The Community for Technology Leaders
Green Image
This paper describes the design and implementation of an application-specific digital architecture aimed at performing the mining of fuzzy databases, a task whose computational weight is very high. The system described here is an array-processor architecture in which the database is split into several units that apply the same operations to different data stored in local memories at the same time. A chip set implementing the architecture has been built using a semicustom approach and it is fully functional. The sustained memory bandwidth of a system featuring 16 units is 5,120 Mbits/s and provides a speedup of 500 times over the software algorithm running on a Sun Sparc2 workstation.
fuzzy coprocessor, database retrieval

R. Guerrieri, G. Baccarani, A. Ferrari and A. Bellettini, "An ASIC Chip Set for Parallel Fuzzy Database Mining," in IEEE Micro, vol. 16, no. , pp. 60-67, 1996.
81 ms
(Ver 3.3 (11022016))