Abstract : Two different approaches to high performance systems for Artificial Intelligence (AI) applications have been set forth at Berkeley: one by the SPUR (Symbolic Processing Upon RISC) group, the other by the Aquarius group. In particular, the SPUR group proposed multiple, homogeneous, RISCs (Reduced Instruction Set Computers) with a functional programming (LISP) paradigm while the Aquarius group proposes multiple, heterogeneous, specialized processors with a logic-programming (Prolog) paradigm. The two groups agreed to meet for the purpose of studying the relative merits of their two approaches and to report on their findings. Meetings were held in October and November of 1984. The meetings consisted of descriptions of the two projects and discussions of the various alternatives available. This report delineates the results of those meetings; in particular, the dissimilar objectives of the two projects are stated and a point by point discussion of the major issues are provided. The two projects are now more similar in certain aspects in which they differed a year ago. The authors describe the two projects as the are now.
A. Despain, R. Katz, Y. Patt
Journal name not available for this finding