High Performance Computing Systems (HPCS)

The HPCS component produces scalable parallel computing systems in collaboration with industry and academia. Unlike dedicated, single processor architectures of the past, scalable parallel systems have the property that increases in size result in proportional improvement in performance. This is achieved by connecting multiple processors and memory units through a scalable interconnection structure. Scalable systems can be configured over a wide range that can deliver high performance computing to users at both small and very large scales.

Because the computing system designs are scalable, they can be used in smaller scale workstations. Such workstations may also have high performance graphics capabilities to enable visualization of a computational result and provide interactive interfaces to the user. These workstations may be linked to local networks connected to the Internet, a network of networks that includes high performance subnets linking higher performance and larger scale computing systems throughout the country.

HPCS focuses on the fundamental scientific and technological challenges of accelerating the advance of affordable scalable parallel high performance computing systems. Critical underlying technologies are developed in prototype form along with associated design tools. This allows evaluation of alternatives as the prototype systems mature. Evaluation continues throughout the research and development process, with experimental results used to refine successive generations of systems.

Scalable computing technologies used in combination with scalable networking technologies provide the technology base needed to address the Grand Challenges and the National Challenges. The necessary software technologies are developed by the ASTA and IITA components.

HPCS is composed of four elements to produce progressively more advanced and mature systems:

I. Research for Future Generations of Computing Systems

This element develops the underlying architecture, components, packaging (integration of electronics, photonics, power, cooling, and other components), systems software, and scaling concepts to achieve affordable high performance computing systems. These efforts ensure that the required advanced technologies will be available for the new systems and provide a foundation for the more powerful systems to follow. This element also produces the basic approaches for systems software, programming languages, and environments for heterogeneous configurations of workstations and high performance servers.

II. System Design Tools

This element develops computer aided design tools and the technology to allow multiple design tools to work together in order to enable the design, analysis, simulation, and testing of system components and modules. These tools make rapid prototyping of new system concepts possible. New design tools will be produced to enable the design of more advanced prototype systems using new technologies as they emerge.

III. Advanced Prototype Systems

Systems capable of scaling to 100 gigaops (billions of operations per second) performance have begun to emerge. Teraops (trillions of operations per second) performance designs will be demonstrated by the mid 1990s. Research in high performance systems focuses on reducing the cost and size of these systems so they can be used for a broader range of applications.

IV. Evaluation of Early Systems

Experimental systems will be placed at sites where researchers can provide feedback to systems and software designers. Performance evaluation criteria for systems and results of evaluations will be made widely available. Scalability enables small to medium size systems to be used for early performance evaluation and software development in preparation for larger scale applications. Larger scale systems are included in the ASTA component for applications such as the Grand Challenges.

HPCS Accomplishments

• Small, medium, and large scale systems developed under the HPCS component have been deployed and are being used in the ASTA component. This includes large systems deployed in various high performance computing centers and some systems installed in heterogeneous configurations.

  The small and medium scale systems are being used to develop algorithms and software, including fundamental building blocks for Grand Challenge problems and a wide variety of new scientific computation models. These prototypes are characterized by very fast routing and component technology, capable of scaling up to 100 gigaops system configurations.

• Scalable systems continue to be evaluated and refined, providing early feedback on hardware, operating systems, compilers, software development tools, input/output systems, and mass storage systems. This process has resulted in rapid upgrades in a commercial system to a scalable operating system based on very small and efficient software called microkernel technology. Extensions such as real time services and distributed and replicated file systems are under development.

• New technologies are providing a scalable, modular approach to mass storage performance and archiving needed in the new large scale parallel computing systems:

  -Prototype scalable mass storage systems that use parallel arrays of inexpensive disk drives to achieve both high aggregate data transmission rates and large storage capacity have been demonstrated. These systems demonstrate an approach that is the basis for a new generation of high performance file servers and mass storage systems that are internal to scalable parallel computing systems.

  -Petabyte mass storage systems, which can hold images from about 50 university libraries, are now available using commodity storage modules with automated robotic transfer to multiple read/write units. These systems help meet the dramatically increasing requirements for mass storage -- from storing library information to storing remotely sensed satellite imagery.

• Evolving advanced component technology is being employed in early experimental computer systems. This technology will form the basis for a new generation of higher performance, physically smaller, and more affordable computing systems. Examples include the following:

  -Single chip nodes that integrate processing, storage and communications, new systems software, and new development environments, have been demonstrated. These have the potential of providing very cost-effective scalable computing using these single chip or fine grained nodes.

  -Multichip modules are being studied in experiments to determine the optimal design for future scalable units.

• Supporting technologies that enable the rapid design, prototyping, and manufacturing of HPCS systems have made an important contribution to HPCS progress. Examples of rapid prototyping facilities used by researchers include:

  -A laser direct write multichip module tool and associated design capability has been developed to reduce the prototyping time of new modules from months to two weeks. This enables designs to be developed more rapidly, and allows for the exploration of more effective and cost-effective alternatives.

  -New algorithms have been incorporated into design systems that extend synthesis to be applied to new technologies such as field programmable gate arrays and various integrated circuit technologies.

  -A model "factory of the future," linking advanced design technologies from workstations to large scalable computing, was completed and coupled to a prototype factory (described in the Case Studies section). These technologies form the basis of a new generation of computational prototyping, exploiting networked and distributed design processes for rapidly prototyping future generations.

• New competitive contractual mechanisms have been developed to enable the timely purchase of experimental systems. These joint government-industry research projects allow experimental use and early evaluation by a variety of user communities, which in turn provides early feedback to the hardware vendors and to developers of associated software technology. Such projects accelerate the maturation of these complex technologies in preparation for their larger scale use by the ASTA component.

Next section

Blue Book Table of Contents

NCO/HPCC Home Page