Finally! It's in the usual place:
http://www.open-mpi.org/software/ompi/v1.6/
Please test, especially with low-registered-memory-available scenarios with
Mellanox OpenFabrics devices.
Here's a list of changes since 1.6:
- A bunch of changes to eliminate hangs on OpenFabrics-based networks.
Users with Mellanox hardware are ***STRONGLY ENCOURAGED*** to check
their registered memory kernel module settings to ensure that the OS
will allow registering more than 8GB of memory. See this FAQ item
for details:
http://www.open-mpi.org/faq/?category=openfabrics#ib-low-reg-mem
- Fall back to send/receive semantics if registered memory is
unavilable for RDMA.
- Fix two fragment leaks when registered memory is exhausted.
- Hueristically determine how much registered memory is available
and warn if it's significantly less than all of RAM.
- Artifically limit the amount of registered memory each MPI process
can use to about 1/Nth to total registered memory available.
- Improve error messages when events occur that are likely due to
unexpected registered memory exhaustion.
- Remove the last openib default per-peer receive queue
specification (and make it an SRQ).
- Switch the MPI_ALLTOALLV default algorithm to a pairwise exchange.
- Increase the openib BTL default CQ length to handle more types of
OpenFabrics devices.
- Lots of VampirTrace fixes; upgrade to v5.13.0.4.
- Map MPI_2INTEGER to underlying MPI_INTEGERs, not MPI_INTs.
- Ensure that the OMPI version number is toleant of handling spaces.
Thanks to dragonboy for identifying the issue.
- Fixed IN parameter marking on Fortran "mpi" module
MPI_COMM_TEST_INTER interface.
- Various MXM improvements.
- Make the output of "mpirun --report-bindings" much more friendly /
human-readable.
- Properly handle MPI_COMPLEX8|16|32.
- More fixes for mpirun's processor affinity options (--bind-to-core
and friends).
- Use aligned memory for OpenFabrics registered memory.
- Multiple fixes for parameter checking in MPI_ALLGATHERV,
MPI_REDUCE_SCATTER, MPI_SCATTERV, and MPI_GATHERV. Thanks to the
mpi4py community (Bennet Fauber, Lisandro Dalcin, Jonathan Dursi).
- Fixed file positioning overflows in MPI_FILE_GET_POSITION,
MPI_FILE_GET_POSITION_SHARED, FILE_GET_SIZE, FILE_GET_VIEW.
- Removed the broken --cpu-set mpirun option.
- Fix cleanup of MPI errorcodes. Thanks to Alexey Bayduraev for the
patch.
- Fix default hostfile location. Thanks to Götz Waschk for noticing
the issue.
- Improve several error messages.
--
Jeff Squyres
jsquy...@cisco.com
For corporate legal information go to:
http://www.cisco.com/web/about/doing_business/legal/cri/
