[mvapich-discuss] Announcing the release of MVAPICH2-X 2.2 GA and OMB 5.3.2
Panda, Dhabaleswar
panda at cse.ohio-state.edu
Fri Sep 9 19:22:44 EDT 2016
The MVAPICH team is pleased to announce the release of MVAPICH2-X 2.2
GA (Advanced MPI Features, Support for OSU INAM and Hybrid MPI+PGAS
(OpenSHMEM, UPC, CAF and UPC++) with Unified Communication Runtime),
and OSU Micro-Benchmarks (OMB) 5.3.2.
MVAPICH2-X 2.2 GA provides support for advanced MPI features
(Dynamically Connected (DC) transport protocol, User Mode Memory
Registration, On-Demand Paging (ODP) and Non-blocking Collectives with
Core-Direct), OSU INAM and hybrid MPI+PGAS (UPC, OpenSHMEM, CAF, and
UPC++) programming models with unified communication runtime for
emerging exascale systems. This library also provides flexibility for
users to write applications using the following programming models
with a unified communication runtime: MPI, MPI+OpenMP, pure UPC, pure
OpenSHMEM, pure UPC++, and pure CAF programs as well as hybrid
MPI(+OpenMP) + PGAS (UPC, OpenSHMEM, CAF, and UPC++) programs.
Features and enhancements for MVAPICH2-X 2.2 GA are as follows:
* Features and Enhancements (since 2.1):
- MPI Features
- Based on MVAPICH2 2.2 GA (OFA-IB-CH3 interface)
- Support for Dynamically Connected (DC) transport protocol
- Support for pt-to-pt, RMA and collectives
- Efficient support for On Demand Paging (ODP) feature of Mellanox
- Support for point-to-point and RMA operations
- Support User Mode Memory Registration (UMR) for high performance
non-contiguous data transfer
- Support for Core-Direct based Non-blocking collectives
- Support for Hybrid mode with RC/DC/UD/XRC
- Support for Intel Knights Landing architecture
- Support for OpenPower architecture
- Support for Intel Omni-Path architecture
- Support for RoCE v2
- OpenSHMEM Features
- Support for RoCE and RoCE v2
- Support for Dynamically Connected (DC) transport protocol
- Support for Intel Knights Landing architecture
- Support for OpenPower architecture
- UPC Features
- Based on GASNET V1.26 and Berkeley UPC 2.20.2 (contains
changes/additions in preparation for upcoming
UPC 1.3 specification)
- Support for RoCE and RoCE v2
- Support for Dynamically Connected (DC) transport protocol
- Support for Intel Knights Landing architecture
- Support for OpenPower architecture
- CAF Features
- Support for RoCE and RoCE v2
- Support for Dynamically Connected (DC) transport protocol
- Support for Intel Knights Landing architecture
- Introducing UPC++ Support
- Based on Berkeley UPC++ v0.1
- Introduce UPC++ level support for new scatter collective
operation (upcxx_scatter)
- Optimized UPC collectives (improved performance for
upcxx_reduce, upcxx_bcast, upcxx_gather, upcxx_allgather,
upcxx_alltoall)
- Unified Runtime Features
- Based on MVAPICH2 2.2 (OFA-IB-CH3 interface). All the runtime
features enabled by default in OFA-IB-CH3 interface of
MVAPICH2 2.2 are available in MVAPICH2-X 2.2
- The advanced MPI features listed above are available with the
unified runtime
- Introduce support for UPC++ and MPI+UPC++ programming models
- Support for OSU InfiniBand Network Analysis and Management (OSU INAM)
Tool v0.9.1
- Capability to profile and report node-level, job-level and
process-level activities for MPI communication (Point-to-Point,
Collectives and RMA) at user specified granularity in conjunction
with OSU INAM
- Capability to profile and report process to node communication matrix
for MPI processes at user specified granularity in conjunction with
OSU INAM
- Capability to profile and report CPU utilization of MPI processes
at user specified granularity in conjunction with OSU INAM
- Memory Utilization
- Inter-node communication buffer usage for RC transport
- Inter-node communication buffer usage for UD transport
Bug Fixes
- Fix corner case in RDMA_CM based startup when UCR support is enabled
- Fix compilation warnings and memory leaks
New features, enhancements and bug fixes for OSU Micro-Benchmarks
(OMB) 5.3.2 are listed here.
* New Features & Enhancements (since 5.3.1)
- Allow specifying very large message sizes (>2GB) for
collective benchmarks
* Bug Fixes
- Fix compilation errors due to missing type casting
For downloading MVAPICH2-X 2.2 GA, OMB 5.3.2 and 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
More information about the mvapich-discuss
mailing list