Computations on the Massively Parallel Processor at the Goddard Space Flight Center

Article Abstract:

The Massively Parallel Processor was installed at the Goddard Space Flight Center in May 1983 and has been used since then with great success. Scientists use the MPP for its great computing speed on any data set that can be put into the form of a multidimensional array. Advanced image analysis algorithms, scientific simulations, and data manipulation techniques are performed on the MPP that would be impractical to do on standard serial architectures. Four important algorithms that have been developed include an algorithm for the automatic determination of depth from stereo imagery; an algorithm that is part of a program that explores production of ordered results from processes that begin with random initial states; an algorithm that uses region growing techniques to do a natural segmentation of images; and an algorithm that performs a bitonic sort of a string of data using the MPP architecture.

Space research and technology, Technology application, Algorithms, Scientists, Imaging systems, Algorithm, United States. Goddard Space Flight Center, Scientific Research, Sorting, Massive Parallelism, Imaging Technology

---

Null-field computations of radar cross sections of composite objects

Article Abstract:

The null-field approach to the computation of radar cross sections (RCSs) of composite objects (electromagnetic scatterers) is extended to several additional classes of such objects, including star-shaped objects and star-shaped objects with one or more concavo-convex parts. The solution of the scattering problem for each shape is solved by determining the total transition matrix for the object. Analysis of numerical implementations of the null-field approach for each object finds that convergence to a solution of the RCS of each object can be achieved for an electromagnetic frequency interval extending into the resonance region. Details of the null-field approach and validation of its application are described.

Radar, Electromagnetic radiation, Object Recognition, Radar Systems, Validation, Data Analysis

---

Three-Dimensional Wave Equation Computations on Vector Computers

Article Abstract:

The vector computer and systems such as the CRAY-1 and the Cyber 205 have made computation of realistic three-dimensional seismic models possible. A complete modeling system requires a program to compute pressure as a function of time and space in addition to data entry, graphic output, post model data analysis, and interpretation programs. A two-dimensional model for the VAX-11 with FPS-100 Array Processors was modified to three dimensions and vectorized for the Cyber 205. Typical output and timing results are presented.

Petroleum industry, Supercomputers, Supercomputer, Seismology, Geophysics, Models, Scientific Computers, Energy Exploration

Similar abstracts:

• Abstracts: Prolog to model based processing of signals: a state space approach. Ultra-high-speed modulation-doped field-effect transistors: a tutorial review
• Abstracts: Array architectures for iterative algorithms. Advanced parallel processing with supercomputer architectures. Task-Oriented Architectures
• Abstracts: Computational models and resource allocation for supercomputers. Hypercube supercomputers

This website is not affiliated with document authors or copyright owners. This page is provided for informational purposes only. Unintentional errors are possible.