Princeton VLSI project

There are three major aspects to our project: ALI, Census, and Testing. ALI, our procedural language for VLSI design and layout, is now up and running. It has already been used by Dobkin and Drysdale to redesign a divider they previously designed on a graphics system at Xerox. In addition, LaPaugh's VLSI class has already used ALI for their VLSI projects. Dobkin is currently beginning to explore ways to create graphics interfaces to the ALI system. The ALI2, the second version of our layout system is coming along well. The census language is a new way to express parallel algorithms that use a fairly loosely coupled method of control. Work is under way to understand the limits and powers of such languages. North is beginning to identify those problem areas that can be successfully mapped onto the census language. We are also thinking about implementations of census, but no implementations is yet started. Our work on testing divides into two areas. Arden is beginning to work on first silicon testing especially with respect to scanning electron microscopes (SEM). In the future we expect that we will be able to use a SEM that the Princeton Siemens group is about to get. LaPaugh and Lipton have begun to work in the area of production testing. They have already successfully been able to completely characterize the testability of prefix computations. Prefix computations arise naturally in a number of places; for instance, in carry look ahead adders. Such characterizations link the self test of these computations with classic semigroup theory.