Analysis and design of a cost-effective associative processor for weather computations

Abstract

An associative processor organization with application to weather computations, and specifically, to a two-level atmospheric general circulation model developed by Mintz and Arakawa is used as the application problem to be implemented for system analysis and design. These system parameter changes include the arithmetic-logic capabilities, the number of processing elements, different speeds, structures, and access techniques of the memory, addition of a high speed temporary storage, and the use of an auxiliary storage unit. Pertinent data are calculated for all these systems and a cost-effective system is finally determined after the comparisons of these analysis results are made. This final system is then evaluated and compared with the original system.

The work reported here was performed at Syracuse University and was supported by The Air Force Rome Air Development Center under Contract F30602-72-C-0281.