The architecture of a large associative processor

This paper will describe features of architectural significance to the segmentability of a processor; it is not intended to be a detailed description of a processor for Information Storage and Retrieval. We regret that the incorporation of some features cannot be defended here because of the length of this paper. They are presented in a report. We first state the types of problems to be processed. This will lead to the overall organization of the processor. In Information Storage and Retrieval, a processor should have the capability to store data which is formatted as ordered sets or unordered sets, and to retrieve all such sets having a specified subset. An unordered set search for a given subset S retrieves all sets containing S. An ordered set search for a given ordered subset S retrieves all ordered sets containing S. A string search for a given string S retrieves all ordered sets (strings) having a substring S. For example, if S = (s1, s2, s3) and S1 = (s1, a, s2, s3), S2 = (a, b, s1), s2, s3, c, d), S3 = (s2, s1, s3) and S4 = (s1, a, b, s2). Then an unordered set search for S would retrieve S1, S2, S3, an ordered set search for S would retrieve S1 and S2, and a string search for S would retrieve S2.