Method for hierarchical parasitic extraction of a CMOS design

This application is related to co-assigned application Ser. No. 09/608,309 filed contemporaneously herewith and incorporated herein by reference. In deep submicron technologies, coupling capacitance significantly dominates the total parasitic capacitance. This causes crosstalk noise to be induced on...

Full description

Saved in:
Bibliographic Details
Main Authors Cano, Francisco A, Savithri, Nagaraj N, Gunturi, Vijaya
Format Patent
LanguageEnglish
Published 26.03.2002
Online AccessGet full text

Cover

Loading…
More Information
Summary:This application is related to co-assigned application Ser. No. 09/608,309 filed contemporaneously herewith and incorporated herein by reference. In deep submicron technologies, coupling capacitance significantly dominates the total parasitic capacitance. This causes crosstalk noise to be induced on quiescent signals which could lead to catastrophic failures. A method is provided for extracting parasitic data in a hierarchical manner from a trial layout of the integrated circuit. Intracellular parasitic data representative each cell type used in the integrated circuit is extracted only once, regardless of the number of times the cell is instantiated in the integrated circuit. For each instance of each cell, a portion of intercell signal lines that are routed over that instance of the cell are cut out in cookie cutter fashion by specifying an area in the trial layout corresponding to the instance of the cell such that the portion of intercell signal lines within the area can be processed apart from the remaining portion of the intercell signal lines. Each cutout portion of the over the cell routing (OCR) is combined with the respective cell instance and OCR parasitic data is extracted with reference to the respective cell. For each cell instance, the intracellular parasitic data derived once for the cell is combined with the OCR parasitic information derived for that cell instance in order to form a coupled simulation model.