17.2. Open MPI v5.0.x series
This file contains all the NEWS updates for the Open MPI v5.0.x series, in reverse chronological order.
17.2.1. Open MPI version 5.0.0rc13
29 September 2023
The MPIR API has been removed
As was announced in the summer of 2017, Open MPI has removed support for MPIR-based tools beginning with the release of Open MPI v5.0.0.
Users who still need legacy MPIR support should see https://github.com/hpc/mpir-to-pmix-guide for more information.
Zlib is suggested for better performance
The Open MPI community recommends building PMIx with Zlib support, regardless of whether you are using an externally-installed PMIx or the bundled PMIx that is included with Open MPI distribution tarballs.
Note that while the Zlib library may be present on many systems by default, the Zlib header files — which are needed to build PMIx with Zlib support — may need to be installed separately before building PMIx.
Unlike previous versions of Open MPI, Open MPI 5.0 and later will prefer an external package that meets our version requirements, even if it is older than our internal version.
To simplify managing dependencies, any required packages that Open MPI v5.0.x bundles will be installed in Open MPI’s installation prefix, without name mangling.
For example, if a valid Libevent installation cannot be found and Open MPI therefore builds its bundled version, a
libevent.sowill be installed in Open MPI’s installation tree. This is different from previous releases, where Open MPI name-mangled the Libevent symbols and then statically pulled the library into
Changes since rc12:
Update PMIx to the
v4.2.6release tag. Hash:
Update PRRTE to the
v3.0.1release tag. Hash:
Lots of documentation updates.
Fixed parameter name in
MPI_Intercomm_merge. Thanks to Yan Wu for the report.
OFI: Update NIC selection to determine optimal interfaces from the current process.
Fix reordering of received data in
Disable builds with
HWLOCversions >= 3.0.0. This is currently not supported.
Fix re-ordering of ranks in
coll/HAN: Fix bug when using
MPI_Type_Dupto propagate errors from inner calls.
Fix the compilation of the monitoring infrastructure.
Various other bug fixes.
All other notable updates for v5.0.0:
ULFM Fault Tolerance support has been added. See the ULFM section.
CUDA is now supported in the
New MCA parameter
ompi_display_comm, enabling a communication report. When set to
mpi_init, display the report when
MPI_Init()is invoked. When set to
mpi_finalize, display the report during
A threading framework has been added to allow building Open MPI with different threading libraries. It currently supports Argobots, Qthreads, and Pthreads. See the
--with-threadsoption in the
configurecommand. Thanks to Shintaro Iwasaki and Jan Ciesko for their contributions to this effort.
New Thread Local Storage API: Removes global visibility of TLS structures and allows for dynamic TLS handling.
Acceleratorframework. CUDA-specific code was replaced with a generic framework that standardizes various device features such as copies or pointer type detection. This allows for modularized implementation of various devices such as the newly introduced ROCm Accelerator component. The redesign also allows for Open MPI builds to be shipped with CUDA support enabled without requiring CUDA libraries.
Added load-linked, store-conditional atomics support for AArch64.
Added atomicity support to the
osc/rdma: Added support for MPI minimum alignment key.
Add ability to detect patched memory to
memory_patcher. Thanks to Rich Welch for the contribution.
coll/ucc: Added support for the
MPI-4.0 updates and additions:
Support for MPI Sessions has been added.
Added partitioned communication using persistent sends and persistent receives.
Added persistent collectives to the
MPI_namespace (they were previously available via the
MPI_Isendrecv()and its variants.
Added support for
Added support for
Added support for
Added error handling for unbound errors to
MPI_Win_get_info()compliant to the standard.
Droped unknown/ignored info keys on communicators, files, and windows.
Initial implementations of
MPI_Info_get_valuelen()are now deprecated.
Issue a deprecation warning when
MPI_Cancel()is called for a non-blocking send request.
Transport updates and improvements
Many MPI one-sided and RDMA emulation fixes for the
This patch series fixs many issues when running with
--mca osc rdma --mca btl tcp, i.e., TCP support for one sided MPI calls.
Many MPI one-sided fixes for the
Added support for
acc_single_intrinsicto the one-sided
Removed the legacy
pt2ptone-sided component. Users should now utilize the
rdmaone-sided component instead. The
rdmacomponent will use BTL components — such as the TCP BTL — to effect one-sided communications.
tcpBTL to use graph solving for global interface matching between peers in order to improve
Improved support for the HPE SS11 network.
Added cache bypass mechanism. This fixes conflicts with Libfabric, which has its own registration cache. This adds a bypass flag which can be used for providers known to have their own registration cache.
sm(shared memory) BTL has been removed. The next-generation shared memory BTL
vaderreplaces it, and has been renamed to be
vaderwill still work as an alias).
Update the new
smBTL to not use Linux Cross Memory Attach (CMA) in user namespaces.
Fixed a crash when using the new
smBTL when compiled with Linux Cross Memory Attach (
XPMEM). Thanks to George Katevenis for reporting this issue.
-mca pmloption to only accept one PML, not a list.
