The Evolution of the PDP-11
C. GORDON BELL and J. CRAIG MUDGE
A computer is not solely determined by its architecture; it reflects the technological, economic, and organizational aspects of the environment in which it was designed and built. In the introductory chapters the nonarchitectural design factors were discussed: the availability and price of the basic electronic technology, the various government and industry rules and standards, the current and future market conditions, and the manufacturing process.
In this chapter one can see the result of the interaction of these various forces in the evolution of the PDP- 11. Twelve distinct models (LSI-11, PDP-11/04, 11/05, 11/20, 11/34, 11/34C, 11/40, 11/45, 11/55, 11/60, 11/70, and VAX-11/780) exist in 1978.
The PDP-l1 has been successful in the marketplace: over 50,000 were sold in the first eight years that it was on the market (1970-1977). It is not clear how rigorous a test (aside from the marketplace) the design has been given, since a large and aggressive marketing organization, armed with software to correct architectural inconsistencies and omissions, can save almost any design.
Many ideas from the PDP-11 have migrated to other computers with newer designs. Although some of the features of the PDP-l1 are patented, machines have been made with similar bus and instruction set processor structures. Many computer designers have adopted a unified data and address bus similar to the Unibus as their fundamental architectural component. Many microprocessor designs incorporate the PDP-l1 Unibus notion of mapping I/O and control registers into the memory address space, eliminating the need for I/O instructions without complicating the I/O control logic.
It is the nature of computer engineering to be goal-oriented, with pressure to produce deliverable products. It is therefore difficult to plan for an extensive lifetime. Nevertheless, the PDP-l 1 evolved rapidly over a much wider range than expected. An outline of a family plan was set forth in a memo on April 3, 1969, by Roger Cady, head of the PDP-11 engineering group at the time (Table 1). The actual evolution is shown in tree form in Figure 1 and is mapped onto a cost/performance representation in Figure 2.