Date: Tue, 26 Nov 1996 00:30:24 GMT Server: NCSA/1.5.2 Last-modified: Fri, 11 Oct 1996 16:00:57 GMT Content-type: text/html Content-length: 6381 The Hypercomputing and Design Project

The Hypercomputing and Design Project


HPCD logo


The Web pages for HPCD are under construction

OVERVIEW

The broad objective of HPCD is to build on top of advances in massively parallel computing (hypercomputing), in computational science, in AI, and in modeling/simulation technology, and to develop a new generation of engineering automation technology that can bring about dramatic gains in productivity of the national industrial base. The top-level goal is to develop hypercomputing methods for attaining order-of-magnitude speedups in the time required for transition from an innovative design concept, or/and a scientific advance, to a useful high-quality product. The focus of the effort is on design of complex engineering systems (such as computers, ships, jet engines), where the design process is strongly dependent upon the use of knowledge from several scientific disciplines. The main technical goals of the project are as follows:
  1. To develop methods that use hypercomputing power, in conjunction with scientific knowledge in a domain, AI, and Visiometric techniques, in order to build computational models (and related simulations) in the domain that range from the highly accurate, but costly, models needed for scientific discovery and design verification to the greatly reduced and approximate, but cheap, models needed to explore large spaces of design alternatives.
  2. To develop methods and tools, based on AI and other related areas of computer and computational science, to help in the setup and control of complex simulations and the automatic analysis and interpretation of massive datasets produced from simulation runs.
  3. To develop methods and tools that use computational models efficiently to manage/control the formulation and solution of complex design problems.

    HPCD is an ARPA/ITO sponsored project. Within the framework of the national HPCC program, the project is intended as a major step towards building a comprehensive national resource for HPC-based computational methods and facilities that can address the 'grand challenge' of design.

    The project is organized as a large scale, multidisciplinary, multi-institutional, consortium involving researchers from academia and industry. Participating universities include Rutgers, USC and Princeton. At present, industrial participants include SAIC, Cambridge Hydrodynamics Inc. (CHI), GE, Lockheed, AeroHydro, IBM, AMD, Sematech, Bellcore, UTRC, Boeing, Schlumberger, and INTEL. Other research collaborations have been established with researchers from several government and industrial labs. The Principal Investigator and Co-PI's are:

    The component efforts of HPCD are designed to address the three main technical goals of the project, by proceeding 'bottom-up' via exploration of selected concrete design problems, and by developing methodologies and tools in a manner which is strongly influenced by the 'pull' of these problems. To achieve cohesion in the overall res-arch effort, the component efforts are organized in two main clusters: Cluster I concentrating on design and manufacturing of computers and of related microstructures; and Cluster II concentrating on design tasks that are critically dependent on Computational Fluid Dynamics (CFD). There is also a third supporting cluster of core work, Cluster III, whose emphasis is on computer/computational science foundations and methodologies that are related to work in the two main clusters. Each cluster includes efforts in two or three research areas. Collaboration among researchers, and synergy among component efforts, are an integral part of the approach in the project.

    RESEARCH AREAS

    Cluster I. Hypercomputing in Computer System Design and Manufacturing

    Area I.1. Tools for design of microlithographic processes in sub-1/4 micron electronic chip manufacturing

    Area I.2. Design of Computers

    Cluster II. Hypercomputing in Design Tasks Supported by Computational Fluid Dynamics (CFD)

    Area II.1. Design of Propulsion systems for aerospace vehicles

    Area II.2. Innovative Ship Design

    Area II.3 Design of 'voice mimic' speech generation systems

    Area II.4. Conceptual Design of Air Vehicles

    Cluster III. Core Research in Hypercomputing and Design

    Area III.1. Artificial Intelligence (AI) methods and systems in design

    Area III.2. Visiometrics and modeling for ship design and vortex flows

    Area III.3 Software Tools and algorithms for HPCD


    Personnel

    HPCD Reports

    Links to Web pages of our collaborators, and other related organizations

    Access Statistics


    Maintained by armhold@www.cs.rutgers.edu

    This page last updated: Fri Oct 11 12:00:42 EDT 1996