[mvapich-discuss] Announcing the Release of MVAPICH2-X 2.3 GA and OSU INAM 0.9.6

[Panda, Dhabaleswar]

The MVAPICH team is pleased to announce the release of MVAPICH2-X 2.3
GA (with support for advanced MPI Features, OSU INAM, and Hybrid
MPI+PGAS) and OSU InfiniBand Network Analysis and Monitoring (INAM)
Tool 0.9.6.

MVAPICH2-X 2.3 provides support for advanced MPI features/support
(XPMEM, enhanced asynchronous progress design, User Mode Memory
Registration (UMR), Dynamically Connected (DC) transport protocol,
Scalable Reliable Datagram (SRD) transport protocol, SHARP and Data
Partitioning-based Multi-Leader Design (DPML) for MPI collectives,
XPMEM-based collectives, Contention Aware Kernel-Assisted MPI
collectives, and Non-blocking Collectives with Core-Direct and SHARP),
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 (since 2.3rc3):

    - MPI Features
        - Based on MVAPICH2 2.3.4
            - OFA-IB-CH3, OFA-IB-RoCE, PSM-CH3, and PSM2-CH3 interfaces
        - Enhanced point-to-point and collective tuning for AMD EPYC,
          Catalyst at EPCC, Mayer at Sandia, Auzre at Microsoft, AWS, and Frontera at TACC

    - MPI (Advanced) Features
        - Optimized support for large message MPI_Allreduce and MPI_Reduce
            - OFA-IB-CH3 and OFA-IB-RoCE interfaces
        - Improved performance for communication using DC transport
            -  OFA-IB-CH3 interface
        - Enhanced support for AWS EFA adapter and SRD transport protocol
            -  OFA-IB-CH3 interface
        - Enhanced point-to-point and collective tuning for AWS EFA adapter
          and SRD transport protocol
            -  OFA-IB-CH3 interface
        - Add multiple MPI_T PVARs and CVARs for point-to-point and collective
          operations
        - Tuning for MPI collective operations for Intel Broadwell, Intel
          CascadeLake, Azure HB (AMD EPYC), and Azure HC (Intel Skylake) systems
            - OFA-IB-CH3, OFA-IB-RoCE, PSM-CH3, and PSM2-CH3 interfaces
    - Support for OSU InfiniBand Network Analysis and Management (OSU INAM)
      Tool v0.9.6

    - Unified Runtime Features
        - Based on MVAPICH2 2.3.4 (OFA-IB-CH3 interface). All the runtime
          features enabled by default in OFA-IB-CH3 and OFA-IB-RoCE interfaces
          of MVAPICH2 2.3.4 are available in MVAPICH2-X 2.3

 * Bug Fixes (since 2.3rc3):
    - Fix issues in UD-Hybrid code path
    - Fix various compilation warnings and memory leaks

OSU INAM monitors InfiniBand clusters in real time by querying various
subnet management entities in the network. It is also capable of
interacting with the MVAPICH2-X software stack to gain insights into
the communication pattern of the application and classify the data
transferred into Point-to-Point, Collective, Remote Memory Access
(RMA), and Lustre I/O. OSU INAM can also remotely monitor several
parameters of MPI processes in conjunction with MVAPICH2-X.

 * Major Features & Enhancements (since 0.9.5):
   - Support to collect and visualize MPI_T based performance data
       - Ability to collect and display "most used" MPI primitives
           - Node, job, and cluster level granularities
           - Live and historical views
       - Ability to collect and display MPI_T based performance data for
         each MPI primitive for different message ranges
           - Node, job, and cluster level granularities
           - Live and historical views
       - Ability to classify blocking and non-blocking data transfers for
         different message ranges
           - Node, job, and cluster level granularities
           - Live and historical views
   - Ability to gather and display Lustre I/O for MPI jobs
       - Support node, job, and cluster level granularities
       - Support live and historical views
   - Enable emulation mode to allow users to test OSU INAM tool in a sandbox
     environment without actual deployment. Emulation mode supports:
       - Node, job, and cluster level granularities
       - Live and historical views
       - Collecting and reporting internal details of MPI jobs
   - Support to search for nodes and jobs for historical node/job and network
     view page
   - Generate email notifications to alert users when user defined events occur
   - Ability to select PBS/SLURM job schedulers at runtime
   - Support for MOFED 4.5, 4.6, 4.7, and 5.0
   - Support for ARM and OpenPower architecture
   - Support for HDR InfiniBand adapters and switches
   - Showing interval of querying data on the web front charts
   - Improving functionality and stability of OSU INAM daemon
   - Redesigned and optimized web front charts with clear legends

 * Bug Fixes (since 0.9.5):
   - Fixed issue for live node where HCA query is enabled
   - Handled missing slashes for phantomJS configuration parameters
   - Fix memory leaks

For downloading MVAPICH2-X 2.3, OSU INAM v0.9.6 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

PS: We are also happy to inform that the number of organizations using
MVAPICH2 libraries (and registered at the MVAPICH site) has crossed
3,100 worldwide (in 89 countries). The number of downloads from the
MVAPICH site has crossed 764,000 (0.76 million).  The MVAPICH team
would like to thank all its users and organizations!!