Browse Prior Art Database

Automatic PLA Generator

IP.com Disclosure Number: IPCOM000089812D
Original Publication Date: 1977-Dec-01
Included in the Prior Art Database: 2005-Mar-05
Document File: 1 page(s) / 12K

Publishing Venue

IBM

Related People

Halliwell, H: AUTHOR

Abstract

A procedure by which efficient PLA personalization can be generated automatically from high level functional specifications is described. The procedure consists of three steps, each based on existing techniques and including modifications and extensions. An assumption is made that the functional specification is coded in some suitable high level language. In this form it is amenable to inspection, and it can be compiled to provide a fast simulation.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 60% of the total text.

Page 1 of 1

Automatic PLA Generator

A procedure by which efficient PLA personalization can be generated automatically from high level functional specifications is described. The procedure consists of three steps, each based on existing techniques and including modifications and extensions. An assumption is made that the functional specification is coded in some suitable high level language. In this form it is amenable to inspection, and it can be compiled to provide a fast simulation.

(1) The first step is to transform this specification into an equivalent random logic design. This need not be an efficient design or conform to any constraints (e.g., powering) required for a real random logic implementation.

(2) The second step is to convert this random logic design into a preliminary version of a PLA personalization. There are various algorithms that can be used to do this, for example, Roth's pi-star algorithm described in [1]. The result of this conversion is not, in general, an efficient PLA personalization, and it requires optimization before it is useful. To assist in this optimization, it is convenient to generate the OFF - as well and the ON array here.

(3) The final step is the optimization which is based on Roth's SHRINK algorithm described in [2] with modifications as described below. This does not strictly guarantee an optimum result, but is very fast and in practice usually produces a personalization as efficient as a manual one. Modification of the existing algo...