285
Bisection trees and half-quad trees:
Memory and time efficient data
structures for VLSI layout editors *
A. Pitaksanonkul, S. Thanawastien and C. Lursinsap
The Center for Advanced Computer Studies, University of Southwestern Louisiana, Lafayette,
LA 70504-4330, U.S.A.
Received 15 October 1988
Abstract. In this paper, we present two new tree structures, bisection trees and half-quad trees
that can be used as data structures for storing VLSI layouts. The bisection tree is based on
the structure of a 2-d tree and the half-quad tree is constructed based on the bisection tree
and the quad tree. These two trees perform as good as a 4-d tree in terms of speed but require
less memory. Specifically, our experimental results show that, for region queries, the bisection
tree has approximately the same nodes visited as the 4-d tree. In addition, the bisection tree
requires less memory space and simpler procedures. For the half-quad tree, it is shown that it
approximately 50% memory of that used by the 4-d tree which is approximately the same as
the memory used by the quad tree. In addition, the half-quad tree achieves a much better
speed than the quad tree.
Keywords. Layout, database
1. Introduction
One of the problems of designing an interactive layout editor in VLSI design
systems is to find an efficient data structure to represent a VLSI layout. A
suitable data structure would save space and running time in various geometrical
operations, such as, reporting all rectangles intersecting a particular area. This
operation is important in the context of implementing the design rule checking,
* This work was partially supported by a grant from the Board of Regents, State of Louisiana,
LEQ SF (1987-89)-RD-A-29.
Elsevier Science Publishers B.V.
INTEGRATION, the VLSI journal 8 (1989) 285-300
0167-9260/89/$3.50 © 1989, Elsevier Science Publishers B.V.