An IC/MCM Timing-Driven Placement Algorithm Featuring Explicit Design Space Exploration

A genetic algorithm for building-block placement oflCs and MCMs is presented which si multaneouslyminimizes layout area and an Blmore-based estimate of the maximum path de lay while trying to meet a target aspect ratio. Explicit design space exploration is performed by using a vector-valued, 3-dimensional cost function and searching for a set of distinct so lutions representing the best tradeoffs of the cost dimensions. From the output solutions, the designer can choose the solution with the preferred tradeoff. This approach eliminates the inherent weight and constraint specification problems of existing multi-objective placement methods, in which a weighted sum is minimized subject to user-defined constraints. Very good experimental results are obtained for various placement problems.