previous | contents | next

system (in this case a computer). Up until now, the RTM diagram has provided a two-dimensional representation schema for behavior and structure. The reader has not been burdened with learning a register-transfer language because they are quite cumbersome for expressing digital systems. Since most- computers behave simply, a one-dimensional representation is usually adequate.

The motivation for introducing ISP notation is that it serves as the formal basis for defining computers. A computer can be defined precisely and relatively easily in ISP. The section on the PDP-8/RTM is based on the ISP description; the specification in ISP is converted into an RTM implementation in a relatively formal way. ISP's of many larger computers (e.g IBM 7090, CDC 6600) can be found in Bell and Newell (1971).


An ISP description can also be converted to a programming language (e.g., Fortran) rather easily, so that a computer may be simulated, and the design and description verified. Subsequent design problems are given to describe computers in ISP, carry out ISP-to-RTM, and ISP-to-Fortran conversions.


A section presents several extended RTM's for use in interfacing an RTM system to a computer. With these interfaces, it is possible to build combined systems in which some operations are carried out within the computer (i.e., programmed) and some operations are done in the specialized system (hardware). Thus, the hardware-software trade-off can be exploited in a design.



In this section, we present three aspects of computers using Crtm-1. -The first section describes the computer in a conventional way, without regard to the fabrication. The second section describes the implementation using RTM's; hence, the reader has another opportunity to understand how the computer operates. Finally, the structure and behavior of Crtm-1 are defined again using ISP. This last section will be referenced in subsequent sections for designing computers based on their ISP descriptions.


A computer is usually considered in terms of a structure constructed from the four components shown in Figure Crtm-1; these are: a D for carrying out arithmetic and logical operations; an M for storing the program and information (data) for the program; a T, for communicating with the outside world, which includes a console for direct human control of the computer; and a K, the interpreter, which defines the behavior of the components through the use of instructions. The interpreter operates by picking up (i.e., fetching) instructions from the computer's memory, examining (i.e., decoding) them and then carrying out (i.e., executing) the operation they specify. After each instruction is interpreted, a register within the interpreter, called the Program Counter (also called the instruction address, the instruction location counter, etc.), selects (accesses) the next instruction to be interpreted. Control then repeats the fetching-decoding-execution process for this new instruction. This simple 3 step sequential process is basically how all current, general purpose, stored program digital computers operate.


previous | contents | next