Array architectures for iterative algorithms

Article Abstract:

Regular mesh-connected arrays are analyzed and found to be isomorphic to a class of regular iterative algorithms, and the method by which algorithms can be translated into arrays is demonstrated as systematic. Several generally recognized systolic arrays are demonstrated to be specific cases of the types of architectures that can be derived by techniques presented, including arrays for Fourier transform, matrix multiplication, and sorting. The technique can be fully automated, produces multiple choices for algorithms, allows regularly iterative algorithms to be mapped into fixed-size mesh-connected processor arrays, and handles a large class of algorithms.

Sorting, Mathematical Proofs, Matrix Computations, Arrays, Fast Fourier Transforms, technical, Mesh Networks

---

Task-Oriented Architectures

Article Abstract:

The factory of the future is a concept of an essentially peopleless, paperless organization. The factory should be capable of producing a wide variety of parts on demand. Speech recognition, speech synthesis and image processing tasks are examples of sensor-interpretation problems. These problems require the development of special purpose processors. A number of issues are to be considered while designing task-oriented systems. A block diagram shows architecture of the Harpy machine.

Software, Technology, Image processing, Design, Hardware, Task analysis, Evaluation, Factory management, Future Technologies

---

Advanced parallel processing with supercomputer architectures

Article Abstract:

Parallel processing techniques and new architectures can boost supercomputer performance. Performance enhancement methods, parallel algorithms, the programming environment, control of concurrency, resource management, compiling techniques, parallel languages, and architectural choices are evaluated for the best solution.

Architecture, Supercomputers, Supercomputer, Parallel processing, Performance improvement (Computers), Performance Improvement

Similar abstracts:

• Abstracts: Scientific modeling with massively parallel SIMD computers. Mapping vision algorithms to parallel architectures
• Abstracts: Algebraic tools for the performance evaluation of discrete event systems. An Algebraic Formula for the Output of a System with Large- Signal Multifrequency Excitation
• Abstracts: Services and Protocols of the Data Link Layer. Operation of higher layer data communication protocols over satellite links
• Abstracts: Lateral resonant tunneling transistors employing field-induced quantum wells and barriers. Parallel Operation of Power Transistors in Switching Amplifiers

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