
This commit was SVN r4181.
 2005-01-27 00:13:44 +00:00
Prabhanjan Kambadur
e851aa9118 Now that teh dynamic loading is working, there is no need to have the components statically built 
This commit was SVN r4180.
 2005-01-27 00:11:34 +00:00
George Bosilca
cc87bde1ea Temporary commit until brian commit the big changes. The 64 bits now works (correctly set the order of low/high 
in the registers).

This commit was SVN r4177.
 2005-01-26 22:47:22 +00:00
Brian Barrett
c14a6d2bd3 * update to work on OS X 
This commit was SVN r4174.
 2005-01-26 20:05:44 +00:00
Prabhanjan Kambadur
b274a72e5e This fix makes it possible to install the components in the right directories 
This commit was SVN r4173.
 2005-01-26 19:43:39 +00:00
Prabhanjan Kambadur
228540071f Changing makefile to make sure that all the dynanmic components are built 
This commit was SVN r4171.
 2005-01-26 19:21:13 +00:00
Rainer Keller
6ee5a29c2f Add a Stacktrace feature, which figures where/what signal has happened 
after MPI-startup.
For this a new mpirun-parameter "mpi_signal" is added, one may specify a
comma-separated list of signals to grab, e.g. mpirun --mca mpi_signal 8,11
will check for SIGFPE and SIGSEGV.
It only finds the first fault (SA_ONESHOT), as after the return the same
fault will occur again.

As printout, the data provided by siginfo_t is printed to STDOUT (yes,
it calls printf ,-]).
Additionally, with glibc, it uses backtrace and backtrace_symbols to 
print the calling stack up to the function in which the signal was raised:

(Rank:0) Going to write to RD_ONLY mmaped shared mem
Signal:11 info.si_errno:0(Success) si_code:2(SEGV_ACCERR)
Failing at addr:0x4020c000
[0] func:/home/rusraink/ompi-gcc/lib/libmpi.so.0 [0x40121afe]
[1] func:./t0 [0x42029180]
[2] func:./t0(__libc_start_main+0x95) [0x42017589]
[3] func:./t0(__libc_start_main+0x49) [0x8048691]

This commit was SVN r4170.
 2005-01-26 19:11:46 +00:00
Prabhanjan Kambadur
46c2c11680 Function prototypes and global symbols need to be protected from C++ name mangling 
This commit was SVN r4165.
 2005-01-26 18:23:26 +00:00
Jeff Squyres
ba77c72820 Update some more svn:ignores 
This commit was SVN r4157.
 2005-01-26 07:42:57 +00:00
Jeff Squyres
9efcdd6149 Try harder not to find an F77 compiler in the F90 section. 
This commit was SVN r4156.
 2005-01-26 07:41:06 +00:00
Jeff Squyres
b35652134f Oops -- update svn:ignore properly. 
This commit was SVN r4155.
 2005-01-26 07:40:42 +00:00
Jeff Squyres
d8c85d053b Update svn:ignore 
This commit was SVN r4154.
 2005-01-26 07:39:23 +00:00
Jeff Squyres
859c3566b1 Remove useless header file ompi.h (it was syntactic sugar for 
including a few other OMPI header files, most of which usually weren't
needed). 

This commit was SVN r4149.
 2005-01-26 05:07:18 +00:00
Jeff Squyres
7d9458b88c Remove an old test that looks like it was designed to protect the user 
from themselves -- but doesn't look valid.  For example, it's legal to
have $FC and $F77 to both be ifort (intel fortran compiler).

This commit was SVN r4148.
 2005-01-26 01:59:42 +00:00
Mitch Sukalski
6cddd3e189 fix for memory regions that are deregistered and then reregistered in ptl_ib_memory.c, and 
removing mpga.h include as it breaks in the latest Mellanox VAPI releases (actually, mpga.h
is moved to mpga/mpga.h...yuck!) 

This commit was SVN r4146.
 2005-01-26 01:24:23 +00:00