The internal machine

  • Authors:
  • Andrew M. Welin

  • Affiliations:
  • North Electric Co., Delaware, Ohio

  • Venue:
  • HLLCA '73 Proceedings of a symposium on High-level-language computer architecture
  • Year:
  • 1973

Quantified Score

Hi-index 0.00

Visualization

Abstract

Two results are presented in this paper: an enlargement of the concept of machine languages and a way to design computers based on an inner machine oriented toward interpretation of machine languages in general. In Part I we discuss attributes of "directly interpretable languages" and processing mechanisms related to their interpretation. "Sequential extraction" allows an instruction to be of arbitrary form, while "linear decoding" allows us to organize the interpreting routines into structures. This leads to powerful machine language level capabilities. Using the mechanisms of Part I, we build in Part II the case for the "internal machine" approach to computer design. The internal machine (unlike microprocessors) is not oriented towards low level (gate, register) manipulation of data, but its operations reflect elementary actions related to the process of interpretation. The advantages of the internal machine include simplicity, programming interface (as opposed to the interface offered by microprocessors), and the ability to handle directly interpretable languages.One can choose among many architectures for the internal machine. An organization based on the execution stack is suggested as the most general solution. The essential "general purpose" organizations of the internal machines lend themselves to extensions and specializations which provide extra efficient handling of specific subprocesses related to interpretation.