SGI Origin 300

Shared-Memory Architecture

The SGI Origin 300 server is the latest advancement in the revolutionary SGI® NUMAflex^TM approach to modular computing. With its small footprint and highly modular design, it is ideal for computational throughput, media streaming, or complex data management. Our SGI Origin 300 server module packs the power of  four MIPS® processors, the high-performance SGI® IRIX® operating system, and one GB of memory into an ultradense, 3.5-inch, rack-mountable chassis. For more computationally intensive applications, we will add a second SGI Origin 300 server and double the capacity to eight MIPS processors and two GB of memory in a single, shared-memory system.

SGI NUMAflex is the third-generation shared-memory system architecture.

NUMAflex can deliver scalable, low-latency memory access and efficient resource management, and is an ideal architectural solution for high-productivity computing. The NUMAflex architecture incorporates a low-latency, high-bandwidth interconnect that is designed to maintain performance as it scales, allowing efficient access to local and remote memory without the bottlenecks associated with switches, backplanes, and other commodity interconnect technologies. In the NUMAflex architecture, the entire memory space is shared. Hence, large models fit into memory with no programming model restrictions. Rather than waiting for all of the processors to complete their assigned tasks, the system dynamically reassigns resources to more complex areas, resulting in faster time to solution.

IRIX Operating System 6.5

The IRIX® operating system is the leading technical high-performance 64-bit operating system based on industry-standard UNIX®. With IRIX, customers can take full advantage of MIPS® processor-based SGI® systems. IRIX 6.5 is SGI's fifth generation of IRIX and is one of the most important and mature UNIX operating system releases in the industry.

Compilers of SGI Origin 300

The MIPSpro compilers represent the latest technology of optimizing and parallelizing compilers from SGI. They include:

• support for C, C++ and Fortran 77/90
• state-of-the-art optimizations and automatic parallelizing technology
• enhancements fo interprocedural analysis (IPA)
• complete support for 32- and 64-bit development and execution

The Automatic Parallelization Option (APO) for C, C++, and Fortran 77/90 enables a variety of applications to benefit from multiprocessor execution. APO can analyze and restructure user applications for parallel execution. User-assisted parallelization is also supported with standards-based directives, enabling users to actively assist the compiler in the parallelization process.

The MIPSpro family of compilers is completely interoperable. Program modules written in Fortran 77, Fortran 90, C and C++ can be linked to form one application for single or multiple processor execution.

In addition, SGI also supports the Java and Ada95 programming languages.

Advanced programming tools and libraries

• ProDev^TM WorkShop provides the specific tools required for debugging large, scalable application code and supports multiprocessor programming interfaces including MPI, OpenMP, and POSIX threads.
• Message Passing Toolkit (MPT) optimizes the performance of MPI codes running over the SGI® NUMAflex^TM architecture.
• SCSL scientific libraries provide the specific library calls needed for technical and scientific applications.
• SGI® SpeedShop^TM helps to identify and eliminate programming bottlenecks.
• Performance Co-Pilot^TM* tracks performance at the system resource level to help identify potential areas for efficiency optimization.