Integration of Tools for the Design and Assessment of High-Performance, Highly Reliable Computing Systems (DAHPHRS)
Abstract
Systems for Space Defense Initiative (SDI) space applications typically require both high performance and very high reliability. These requirements present the system engineer evaluating such systems with the extremely difficult problem of conducting performance and reliability trade-offs over large design spaces. A controlled development process supported by appropriate automated tools must be used to assure that the system will meet design objectives. This report will describe an investigation which examined methods, tools, and techniques necessary to support performance and reliability modeling for SDI systems development. Models of the JPL Hypercube, the Encore Multimac, and the C.S. Draper Lab Fault-Tolerant Parallel Processor (FTPP) parallel computing architectures using candidate SDI weapons-to-target assignment algorithms as workloads were built and analyzed as a means of identifying the necessary system models, how the models interact, and what experiments and analyses should be performed. As a result of this effort, weaknesses in the existing methods and tools were revealed and capabilities that will be required for both individual tools and an integrated toolset were identified.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1991
- Accession Number
- ADA252777
Entities
People
- Charlotte O. Scheper
- Harold L. Waters Iii
- Robert L. Baker
Organizations
- RTI International