PARALLEL COMPUTING STRUCTURES AND ALGORITHMS FOR LOGIC DESIGN PROBLEMS.
Abstract
Solution of large logic design problems include manipulation, simplification, minimization and realization of logic equations, state table reduction, state assignment and realization of sequential machines, digital system simulation, and wiring simplification. To efficiently solve logic design problems a computing structure that fits the problem structure is needed. Parallel versions of eleven algorithms in the above areas are developed to help determine a more efficient computing structure. Several types of parallel computing structures are examined, and the associative structure is found to be best suited for logic design problems. A restructurable system and a pseudo-restructurable system are proposed to overcome some of the problems which exist in present associative structures. A machine instruction set is proposed for the two systems. Advantages of associative systems are demonstrated by three machine language programs - iterative consensus, row and column dominance, and function multiplication. Programming is simplified because many of the needed operations can be done in parallel with a single instruction and because of similarities between the algorithm structures and the system structure. Parallel languages are also examined, and a parallel language for solution of logic design problems is proposed. A specialized associative computing system is shown to effectively solve many logic design problems. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 11, 1969
- Accession Number
- AD0699218
Entities
People
- C. H. Roth Jr
- Roy Miller Matney Ii
Organizations
- University of Texas at Austin