skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: The creation of an alternating sequential/parallel Lisp machine

Miscellaneous ·
OSTI ID:5897173

The theory, design, and fabrication of an alternating sequential/parallel Lisp machine (ASPLM) is described. The ASPLM provides an environment for symbolic computation, capable of solving artificial intelligence problems using parallel algorithms. The ASPLM is an innovative hardware/software system, resulting from the integration of the following new ideas: an extensible hardware architecture which directly supports the alternating sequential/parallel (ASP) methodology, a coordinated garbage collection mechanism for multicomputers, an efficient Lisp identifier-passing scheme, a tokenized protocol for the transmission of symbolic Lisp expressions, and Lisp language extensions which support parallel programming. This dissertation covers the following topics: Background - The critical decisions in designing the ASPLM architecture were based on a survey of existing parallel computer architectures. A comparison between the ASPLM and other modern Lisp machine architectures demonstrates the relevance and importance of this research. ASPLM architecture - The fundamental components of the ASPLM architecture are described. The host-slave communications scheme, which shows how symbolic expressions can be efficiently processed on this architecture, is presented. Hardware design - Important aspects of the ASPLM hardware organization are discussed. A complete description of the host computer, the host interface, the bus repeater, and the slave processors is provided. Schematic diagrams and photographs of the ASPLM hardware are included. Software design - An analysis of the software organization reveals the close interrelationships between the various component modules which constitute the ASPLM Lisp system. The ASPLM Lisp system software is written in the C language and 68000 assembler. All source code is included.

Research Organization:
Cincinnati Univ., OH (USA)
OSTI ID:
5897173
Resource Relation:
Other Information: Thesis (Ph.D)
Country of Publication:
United States
Language:
English