Scali, the leader in higher performance computing software, has released the first benchmark data using AMD’s newly released Quad-Core Opteron model 2350 processor with the Mellanox ConnectX interconnect.The results highlight the increased importance of high performance MPI and InfiniBand interconnects as the core count increases in current and future systems.
Based on the LS-Dyna neon_refined_revised TopCrunch benchmark, Scali and Mellanox were able to compare performance of multi-core platforms running the same application. The Scali™ MPI Connect™ based-version, running in conjunction with Mellanox 20Gb/s ConnectX interconnect, delivered superior performance, in excess of 30%, on AMD’s Quad-Core Opteron when compared to similar benchmarks published on www.topcrunch.org. These performance gains allow customers to complete more jobs per day and increase utilization of high performance computing clusters.
Scali MPI Connect™: Scali MPI Connect consistently outperforms other solutions at both application run time and application scaling, enabling you to maximize the performance of your applications within your environment. Focused on accelerating and scaling your applications, Scali MPI Connect is a fully integrated message passing interface (MPI) solution, enabling your applications to run at maximum performance through its unique, high performance MPI implementation. Additionally, Scali MPI Connect allows you to take advantage of leading interconnect technologies to build high performance clusters. MPI Connect lets application developers support a wider range of customer environments while lowering the number of binaries required. Scali MPI Connect delivers faster performance through application enhancement.
Why Scali MPI Connect?
Performance, performance, performance. Scali MPI Connect consistently outperforms other MPI implementations in benchmark after benchmark. Scali runs faster. If you find another MPI implementation that beats Scali (on equal terms, of course),
Scalability Scali MPI Connect consistently produces application scalability to higher node counts than other MPI implementations.
Broad interconnect support Scali MPI Connect supports a full range of interconnects ranging from legacy Ethernet to high-performance interconnects such as InfiniBand, InfiniPath, and Myrinet
Ease of development Scali MPI Connect enables you to use a single MPI implementation, and a single executable, to support a wide array of interconnects.
Key Features
Multi-core awareness Scali MPI Connect has specific optimizations and tuning parameters to maximize performance for multi-core environments. This allows applications to leverage the right balance of processors, memory availability, and network capacity to achieve optimal performance.
Comprehensive ebugging and troubleshooting tools Scali MPI Connect has an extensive set of tools including verification and test tools, application tracing and timing facilities, and performance counters with message sizes and counts. These tools enable application developers and users to maximize application performance.
Tracing and monitoring MPI-related monitoring presentation can be selected through environment variables to determine presentation of timing and trace information, with no recompilation or re-linking of the application. There is no need for re-compilation or re-linking the application.
Multithread safe Multithreaded applications can fully exploit Scali MPI Connect and multiple threads can simultaneously request services and conduct communication.
Automatic selection of optimal network Through the use of a priority network list built from system configuration files, user environment variables, user command line options and library hard-coded defaults, Scali MPI Connect can dynamically select the optimal network connection between each node and each other node within a cluster at runtime. This maximizes network efficiency so that processes get the data they need quickly.
Network failover capability Scali MPI Connect has the ability to shift from a high performance network (RDMA-type) network to a standard TCP/IP network without causing disruption to a running job, ensuring that jobs run to completion regardless of network instabilities.
Simplified application development Scali MPI Connect enables you to use a single MPI implementation, and a single executable, to support a wide array of interconnects, with dynamic runtime selection of interconnect. With Scali MPI Connect, you can run a single application across multiple architectures, Linux distributions and interconnects. It eliminates need to recompile applications and MPI libraries, while providing full support for dynamically linked libraries. These means there is no need to recompile for every compiler/interconnect combination.
Linux command line replication Command line arguments to the application are automatically provided to all MPI processes, avoiding tedious parsing and broadcasting of parameters to other MPI processes.
MIMD support The Multiple Instruction - Multiple Data (MIMD) model is supported through provisions that launch different executables which constitute the whole MPI application.
Support for popular debuggers Scali MPI Connect fully supports Scali MPI Connect fully supports Etnus TotalView analysis tools, Allineaís distributed debugging tool (DDT), and standard GNU gdb.
Heterogeneous cluster support Application processes are distributed at runtime across nodes, leveraging a range of different hardware architectures, such as x86, x86-64, or EM64T.
