seem to pass now the test without coredumps etc.
Checking for MPI_Alltollw(intra)...working
Checking for MPI_Exscan:
testing MPI_MAX...........working
testing MPI_MIN...........working
testing MPI_SUM...........working
testing MPI_PROD..........working
testing MPI_LAND..........working
testing MPI_LOR...........working
testing MPI_LXOR..........working
testing MPI_BAND..........working
testing MPI_BOR...........working
testing MPI_BXOR..........working
Checking for MPI_Bcast.............working
Checking for MPI_Allreduce.........working
Checking for MPI_Reduce............working
Checking for MPI_Barrier...........working
Checking for MPI_Gather............working
Checking for MPI_Gatherv...........working
Checking for MPI_Scatter...........working
Checking for MPI_Scatterv..........working
Checking for MPI_Allgather.........working
Checking for MPI_Allgatherv........working
Checking for MPI_Alltoall..........working
Checking for MPI_Alltoallv.........working
Checking for MPI_Alltollw(inter)...working
Checking for MPI_Reduce_scatter....working
Checking for MPI_Bcast.............working
Checking for MPI_Allreduce.........working
Checking for MPI_Reduce............working
Checking for MPI_Barrier...........working
Checking for MPI_Gather............working
Checking for MPI_Gatherv...........working
Checking for MPI_Scatter...........working
Checking for MPI_Scatterv..........working
Checking for MPI_Allgather.........working
Checking for MPI_Allgatherv........working
Checking for MPI_Alltoall..........working
Checking for MPI_Alltoallv.........working
Checking for MPI_Alltollw(inter)...working
Checking for MPI_Reduce_scatter....working
This commit was SVN r1903.
- because we had the scenario, that fragments for a communicator, which was not
not yet set up on all procs, arrived and caused problem, we introduced
a comm_activate function call, which executes a kind of barrier (using
the allreduce functions used for the comm_cid allocation).
Setting up the coll-component has moved *after* this barrier, since
some coll-modules (e.g. the MagPIe component) might want to communicate
using this communicator already (e.g. a comm_split).
- adding a new file comm_dyn.c, which basically abstracts the required functionality
for connect-accept, and therefore is the 'magic' code (from the MPI point of view)
for all dynamically created communicators.
This commit was SVN r1900.
since it is easier to specify the data type for an int (in both
OOB terms as well as in MPI terms), if we have to send the result
data length around.
This commit was SVN r1899.
* Add hostfile component for the LLM (reads hostfiles, returns array of
node identifiers
NOTES:
- This will require the full autogen / configure / make.
- You now need flex to build Open MPI from Subversion. The versions
available on most Linux boxen and OS X is more than new enough. You
do *not* need flex to build from a nightly or release tarball.
This commit was SVN r1890.
(e.g., Portland compilers don't have/know about <stdbool.h>, but
AC_CHECK_HEADERS will typically find the GNU/gcc <stdbool.h>).
Reviewed by Brian.
This commit was SVN r1889.
All the user interface functions are now in mca/oob/base/base.h
Anyone who uses the oob should just include this file.
All component related functions have been moved to mca/oob/oob.h
The reason for this change was to make the user interface more clear.
This commit was SVN r1884.
-properly initialize variables in the oob_tcp_msg struct
-properly close peer sockets in the tcp oob
-fix compare in bucket allocator to use the correct variable
-remove duplicate free in teg
-updated the oob tests
-add more output to tcp oob when there are failures
This commit was SVN r1866.
They do not work yet necessarily (although some tests pass already),
but they do not break anything as far as I can see and anticipate.
This commit was SVN r1857.
