[mvapich-discuss] Announcing the release of MVAPICH2 2.0.1, MVAPICH2-X 2.0.1, and OMB 4.4.1
Panda, Dhabaleswar
panda at cse.ohio-state.edu
Thu Oct 30 23:50:44 EDT 2014
The MVAPICH team is pleased to announce the release of MVAPICH2 2.0.1,
MVAPICH2-X 2.0.1 (Hybrid MPI+PGAS (OpenSHMEM) with Unified
Communication Runtime) and OSU Micro-Benchmarks (OMB) 4.4.1.
These are primarily bug-fix releases. MVAPICH2 2.0, MVAPICH2-X 2.0, and
OMB 4.4 users are requested to upgrade to these latest versions.
Features, Enhancements, and Bug Fixes for MVAPICH2 2.0.1 are listed
here.
* Features and Enhancements (since 2.0 GA):
- Support for PMI-2 based startup with SLURM
- Enhanced startup performance for Gen2/UD-Hybrid channel
- Add -export-all option to mpirun_rsh
- GPU support for MPI_Scan and MPI_Exscan collective operations
- Optimize creation of 2-level communicator
- Collective optimization for PSM-CH3 channel
- Increase counter length for shared-memory collectives
- Tuning for IvyBridge architecture
- Support for additional MPI-T performance variables (PVARs)
in the CH3 channel
- Standardizing the MPI-T PVAR names
- Link with libstdc++ when building with GPU support
(required by CUDA 6.5)
- Increase number of connect attempts for RDMA_CM
* Bug-Fixes (since 2.0 GA):
- Fix issues in MPI_Unpack operation with CUDA
- Fix issues with DPM and PMI-1 APIs
- Fix issue with DPM and non-blocking collectives
- Fix issue with detecting the use of mpirun_rsh in SMP jobs
- Add check for pending operations in one-sided channel in flush_all
- Thanks to Hajime Fujita for the report
- Fixing threads deadlock scenario with collectives
- Fix error in large message (>2GB) transfers in CMA code path
- Fix issues with optimizations for broadcast and reduce collectives
- Fix hang at finalize with Gen2-Hybrid/UD channel
- Fix issues for collectives with non power-of-two process counts
- Thanks to Evren Yurtesen for identifying the issue
- Make ring startup use HCA selected by user
- Fix corner error case with RDMA_CM
- Fix several compile time warnings
- Fix memory leaks in OFA-IB-CH3 and OFA-IB-Nemesis channels
MVAPICH2-X 2.0.1 software package provides support for hybrid MPI+PGAS
(UPC and OpenSHMEM) programming models with unified communication
runtime for emerging exascale systems. This software package provides
flexibility for users to write applications using the following
programming models with a unified communication runtime: MPI,
MPI+OpenMP, pure UPC, and pure OpenSHMEM programs as well as hybrid
MPI(+OpenMP) + PGAS (UPC and OpenSHMEM) programs.
Features and enhancements for MVAPICH2-X 2.0.1 are as follows:
* Features and Enhancements (since 2.0 GA):
- MPI Features
- Based on MVAPICH2 2.0.1 (OFA-IB-CH3 interface)
- Unified Runtime Features
- Based on MVAPICH2 2.0.1 (OFA-IB-CH3 interface). All the runtime
features enabled by default in OFA-IB-CH3 interface of
MVAPICH2 2.0.1 are available in MVAPICH2-X 2.0.1
Bug Fixes for OSU Micro-Benchmarks (OMB) 4.4.1 are
listed here.
* Bug Fixes (since OMB 4.4)
- adding missing MPI3 guard for WIN_ALLOCATE
- capture getopt return value in an int instead of char
For downloading MVAPICH2 2.0.1, MVAPICH2-X 2.0.1, OMB 4.4.1,
associated user guides, quick start guide, and accessing the SVN,
please visit the following URL:
http://mvapich.cse.ohio-state.edu
All questions, feedback, bug reports, hints for performance tuning,
patches and enhancements are welcome. Please post it to the
mvapich-discuss mailing list (mvapich-discuss at cse.ohio-state.edu).
Thanks,
The MVAPICH Team
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cse.ohio-state.edu/pipermail/mvapich-discuss/attachments/20141031/4cb3c9a9/attachment.html>
More information about the mvapich-discuss
mailing list