Browse Prior Art Database

Hybrid Macromodular Network Analysis

IP.com Disclosure Number: IPCOM000083551D
Original Publication Date: 1975-Jun-01
Included in the Prior Art Database: 2005-Mar-01
Document File: 2 page(s) / 13K

Publishing Venue

IBM

Related People

Rabbat, NB: AUTHOR

Abstract

As advances are made further into large-scale integration (LSI), the data associated with a design becomes voluminous and complex. The problem of analyzing complete chip performance outstrips the ability of traditional circuit analysis programs to determine the transient response, power dissipation, noise sensitivities, etc., of thousands of circuits per chip.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 2

Hybrid Macromodular Network Analysis

As advances are made further into large-scale integration (LSI), the data associated with a design becomes voluminous and complex. The problem of analyzing complete chip performance outstrips the ability of traditional circuit analysis programs to determine the transient response, power dissipation, noise sensitivities, etc., of thousands of circuits per chip.

Modern circuit analysis programs can accommodate special functions. Thus the capability of existing programs can be extended by providing a presolved system of differential equations symbolically for commonly used circuits. Using these presolved equations, the ability of these programs is greatly enhanced to solve configurations with large numbers of such common circuits.

To illustrate the savings of the presolved macromodular approach compared to traditional analysis, consider using an existing circuit analysis program for a configuration which includes 15 transistor-transistor logic (TTL) networks. In a discrete description there are N nodes = 212, and E branches = 540; the total link and cutset variables are 7, 192 and 696, compared with 33 and 102 employing the macromodular approach which is an average saving of 1:100. These savings increase exponentially with the number of circuits, permitting accurate statistical analysis of LSI circuits.

While initial effort is required to generate these presolved networks symbolically, they give the user great power to look at a large assembly of LSI circuits. Each circuit is solved for its time-dependent output as a function of inputs, and only the solution is entered into the main program symbolically. A circuit manual is provided for the user with a corresponding presolved system of macro equations and their required...