From panda at cse.ohio-state.edu Wed Mar 15 00:40:48 2006 From: panda at cse.ohio-state.edu (Dhabaleswar Panda) Date: Wed Mar 15 00:40:53 2006 Subject: [mvapich] Announcing the release of MVAPICH 0.9.7 with support for Gen2, SRQ with Flow Control, Fault Tolerance and anonymous SVN access Message-ID: <200603150540.k2F5emA8021089@xi.cse.ohio-state.edu> You are receiving this e-mail because of: 1) being a member of more than 330 organizations world-wide who have downloaded MVAPICH/MVAPICH2 from the OSU web site and/or 2) your interest in the MVAPICH/MVAPICH2 package The MVAPICH team is pleased to announce the release of MVAPICH 0.9.7 with the following new features: - Shared Receive Queue (SRQ) support with flow control: The new design uses significantly less memory for MPI library - less than 300MBytes of MPI internal buffers per process for clusters with 16K processes!! Performance benefits of the proposed scheme for sample applications (HPL and NAS LU) can be obtained from MVAPICH project's page -> Performance. - OpenIB/Gen2 support: All features (such as RDMA-based collectives, multi-rail, scalable MPD-based startup, etc.) available with VAPI are now available with OpenIB/Gen2. - MVAPICH-Gen2 also enables the use of IBM ehca adapter and PathScale adapter through OpenIB/Gen2 support - Support for Fault Tolerance: Mem-to-mem reliable data transfer (detection of I/O bus error with 32bit CRC). Additional fault tolerance support (such as checkpoint restart, automatic path migration (APM), etc.) will be introduced in successive releases. - Advanced AVL tree-based Resource-aware registration cache - Tuning and Optimization of various collective algorithms for a wide range of system sizes - Multi-rail communication: Various schemes (Multiple queue pairs per port, Multiple ports per adapter, and Multiple adapters). Flexible scheduling policies: round-robin for small and non-blocking communication and striping for large blocking communication MVAPICH 0.9.7 release supports Gen2, VAPI and uDAPL transport interfaces. It also has support for the standard TCP/IP (provided by MPICH stack). It is tested with the following architecture, OS, compilers and InfiniBand adapters: - Architecture: EM64T, Opteron, IA-32, IBM PPC and Mac G5 - Operating Systems: Linux, Solaris, AIX and Mac OSX - Compilers: gcc, intel, pathscale and pgi - InfiniBand Adapters: - Mellanox adapters with PCI-X and PCI-Express (SDR and DDR with mem-full and mem-free cards) - PathScale adapter (through OpenIB/Gen2 support) - IBM ehca adapter (through OpenIB/Gen2 support) More details on all features and supported platforms can be obtained by visiting the project's web page -> Overview -> features. MVAPICH 0.9.7 is being distributed as a single integrated package (with MPICH 1.2.7 and MVICH). It is available under BSD license. Starting with this 0.9.7 release, MVAPICH team is also pleased to announce the availability of the code base through anonymous SVN access. Nightly tarballs are also available. A new mvapich-commit mailing list has been established for users, developers and vendors to keep track of all commits happening to the SVN. (SVN access to MVAPICH2 code base will be coming soon.) MVAPICH 0.9.7 continues to deliver excellent performance. Performance numbers for all platforms and system configurations and operations can be viewed by visiting `Performance' section of the project's web page. An enhanced and detailed `User Guide' is now available (in both html and pdf forms) from the FAQ page. For downloading MVAPICH 0.9.7 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 OpenIB SVN. All feedbacks, including bug reports, hints for performance tuning, patches and enhancements are welcome. Please post it to the newly established mvapich-discuss mailing list. Thanks, MVAPICH Team at OSU/NBCL ---------- PS: If you would like to be removed from this mailing list, please end an e-mail to mvapich_request@cse.ohio-state.edu. ====================================================================== 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 AMD, Apple, Appro, IBM, Intel, Mellanox, Microway, PathScale, SilverStorm and Sun Microsystems. Other technology partner includes Etnus. ======================================================================