For future parallel data processing and 5th generation computing systems, novel associative circuits and processing elements are proposed for research and development. These VLSI circuits will ,,,,feature high density, high performance, radiation hardness and wide operating temperature range. The most promising approach for intelligent non-von Neuman type of systems has been based on associative or content addressable memories, but their implementation has been hindered by the uneconomical size and by the susceptability to device parameter variations of the associative memories. The main objective of this project is to provide associative memories and processing elements for advanced adaptive array processors, which feature the density, performance and reliability of comparable static random access memories. The effort in Phase I will include feasibility establishment of novel associative memory circuits by computer-aided analyses and simulations as well as the design of a test chip. In Phase 11 a complete content addressable memory and an associative processing element will be developed in addition to the fabrication and evaluation of the test chip. The results will provide key elements for intelligent knowledge based robots, controlling and processing systems as well as for large computing problems. In systems, the characteristics of speed, power, tolerance to extreme environments, weight, reliability, maintenance, and life time will be improved significantly.The potential commercial application as described by the awardee: This research program could lead to a breakthrough in designs of intelligent systems. Robots, controllers, and numerous processors require associative VLSI circuits for knowledge-based systems operating with high performance or in extreme environments. Solutions of large computing problems need content addressable circuits for parallel data processing. Microcirc Associates is a professional organization of experts in the field of microelectronics for research, development, and design of advanced semiconductor circuits.
