Implementation Tool Machine Learning Features Evaluation

In recent trends, machine learning in EDA tool features is a hotcake. It covers floorplanning, placement optimization, congestion optimization, routing optimization, as well as design space optimization with artificial intelligence capability. With machine learning capability, the design is expected to converge faster with a higher quality of result in timing, power, and routing utilization. This paper will review the current artificial intelligent state of the arts, in chip design optimization. The paper will discuss the type of artificial model deployed in the field as well as the preferred model in the field. The evaluation is done on commercially supported artificial intelligent features in the EDA tool, at various implementation stages, including stability tests of the commercial artificial intelligent features. From the finding, machine learning does improves the predictability of quality of result however, there is a need for forward looking controller to ensure the optimum quality of result from combined machine learning features. Hopefully, the finding of existing artificial intelligent features supported and limitations will benefit the design community.