Microcode simulator for Apple Macintosh

Abstract

No fewer than three microcode simulators have been described in this bulletin since 1987. Carlson's [1] runs on an IBM mainframe, Donaldson's [3] on DOS machines and Dunworth's [4] on DOS or unix machines. All have a traditional textual interface. We wanted a visual interface that would display a circuit diagram (figure 1) and let a user follow the execution of a microprogram by watching the action of the gates. The first version became operational in the fall of 1990 and has survived two academic years of use by undergraduates.Microcode has traditionally been written using an obtuse mnemonic with such instructions as "a=7, b=3, ab, add, ramf, mar=y". We did not want students to have to contend with that kind of syntax, so our users write microcode by responding to a series of dialogs . The user makes a choice either by the "point and click" technique or by typing a keystroke equivalent. For example, the keystroke corresponding to the addition operation is "+".When the input #1 dialog comes up, the user can put register 7 on input bus "a" by typing "7". Keystrokes are buffered, allowing "power programmers" to type as fast as they wish. The dialogs can keep up with most programmers on any of the various Mac-II models. Even on the slower Mac SE, the standard sequence of eight dialogues can be completed in approximately four seconds, which is considerably less time than it takes to flawlessly type "a=7, b=3, ab, add, ramf, mar=y". After two years of field testing, several improvements suggested by students have been incorporated into the editor. Students, with the exceptions of a few intransigent DOS zealots, are now generally satisfied with the operation of the microcode editor.