An efficient list representation and access scheme for artificial intelligence applications

摘要

The rapid increase in artificial intelligence utilization has increased the demand for list-oriented processing. For an architecture to support such environments, an efficient list representation and access scheme must be developed. Such a scheme, called implied CDR coding, which optimizes memory utilization by taking advantage of the linearity property of lists, while maintaining ease of implementation of list instruction primitives, is proposed. The further enhance performance, a hardware-supported pointer-access algorithm is introduced for use when list-access primitives encounter pointers within the list.