A research on optimum-searching quadratic optimization for very large-scale standard cell placement

Abstract

This paper presents a research on how to find the optimal or near optimal solution in the specific standard cell placement of very large-scale integration (VLSI) designs. The optimization methods currently used are faced with big difficulties in finding better placement result. This paper studies the quadratic optimization that is frequently adopted in this field and analyzes the difficulty it faces. Meanwhile, this paper suggests a very efficient quadratic programming based optimization method. This method employs the tactic of hybrid search based on multi-space search and search space traversing to try to find the optimum or near-optimum of the problem. Experimental results show that this method can make up with the shortcomings of the current quadratic optimization used in placement to a far extent and gains much placement improvement.