COMPACTION WITH AUTOMATIC JOG INTRODUCTION
暂无分享,去创建一个
This thesis presents an algorithm for one-dimensional compaction of VLSI layouts. It differs from older methods in treating wires not as objects to be moved, but as constraints on the positions of other circuit components. These constraints are determined for each wiring layer using the theory of planar routing. Assuming the wiring layers can be treated independently, the algorithm minimizes the width of a layout, automatically inserting as many jogs in wires as necessary. It runs in time O(n ) on input of size n. Several heuristics are suggested for improving the algorithm''s practical performance. The compaction algorithm takes as input a data structure called a sketch, which explicitly distinguishes between flexible components (wires) and rigid components (modules). The algorithm first finds constraints on the positions of modules that ensure enough space is left for wires. Next, it solves the system of constraints by a standard graph-theoretic technique, obtaining a placement for the modules. It then relies on a single-layer router to restore the wires to each circuit layer. An efficient single-layevery wire, though not the number of jogs. As given, the compaction algorithm applies only to a VLSI model that requires wires to run a rectilinear grid. This restriction is needed only because the theory of planar routing (and single-layer routers) has not yet been extended to other models. The compaction algorithm''s correctness proof elucidates the assumptions on which the algorithm depends, so that the algorithm is easily generalized once the necessary theoretical machinery is in place.
[1] M. Y. Hsueh,et al. Symbolic layout and compaction of integrated circuits , 1980 .
[2] Robert E. Tarjan,et al. Fibonacci heaps and their uses in improved network optimization algorithms , 1987, JACM.
[3] R. Y. Pinter,et al. THE IMPACT OF LAYER ASSIGNMENT METHODS ON LAYOUT ALGORITHM FOR INTEGRATED CIRCUITS , 1983 .