previous | contents | next

448 EVOLUTION OF COMPUTER BUILDING BLOCKS

CONCLUSIONS

The concept of using high level building blocks is not new, but we think this particular implementation of a set of simple blocks is quite useful to many digital systems engineers. The design time using this approach is significantly less than with conventional logical design. The modules are especially useful for teaching digital system design. We have solved many benchmark designs with reasonably consistent results. The modules can be applied quickly and economically where there are between 4 and 100 control steps, a small read-write memory (100 words), and perhaps some read-only memory. Larger system problems are usually solved better with a stored program computer, although such a computer can be designed using RTMs. The user need only be familiar with the concept of registers and register operations on data, and have a fundamental understanding of a flowchart.

ACKNOWLEDGEMENT

These modules were formally proposed in March 1970 in a form essentially described herein by one of the authors, C. G. Bell. In June 1970 the project was seriously started by constructing the computer of the previous example using them. The authors gratefully acknowledge the organization and management contributions of F. Gould, A. Devault, and S. Olsen (Digital Equipment Corporation) without whose goal-oriented commitment the RTMs could not have been built. The authors are also indebted to Mrs. D. Josephson of Carnegie-Mellon University for typing the manuscript.

previous | contents | next