[openfabrics-ewg] Announcing the release of MVAPICH2 0.9.8 with Checkpoint/Restart, iWARP, RDMA CM-based connection management and optimized collective support

[Dhabaleswar Panda]

The MVAPICH team is pleased to announce the availability of MVAPICH2
0.9.8 with the following NEW features:

 - Checkpoint/Restart support for application transparent systems-level
   fault tolerance. BLCR-based support using native InfiniBand Gen2
   interface is provided. Flexible interface to work with different
   file systems. Tested with ext3 (local disk), NFS and PVFS2.  

   Performance of sample applications with checkpoint-restart using 
   PVFS2 and Lustre can be found here:

   http://nowlab.cse.ohio-state.edu/projects/mpi-iba/performance/mvapich2/application/MVAPICH2-ckpt.html

 - iWARP support: Incorporates the support for OpenFabrics/Gen2-iWARP.
   Tested with Chelsio T3 (10GigE) and Ammasso iWARP adapters and
   drivers.

 - RDMA CM-based Connection management support

 - Shared memory optimizations for collective communication operations. 
   Efficient algorithms and optimizations for barrier, reduce and
   all-reduce operations. Exploits the multi-core optimized shared
   memory point-to-point communication support introduced in MVAPICH2
   0.9.6.
 
   Performance of sample collective operations with this new feature
   can be found here:

   http://nowlab.cse.ohio-state.edu/projects/mpi-iba/perf-coll.html 

 - uDAPL support for NetEffect 10GigE adapter. Tested with  
   NetEffect NE010 adapter.

More details on all features and supported platforms can be obtained
by visiting the following URL:

http://nowlab.cse.ohio-state.edu/projects/mpi-iba/mvapich2_features.html

MVAPICH2 0.9.8 release is tested with the latest OFED 1.1 stack.  It
continues to deliver excellent performance.  Sample performance
numbers include:
 
  - OpenFabrics/Gen2 on EM64T dual-core with PCI-Ex and IBA-DDR:
      Two-sided operations: 
        - 2.81 microsec one-way latency (4 bytes)
        - 1561 MB/sec unidirectional bandwidth 
        - 2935 MB/sec bidirectional bandwidth  

      One-sided operations:
        - 4.92 microsec Put latency 
        - 1569 MB/sec unidirectional Put bandwidth 
        - 2935 MB/sec bidirectional Put bandwidth 

  - OpenFabrics/Gen2 on EM64T dual-core with PCI-Ex and IBA-DDR (Dual-rail):
      Two-sided operations:
        - 2.81 microsec one-way latency (4 bytes)
        - 3127 MB/sec unidirectional bandwidth
        - 5917 MB/sec bidirectional bandwidth

      One-sided operations:
        - 4.37 microsec Put latency
        - 3137 MB/sec unidirectional Put bandwidth
        - 5917 MB/sec bidirectional Put bandwidth

  - OpenFabrics/Gen2 on Opteron single-core with PCI-Ex and IBA-DDR: 
      Two-sided operations:
        - 3.01 microsec one-way latency (4 bytes) 
        - 1402 MB/sec unidirectional bandwidth
        - 2238 MB/sec bidirectional bandwidth 

      One-sided operations:
        - 4.65 microsec Put latency
        - 1402 MB/sec unidirectional Put bandwidth
        - 2238 MB/sec bidirectional Put bandwidth

Performance numbers for all other platforms, system configurations and
operations can be viewed by visiting `Performance' section of the
project's web page.

With the ADI-3-level design, MVAPICH2 0.9.8 delivers similar
performance for two-sided operations compared to MVAPICH 0.9.8.
Organizations and users interested in getting the best performance for
both two-sided and one-sided operations and also want to exploit
advanced features (such as fault tolerance with checkpoint/restart,
iWARP, RDMA CM connection management, multi-threading, integrated
multi-rail, multi-core optimization, memory hook support and optimized
collectives) may migrate from MVAPICH code base to MVAPICH2 code base.

For downloading MVAPICH2 0.9.8 package and accessing the anonymous
SVN, please visit the following URL:

http://nowlab.cse.ohio-state.edu/projects/mpi-iba/

A stripped down version of this release is also available at the
OpenFabrics SVN.

All feedbacks, including bug reports and hints for performance tuning,
are welcome. Please post it to the mvapich-discuss mailing list.

Thanks, 

MVAPICH Team at OSU/NBCL 

======================================================================
MVAPICH/MVAPICH2 project is currently supported with funding from
U.S. National Science Foundation, U.S. DOE Office of Science,
Mellanox, Intel, Cisco Systems, Sun Microsystems and Linux Networx;
and with equipment support from Advanced Clustering, AMD, Apple,
Appro, Dell, IBM, Intel, Mellanox, Microway, PathScale, SilverStorm
and Sun Microsystems. Other technology partner includes Etnus.
======================================================================