Marco Negrete on Structured VLSI

Viewpoints †" Marco Negrete on Structured VLSI Design The rewards of very large-scale integration are great, but breakthroughs are needed in design capabilities. VLSI Daniel Strategies and Tools, by William J. Haydamack and Daniel J. Griffin A survey of present approaches and possible future directions at Hewlett-Packard. In this Issue: If the decade of the sixties was the decade of electronics and the seventies the decade of i the computer, what will characterize the decade of the eighties? This issue explores a likely candidate: the very large-scale integrated (VLSI) circuit. The size of the VLSI challenge †" the design such production of tiny chips containing hundreds of thousands of transistors †" is such that it design tax the creativity and ingenuity of everyone involved. To reduce the design time for such design circuits to a reasonable amount, the complete design process must be ad-_ dressed. It is not enough simply to create tools for VLSI designers. The creation of sophisti-. ¿LI üü^JS cated design tools will help, but a balanced strategy encompassing tools, methodologies and quick-turnaround fabrication must he employed if the design time is to be significantly reduced. In this that you will have an opportunity to review some of the VLSI design concepts that have worked for Hewlett-Packard Company. Design strategies and design tools are continuously evolving at Hewlett-Packard divisions and in Hewlett-Packard Laboratories. The combination of industrial development and university research is quickly moving 1C design capability into the hands of hundreds of very creative engineers who have not had access impact development additional degree of freedom before. This design strength will impact new product development throughout the world. With this new strength, we may find that in the future we will be limited more by the capital required to build 1C facilities than by process technologies or design resources. WITH THE CAPABILITY of putting 450.000 devices on a chip, enough to implement a 32-bit microcomputer, designers are well into the VLSI era (see box, page 11). If the over of 2-to-l increases in density per year established over the last decade continues, we will soon be talking about ultra-LSI. The rewards are great as higher levels of integration are achieved. The most obvious benefit is the dramatic decrease in the cost per function as the cost of a chip is amortized over a larger number of functions. Equally important are the increases in …