Deprecations and removals:
ORTE, the underlying Open MPI launcher has been removed, and replaced with the PMIx Reference RunTime Environment (
PMI support has been removed from Open MPI; now only PMIx is supported. Thanks to Zach Osman for contributing.
The following components have been removed, and are replaced by UCX support: PML
psmcomponent has been removed and is no longer supported.
Removed all vestiges of Checkpoint Restart (C/R) support.
32 bit atomics are now only supported via C11 compliant compilers.
Explicitly disable support for GNU gcc < v4.8.1 (note: the default gcc compiler that is included in RHEL 7 is v4.8.5).
Various atomics support removed: S390/s390x, Sparc v9, ARMv4 and ARMv5 with CMA support.
The MPI C++ bindings have been removed.
--amcaoptions have been deprecated.
libompitracecontributed library has been removed. This library was incomplete and unmaintained. If needed, it is available in the v4.x series.
The rankfile format no longer supports physical processor locations. Only logical processor locations are supported.
32-bit builds have been disabled. Building Open MPI in a 32-bit environment is no longer supported. 32 bit support is still available in the v4.x series.
Hardware Locality updates:
Open MPI now requires Hardware Locality v1.11.0 or later.
The internally-bundled Hardware Locality shipped with Open MPI has been updated to v2.7.1.
Open MPI builds Hardware Locality with
Documentation updates and improvements:
Many, many people from the Open MPI community contributed to the overall documentation effort — not only those who are listed in the Git commit logs — including (but not limited to):
Build updates and fixes:
Various changes and cleanup to fix, and better support the static building of Open MPI.
Change the default component build behavior to prefer building components as part of the core Open MPI library instead of individual DSOs. Currently, this means the Open SHMEM layer will only build if the UCX library is found.
autogen.plnow supports a
-joption to run multi-threaded. Users can also use the environment variable
autogen.plto support macOS Big Sur. Thanks to @fxcoudert for reporting the issue.
Fixed bug where
autogen.plwould not ignore all excluded components when using the
Fixed a bug the
buildrpm.shwhich would result in an rpm build failure. Thanks to John K. McIver III for reporting and fixing.
C++compiler requirement to build Open MPI.
Updates to improve the handling of the compiler version string in the build system. This fixes a compiler error with clang and armclang.
Added OpenPMIx binaries to the build, including
pmix_info. Thanks to Mamzi Bayatpour for their contribution to this effort.
Open MPI now links to Libevent using
Added support for setting the wrapper C compiler. This adds a new option:
Fixed compilation errors when running on IME file systems due to a missing header inclusion. Thanks to Sylvain Didelot for finding and fixing this issue.
Add support for GNU Autoconf v2.7.x.
Other updates and bug fixes:
Updated Open MPI to use
common/ompio: implement pipelined read and write operation. This new new code path shows significant performance improvements for reading/writing device buffers compared to the previous implementation, and reduces the memory footprint of Open MPI IO (“OMPIO”) by allocating smaller temporary buffers.
Fixed Fortran-8-byte-INTEGER vs. C-4-byte-int issue in the
mpi_f08MPI Fortran bindings module. Thanks to @ahaichen for reporting the bug.
Fixed Fortran keyword issue when compiling
oshmem_info. Thanks to Pak Lui for finding and fixing the bug.
Added check for Fortran
ISO_FORTRAN_ENV:REAL16. Thanks to Jeff Hammond for reporting this issue.
Fixed Fortran preprocessor issue with
CPPFLAGS. Thanks to Jeff Hammond for reporting this issue.
MPI module: added the
TYPE(MPI_*)types for Fortran. Thanks to George Katevenis for the report and their contribution to the patch.
Fixed a typo in an error string when showing the stack frame. Thanks to Naribayashi Akira for finding and fixing the bug.
Fixed output error strings and some comments in the Open MPI code base. Thanks to Julien Emmanuel for tirelessly finding and fixing these issues.
uctBTL transport now supports
UCXv1.9 and higher. There is no longer a maximum supported version.
Updated the UCT BTL defaults to allow NVIDIA/Mellanox HCAs (
mlx5_0) for compatibility with the one-sided
Fixed a crash during CUDA initialization. Thanks to Yaz Saito for finding and fixing the bug.
MPI_Comm_spawn()support has been fixed.
PowerPC atomics: Force usage of ppc assembly by default.
The default atomics have been changed to be GCC, with C11 as a fallback. C11 atomics incurs sequential memory ordering, which in most cases is not desired.
Various datatype bugfixes and performance improvements.
Various pack/unpack bugfixes and performance improvements.
Various OSHMEM bugfixes and performance improvements.
New algorithm for Allgather and Allgatherv has been added, based on the paper “Sparbit: a new logarithmic-cost and data locality-aware MPI Allgather algorithm”. Default algorithm selection rules are unchanged; to use these algorithms add:
--mca coll_tuned_allgather_algorithm sparbitand/or
--mca coll_tuned_allgatherv_algorithm sparbitto your
mpiruncommand. Thanks to Wilton Jaciel Loch and Guilherme Koslovski for their contribution.
Updated the usage of
.gitmodulesto use relative paths from absolute paths. This allows the submodule cloning to use the same protocol as Open MPI cloning. Thanks to Felix Uhl for the contribution.