Placement is one of the most important problems in electronic design automation (EDA). An inferior placement solution will not only affect the chip’s performance but might also make it nonmanufacturable by producing excessive wirelength, which is beyond available routing resources. Although placement has been extensively investigated for several decades, it is still a very challenging problem mainly due to that design scale has been dramatically increased by order of magnitudes and the increasing trend seems unstoppable. In modern design, chips commonly integrate millions of gates that require over tens of metal routing layers. Besides, new manufacturing techniques bring out new requests leading to that multi-objectives should be optimized simultaneously during placement.

Our research provides high performance algorithms for placement problem. We propose (i) a high performance global placement core engine POLAR; (ii) an efficient routability-driven placer POLAR 2.0, which is an extension of POLAR to deal with routing congestion; (iii) an ultrafast global placer POLAR 3.0, which explore parallelism on POLAR and can make full use of multi-core system; (iv) some efficient triple patterning lithography (TPL) aware detailed placement algorithms.