Architectural Evaluation of a Semantic Network Machine
This paper describes the architecture of the semantic network machine IXM and simulation results on its performance. IXM consists of an associative network with a large number of processing elements connected to it. Each processing element has two associative memories in order to take advantage of the possibility of parallel execution of basic operations in semantic network: association, set operation, and marker propagation. The associative network is a functional network whose node processor has also an associative memory for parallel marker propagation. Execution time of IXM does not necessarily depend on the size of semantic network because the basic operations of IXM can be executed in parallel.
KeywordsPyramid Cese Paral
Unable to display preview. Download preview PDF.
- S.E. Fahlman, “Design sketch for a million NETL machine”, Proc. of First Annual National Conf. on AI, 1980.Google Scholar
- G.E. Hillis, “The connection machine”, MIT Press, 1985.Google Scholar
- D.I. Moldovan and Yu-Wen Tung, “”Semantic Network Array Processor”, Univ. of Southern California, Tech. Rep. PPP-84–2,1984.Google Scholar
- K. Handa, T. Higuchi, A. Kokubu and T. Furuya, “Flexible Semantic Network for knowledge Representation”, to appear in Journal of Information Processing Society of Japan.Google Scholar
- T. Higuchi, K. Handa, T. Furuya and A. Kokubu, “A Semantic Network Language Machine”, Proc. of EURO MICRO, 1985.Google Scholar
- T. Higuchi, T. Furuya, H. Kusumoto, K. Handa, A. Kokubu, “The IX Super Computer for Knowledge Based System”, FJCC,1986.Google Scholar