The ASCI Red TOPS Supercomputer

Introduction
The ASCI Red TOPS (Tera-OPerations per Second) Supercomputer is the first step in the ASCI Platforms Strategy, which is aimed at giving researchers the five-order-of-magnitude increase in computing performance over current technology that is required to support "full-physics," "full-system" simulation by early next century. This supercomputer, installed at Sandia National Laboratories, is a massively parallel, MIMD (Multiple Instruction, Multiple Data) computer. It is noteworthy for several reasons. It was the world's first TOPS supercomputer. I/O, memory, compute nodes, and communication are scalable to an extreme degree. Standard parallel interfaces make it relatively simple to port parallel applications to this system. The system uses two operating systems to make the computer both familiar to the user (UNIX) and non-intrusive for the scalable application (Cougar). And it makes use of Commercial Commodity Off The Shelf (CCOTS) technology to maintain affordability.

Hardware
The ASCI TOPS system is a distributed memory, MIMD, message-passing supercomputer. All aspects of this system architecture are scalable, including communication bandwidth, main memory, internal disk storage capacity, and I/O.

Artist's Concept

Software
Software on the TOPS Supercomputer is a combination of operating systems tailored for specific tasks and standard programming tools to make the computer both familiar to the user and non-intrusive for the scalable application. To the application programmer, the system looks like a UNIX-based supercomputer. All the standard facilities associated with a UNIX workstation will be available to the user.
The operating system used for the Service, I/O, and System Partitions is Intel's distributed version of UNIX (POSIX 1003.1 and XPG3, AT&T System V.3 and 4.3 BSD Reno VFS) developed for the Paragon XP/S Supercomputer. The Paragon OS presents a single system image to the user. This means that users see the system as a single UNIX machine despite the fact that the operating system is running on a distributed collection of nodes.
The operating system in the Compute Partition is Cougar. Cougar is Intel's port of Puma, a light-weight operating system for the TOPS, based on the very successful SUNMOS system for the Paragon. (SUNMOS, and subsequently Puma, were developed by Sandia National Laboratories and the University of New Mexico.) System services and support for the interactive user are provided by a host OS (in this case, the Paragon OS running in the Service Partition). All access to hardware resources comes from the Q-Kernel, the lowest-level component of Cougar. Above the Q-Kernel sits the process control thread (PCT), which runs in user space and manages processes. At the highest level is the user's applications. As with most MPP systems, the basic programming model in Cougar is based on message passing.
FORTRAN77, FORTRAN90, C and C++ are supported. The interactive debugger and performance analysis tools understand these languages and map onto original source code.

Conclusion
The ASCI platform effort bridges the gap between giga-scale and tera-scale computing to accommodate the five-order-of-magnitude increase in performance required by "full-physics", "full-system" simulation.