Computer Structures

This document, "Computer Structures: What Have We Learned From the PDP-11?", published on November 8, 1975, reflects on the PDP-11's design and its impact. The authors, Gordon Bell and William D. Strecker, discuss the PDP-11's extensive range in memory size and system price, noting that it stressed the basic minicomputer architecture. The UNIBUS, a key interconnection standard for the PDP-11, has become a de facto standard for many micro and minicomputer systems. The paper details the architectural experience gained through the PDP-11's design and use by examining its environment, including initial goals, constraints, technology, the development organization, and user impact.

The authors highlight that a computer is a product of its total environment, not solely its architecture. They identify key factors affecting design: basic technology, the development organization's nature, and the influence of user groups and government standards. The document also touches upon the evolution of computer technology, the challenges of designing for a long product lifetime, and the pressures from market and implementation groups.

The paper delves into specific design goals and constraints of the PDP-11 in 1970, addressing issues like limited address capability, insufficient registers, lack of stacks, slow context switching, and byte handling. It also discusses the technology that influenced the PDP-11's design, particularly semiconductor memory, and how it shaped various models.

Furthermore, the document explores the organization of people involved in the design process, the evolution of PDP-11 models, and the concept of multiprocessor computer structures. It concludes with a summary of problems encountered in extending the machine's range and a bibliography of relevant works. The paper suggests that while the PDP-11 has been successful, design faults occurred due to late initiation, and future evolution will be interesting to observe, with ultimate testing being its actual use.